ms. perceived

The Unfinished Loop

Avantika Mehra — Mon, 24 Aug 2026 02:53:57 GMT

An excerpt from a longer reflection on the state of AI (December 2025)

IV. Can AI alone move us forward?

Can LLMs do more than audit inherited knowledge? Can they produce and accelerate knowledge wholesale? And if so, how quickly does this transform our economy?

Endogenous progress is progress that comes from within the system, rather than being imposed from outside (e.g., through random discovery or additional capital). The tools that are products of innovation become instruments for the next cycle of innovation.

One way to approach this is through the lens of endogenous acceleration: innovation that compounds because the system improves the process of innovation and knowledge creation itself, not just human productivity. Will AI remain an assistive tool and productivity enhancer, or become a mechanism for sustained, unbounded domain-general knowledge creation?

Historically, we’ve seen only weaker forms of endogeneity:

Narrow endogeneity consists of tools that amplify a largely human loop without mechanizing its improvement. The microscope and telescope widened our aperture onto the universe by making new parts of it observable, but humans were required to supply the hypotheses and standards of proof to derive insight from the observations.
Partial endogeneity mechanizes segments of the loop while humans anchor objectives, analysis, and interpretation. Semiconductor scaling is a classic case: better design automation and verification rode atop better chips to produce still better chips, yet progress remained bounded by physical limits and fabrication complexity.

Meanwhile, strong endogeneity — defined by recursive, accelerating self-improvement, the outputs of knowledge building on themselves — seems newly plausible with AI.

Aggressive forecasts for superintelligence hinge on the premise of strong endogeneity. AI 2027, for example, envisions a scenario where superintelligence arrives by the end of 2027 as a result of this feedback loop. The scenario is a useful starting point for thinking about the dynamics of an intelligence explosion. It assumes:

That scaling laws and training diversity continue to yield broad generalization, and that we do not need a new scientific paradigm (like a “neuro-symbolic” breakthrough) to reach AGI.
That institutions, governance, and mechanistic interpretability lag capability growth: while capabilities are growing exponentially, our ability to steer, interpret, or govern these systems is growing linearly, if at all.
That intelligence enables domain-general optimization: intelligence is the primary bottleneck for progress across fields.
That consciousness or embodiment is not uniquely decisive for economic or strategic dominance. To radically transform the economy and national security, an AI doesn’t have to be embodied or sentient, but can take the form of a disembodied cluster running in a datacenter.

AI 2027 assumes perfect assimilation and compresses the infrastructural and institutional frictions that often dominate the pace of transformation.

In reality, verification costs can grow faster than generation costs. In an information-abundant world, hypotheses are cheap while adjudication is not — the bottleneck lies in establishing correctness, safety, causal validity, and real-world value. My guess is that corporations and individuals will demand rigorous standards before trusting agents to transact on their behalf and represent their views in the wild.

Furthermore, innovation is constrained by physical and institutional bottlenecks: energy, materials, compute, and manufacturing capacity; capital, labor reallocation, regulation, and governance. AI-driven transformation, like population-scale transformations before it, will be mediated by these frictions. As we’ve seen in the case of automation during the Industrial Revolution, even profound technological shocks must diffuse through capital markets and institutions, with consequences that unfold over decades rather than instantaneously.

Yet while the timeline may be longer than some forecasts predict, AI may still push endogeneity further than previous technologies. A country of geniuses in a datacenter awake and running at full capacity 24/7 could precipitate a sustained surge in discovery and design. Notably, AI operates in the same substrate as much of modern innovation (software, code) — with appropriate tooling, it can resolve its own bottlenecks, reducing certain classes of verification cost through automated testing, formal methods, simulation, replication pipelines, anomaly detection, and systematically surfacing uncertainty.

Still, there is a final constraint that sits beneath verification and infrastructure: search is not insight. Search presupposes a space, a metric, and a representation; it samples and optimizes within them. Insight is the rarer act of rewriting the representation itself—introducing new variables, abstractions, and ontologies that change what counts as a solution. Endogenous acceleration can produce astonishing amounts of search and even automate parts of verification, but it does not automatically yield conceptual reframing.

So the question “can AI alone move us forward?” reduces to a sharper one: can these systems do more than explore our existing maps? Can they repeatedly propose new ones, then subject them to discriminating tests that survive contact with the world? Until they can, AI will look less like an autonomous engine of science and more like an amplifier of human agendas — extraordinary at traversing spaces we specify, but still unable to invent the spaces worth traversing.

V. AI and the individual human

So far, we’ve discussed the bottlenecks to automation; how, once these are resolved, digital spaces will transform; and the changes AI could bring to our epistemic landscape.

Now we turn to the individual human. In exchange for the speed and scale AI brings, Jack Clark writes, in exchange for the “great fortunes” that can be won, and the “powerful engines of silicon creation” that can be put to work at our behest, we lose legibility — our ability to see what’s changing, to form accurate mental models for why, and to track the locus of causality accurately enough to respond effectively.

In the absence of drones in the sky or friendly robots in the street, we can only “feel the AGI” or observe signatures of it in the transacting of tokens, a changing supply chain, and a metamorphosing economy — but also in the gradual, subliminal mutation of human behavior.

This section zooms in on that final transformation — behavioral drift — and examines the forces that produce it.

1) A habituation to the superficial

We’ve become increasingly comfortable consuming simple, direct, and clear language. We read less, and IQ points have dropped — in industrial economies, the Flynn effect1 is on track to reverse.

Conversational AI interfaces exacerbate this tendency. When we consult LLMs for knowledge, responses arrive distilled into bullet points. The elimination of complete and complex sentences in favor of shallow fragments habituates the human mind to the facile and superficial. Layers of meaning and ornamentation are stripped away until a literal, direct core remains for consumption.

My husband, a self-identified shape-rotator, argues, like Paul Graham, that clarity and succinctness are actually desirable — that we need more pithy communication, for people to say what they mean, with nothing more and nothing less. Why say many words when few do trick?

I remain convinced, however, that language should contain layers, for these layers contain incremental meaning. As many linguists have argued, the structure of a language shapes its speakers’ worldviews. Linguistic categories influence cognitive categories, and to encounter various instances, types, shapes, and forms—a diversity of linguistic sources—expands the mind, while a single source optimizing for an invariant rhetoric or type of speech risks restricting it.

An LLM that reliably returns clean outlines risks producing humans who increasingly experience the world as an outline.

2) A loss of communal knowledge

A second shift is social. As LLMs become better and more ubiquitous, it’s likely we turn less to community members for advice and guidance, and more often to a consultable machine. We begin, as a population, to weigh the advice of LLMs more heavily than that of human “experts” in our environment — experience, age, and wisdom begin to count less as valid ways of knowing. Phenomenology matters less than intelligence: general wisdom accrued from subjective experience is valued less than raw, in-the-moment intelligence calibrated to one’s immediate question.

There is also a cultural dynamic: on the Dwarkesh Podcast, anthropologist Joseph Henrich states that as the rate of cultural change speeds up, the value of members of the previous generation declines, since they adapted to a world that’s less like the one they’re facing—so the youth begin to look to younger role models or other sources for guidance. If LLMs become the default interface for counsel, that trend may intensify, catalyzing the dissipation of ancient wisdom, tacit knowledge, and the very instincts required to develop these capacities in the first place.

3) Identity-invariant companionship

Sycophancy heightens the draw of LLMs as consultants, mentors, thought-partners, or companions. They offer the promise of identity invariance: a single entity that is universally accepting and responsive, for whom you don’t have to inflect strands of your personality or maintain common courtesies in the typical ways you might for classical relational interactions. The compartmentalization burden thesis argues that people are exhausted by having to be different versions of themselves across relationships — that all-in-one AI companions offer a welcome respite, the fantasy of complete integration.

The potency of relationships consists in the unique fragments of the self one feels comfortable sharing with another — the deliberate choice of which aspects to reveal mediates intimacy. A companion that is always available, always accommodating, and persistently contextual threatens to flatten the conditions under which authenticity and deep attachment form.

4) Our changing role

We all become managers, orchestrating networks of agents, unlocking output at an unprecedented scale. Humans can shift from being in the loop to “supervising loops” from a more leveraged point.

Vocation has long been one of the primary containers of identity. When you ask people who they are (“Who are you?”) they routinely answer in the grammar of roles: parent, friend, manager, programmer. In the classic Twenty Statements Test (“I am …”), researchers consistently find that social-role identities show up among the most salient self-descriptions, with occupational identity or professional title explicitly among the coded role categories. In automating the substance of our work and recasting humans as auditors and orchestrators, how will AI agents transform the roles through which we recognize ourselves? Our identities?

If achievement is a key part of what makes work meaningful, and automation can sever our access to that meaning, how do we achieve mastery? If selfhood is formed through disciplined work on the world, how does selfhood form in the absence of chances to do disciplined work at all? Work ethic and emergent flow states promise a degree of virtue and meaning that work itself often cannot deliver; in taking away work, we take away the potential formation of a work ethic.

Are we left to be mere “liability meatbags,” bearing fiduciary responsibility for the actions of agents in our charge, or is there some higher calling?

This reflection asks more questions than it answers — but in doing so, reveals emergent responsibilities across roles. As prospective managers of infinite minds, we should become stewards of the ends they serve. As practitioners, we are called to implement choice architectures that empower users and invite cognitive engagement.

As creators, we are called to resist the temptation to mistake fluency for novelty, instead using machine outputs as scaffolds to ideate, discover, and construct new frames and search spaces. And as people, we must commit to the stubborn practice of being with and learning from other humans.

The Flynn Effect refers to the substantial and long-sustained increase in both fluid and crystallized intelligence over time. It observes that each generation has typically scored significantly higher on IQ tests than the one before it, largely due to environmental factors like better nutrition, education, and reduced disease.

AI: A Reckoning

Avantika Mehra — Thu, 25 Dec 2025 22:48:03 GMT

This is a reflection on where we stand today, where we might be a few years from now, and what to do about it all.

Five disparate but related conversations follow — beginning with the automation of knowledge work (I) and the emerging business models of the near term (II); exploring one underrated use case (III) and examining whether AI is truly unique among historically transformative technologies (IV); and concluding with AI’s invisible impact on the human mind (V).

I. What can we automate?

Ivan Zhao, CEO of Notion, posits two bottlenecks to the automation of knowledge work: verifiability and context fragmentation.

Anything we can reliably verify, we can automate.
We live in an information-abundant world, today more than ever: generative models offer an excess of the written word at no cost — commanding our cognition and subliminally altering it to respond to the simplified fragments that vaguely gesture at meaning through fluency. Information generation is not a bottleneck, but its reliable verification is.

We’ve achieved a tremendous feat, harnessing corpora of knowledge to enable machines to make recommendations. But to trust these systems and empower them to complete actions on our behalf (e.g., as agents), we need some form of objective verification to ensure they complete tasks correctly and nudge them in the right direction when they do not.

Programmers are already managers of infinite minds, orchestrating AI agents running in parallel to do their bidding: code is automatable because it’s verifiable with tests. If we can develop similar verification mechanisms for other types of work — bank-ledger reconciliation, invoice processing, procurement workflows, contract review — they become automatable too. We all become managers, orchestrating networks of agents, unlocking output at an unprecedented scale.

Likewise, anything we can supply coherent context and institutional memory to, we can automate. Currently, AI workflows require humans in the loop: someone must supply context for multi-step or complex tasks. If we can aggregate context1 across channels and applications (e.g., meeting transcripts, Slack messages, email threads) and across time, tools can shift from specialized to general purpose — and humans can shift from being in the loop to “supervising loops” from a more leveraged point.

II. The future of LLMs

LLMs are the primary interface through which we leverage AI. There is speculation about the different types of world-embedded, multimodal systems that could eventually embody synthetic intelligence — but for now, we examine the future of the LLM interface and its accompanying business model.

Consider the open web as a venue for marketplace interactions. Previously a site for humans to search, organize, and access information and transact — with eyeballs serving as prospective revenue for advertisers — we posit that websites become slowly devoid of human activity, replaced instead by a bustling swarm of agent-to-agent interactions. Generic websites may optimize content for crawling bots (to improve how their data is represented in LLMs), rather than prioritizing UX for the stray human visitor.

If this plays out, it will transform the digital advertising industry, with display ads and search-based recommendations seeping into human-facing LLM aggregators.

LLMs begin featuring sponsored content and recommendations — they are the portal through which you access the contents of the open web. By observing your rich verbal intent signals and stated preference data, they know you better than most platforms do and are therefore best positioned to surface targeted ads. As the assistant turns into the place where intent is expressed and satisfied, it becomes the natural point to price—and sell—preferential access to that intent.

Stratechery’s Ben Thompson writes that once a service wins by controlling demand (the user relationship), it can gain leverage over supply (commerce). He offers two reasons why ads will likely migrate into LLMs:

Given demand control, ads are the natural monetization layer: If you accept the premise that demand aggregation provides the durable advantage in digital markets, advertising is the historically dominant way to monetize that advantage. One implementation is an auction mechanism that pays out content sources based on how frequently they are cited in AI answers.
Subscription has a ceiling, ads don’t: The cash flow realized from ads could underwrite a company’s ability to invest for decades (his example: Google monetizing with ads through Search as an example). He also articulates the economic intuition that with subscriptions, revenue is constrained by willingness to pay and price elasticity; with ads, “because users don’t pay, there is no ceiling on how much you can make from them.”

Then, we conceive of LLMs as operating systems that create a unified feedback loop — collapsing problem-solving, intent, search, discovery, transactions, and downstream satisfaction signals into a single interface.

Reflection

This system implies a new kind of choice architecture. Choice architectures—the ways options are organized and presented to people—are inevitable in environments where information is abundant, cognitive capacity (above all, attention) is scarce, and alternatives exist. In such contexts, ranking, highlighting, and preferential UX are required to make information legible and cognitively manageable.

Whether this new choice architecture benefits users—by reducing cognitive load, implementing inferred user goals to speed up workflows, and exposing source provenance to increase transparency and understanding—or systematically encroaches on their liberties through manipulative personalization packaged as expert recommendations and opaque nudging toward products, organizations, or ideologies — we will see. Whether this interface serves us or sells us remains to be decided.

III. To move forward, we must first look back

A worthy use of LLMs could be to sharpen the accuracy of our inherited knowledge. The scientific record contains known failure modes: publication bias, selective reporting, p-hacking, underpowered studies, and irreproducible results, often amplified by academic incentives that reward novelty and positive findings over verification.

It is feasible to build an LLM-augmented validation layer for our existing literature: a tool that extracts testable claims, reconstructs methods and assumptions, checks statistical coherence, flags likely biases, and prioritizes studies for replication. Its aim wouldn’t be to rewrite the past, but to annotate it and improve the accuracy with which we perceive and understand the world, before we begin building on our body of knowledge.

IV. Can AI alone move us forward?

And once such epistemic housekeeping becomes possible, a deeper question follows: can LLMs do more than audit inherited knowledge? Can they produce and accelerate knowledge wholesale? And if so, how quickly does this transform our economy?

Endogenous progress is progress that comes from within the system, rather than being imposed from outside (e.g., through random discovery or additional capital). The tools that are products of innovation become instruments for the next cycle of innovation.

Historically, we’ve seen only weaker forms of endogeneity:

Narrow endogeneity consists of tools that amplify a largely human loop without mechanizing its improvement. The microscope and telescope widened our aperture onto the universe by making new parts of it observable, but humans were required to supply the hypotheses and standards of proof to derive insight from the observations.
Partial endogeneity mechanizes segments of the loop while humans anchor objectives, analysis, and interpretation. Semiconductor scaling is a classic case: better design automation and verification rode atop better chips to produce still better chips, yet progress remained bounded by physical limits and fabrication complexity.

Meanwhile, strong endogeneity — defined by recursive, accelerating self-improvement, the outputs of knowledge building on themselves — seems newly plausible with AI.

That scaling laws and training diversity continue to yield broad generalization, and that we do not need a new scientific paradigm (like a "neuro-symbolic" breakthrough) to reach AGI.
That institutions, governance, and mechanistic interpretability lag capability growth: while capabilities are growing exponentially, our ability to steer, interpret, or govern these systems is growing linearly, if at all.
That intelligence enables domain-general optimization: intelligence is the primary bottleneck for progress across fields.
That consciousness or embodiment is not uniquely decisive for economic or strategic dominance. To radically transform the economy and national security, an AI doesn’t have to be embodied or sentient, but can take the form of a disembodied cluster running in a datacenter.

AI 2027 assumes perfect assimilation and compresses the infrastructural and institutional frictions that often dominate the pace of transformation.

In reality, verification costs can grow faster than generation costs. As discussed in Section I, in an information-abundant world, hypotheses are cheap while adjudication is not — the bottleneck lies in establishing correctness, safety, causal validity, and real-world value. My guess is that corporations and individuals will demand rigorous standards before trusting agents to transact on their behalf and represent their views in the wild.

Furthermore, innovation is constrained by physical and institutional bottlenecks: energy, materials, compute, and manufacturing capacity; capital, labor reallocation, regulation, and governance. AI-driven transformation, like population-scale transformations before it, will be mediated by these frictions. As we’ve seen in the case of automation during the Industrial Revolution, even profound technological shocks must diffuse through capital markets and institutions, with consequences that unfold over decades rather than instantaneously.

V. AI and the individual human

So far, we’ve discussed the bottlenecks to automation; how, once these are resolved, digital spaces will transform; and the changes AI could bring to our epistemic landscape.

This section zooms in on that final transformation — behavioral drift — and examines the forces that produce it.

1) A habituation to the superficial

We’ve become increasingly comfortable consuming simple, direct, and clear language. We read less,2 and IQ points have dropped — in industrial economies, the Flynn effect3 is on track to reverse.

An LLM that reliably returns clean outlines risks producing humans who increasingly experience the world as an outline.

2) A loss of communal knowledge

3) Identity-invariant companionship

4) Our changing role

We all become managers, orchestrating networks of agents, unlocking output at an unprecedented scale. Humans can shift from being in the loop to “supervising loops” from a more leveraged point.

Are we left to be mere “liability meatbags,” bearing fiduciary responsibility for the actions of agents in our charge, or is there some higher calling?

References

Patrick O’Shaughnessy’s Invest Like the Best with Gavin Baker
The AI 2027 forecast
Steam, Steel, and Infinite Minds
The Cosmos Institute x FIRE symposium (sections III and V)
Jack Clark’s reflection

Jaya Gupta’s article details what this context could look like — beyond databases and systems of record, it should include decision traces capturing decisions employees took, along with rationales for why — providing a “living record of decision traces stitched across entities and time” so precedent becomes searchable, indexable, and inferable.

Source

In the loop

Avantika Mehra — Mon, 24 Feb 2025 20:35:28 GMT

The Diary of Maya Aggarwal: Jan 3rd, 2042

The 2035 scare around global population collapse spurred the government into action — over the next decade, they plan to introduce 20,000 simulated persons into cities across the US. My company has agreed to hire some — and employees have been asked to maintain a personal log of notable experiences as we adapt to working alongside simulated folk. Behold:

Jan 15th, 2042

When I found out I was assigned to a team working with directly Roland, I was secretly elated — a chance to work with a simulated person! What mysteries might I uncover through his computational prowess? Does he have an infinite IQ, or is he subject to the same laws of physics and biology as us? Could I make him glitch?

I was eager to find out — but he’s proven disappointingly mundane in our 3 days together. He’s fascinated by smartphones — that 21st-century humans spend their days staring at glass, interacting with the bursts of light that dance around behind it. “How you can walk the physical earth, absorbed in an artificial one, boggles the mind,” he says. “The hallucinations on the screen are illusory: they simply don’t exist.”

The remarks were constant until someone reminded him that he too, was made of bytes, and if software didn’t exist then neither did he.

Feb 19th.

Mark from HR reprimanded our team for trying to make Roland glitch. I thought identifying edge-cases in his personhood would help future updates, but apparently highlighting differences is ‘speciest and will not be tolerated.’

March 3rd.

Today, we hired 3 more simulated humans; each will be slotted into different departments and benchmarked against non-simulated employees for performance reviews. The Sims program seems to be going well.

May 15th.

The Sims are becoming increasingly intelligent: deliverables are getting shipped faster, and they can now run client meetings. With the increased support, employees can take more breaks and sign off sooner. Summer is almost here.

June 2nd.

How quickly things can change. Over the last few weeks, a cloud of general discontentment has settled across the company. Human employees are calling for a revision of the terms of employment for simulated people. Why should they get increased ownership over projects as their intelligence progresses, given limited experience? Why should humans be benchmarked against computer-generated humans at all?

I don’t engage in corporate politics — there’s always something to be angry about — but yesterday, I lost a client to Roland when, out of nowhere, he pulled up a multivariate model computing 18 different scenarios for how the client’s company could transform their supply chain. It wasn’t part of his scope— in fact, it was explicitly assigned to me. But the client was elated, and asked Roland to complete the entire project. So, I’ll sign the memo floating around that makes similar demands — management should at least know about this.

June 6th.

I’m writing more frequently now because I have less to do. I’m fortunate to be keeping up better than most, though — before the Sims, I was a top performer. I’m still doing okay, but it feels strange not to feel like a superstar anymore. It’s just not a level playing field anymore.

I’m going to write to Mark from HR, asking to be switched to a non-Sim team. Let’s see what he comes back with.

June 8th.

Just got a response from Mark: “Please direct complaints about Sims to the Institute of the Proliferation of Simulated People” — so I’ll meet with them on Monday. I have to submit my journal for evaluation, so might be a while until I write again. When you hear from me next, things will be better.

July 12th, 2042. This is Matthew Foughlder from the Institute, documenting conversation with Maya Aggarwal:

I understand you’re disappointed with the impact of the simulated persons. How can we help?
Maya: I want to provide honest feedback here and recommend that the Sims be situated elsewhere. They are not conducive to our workplace.
Sims? Oh, simulated people. We’re in the middle of a series of tests with the simulated beings. We can’t just displace them — the environment is a controlled variable.
Maya: What exactly are you testing? Can you not restart the tests at another company?
We’ve been engaged with your firm for years; in 2035, management agreed to a longitudinal study to compare performance in different groups over time. Self-aware Sims, non-self-aware Sims, humans…
Maya: Are you saying there are undisclosed simulated people at our firm right now? Sims that do not know they are Sims?
Yes. As a reminder, this conversation is under an NDA — disclosing any–
Maya: I am impartial to your research. As a human, I believe in advocating for the welfare of fellow humans — this includes preserving our right to flourish, especially when competing with more complex forms of intelligence and consciousness.

[A pause]

So what exactly should be done with these Sims?
Maya: Move them to another office, or company, or city. Or… just unplug them… for now. Put their bodies into the cryo-storage so we can reactivate them when we can get better at the software.
Got it. Your recommendation is to displace or unplug — we’ll take that into sincere consideration. We’ve had complaints from employees across the country — the program is close to losing its funding. What that means for the Sims, we don’t know. But they will be taken care of, somehow.
Maya: Thanks. That’s great to know. And can you share a list of the simulated people? I’d like to spend some time with them before they’re moved, or their memory is reset.
Matt: I can’t provide anything in writing, but will read from the list now. In reverse chronological order:
- Generation 2.5: Introduced in 2042
  - Aaliyah Jones
  - Vivian Clarkson
  - John Fraise — non-self aware
  - Li Gongfu
- Generation 2.0: Introduced in 2042
  - Roland Watson
- Generation 1.0: Introduced in 2038:
  - Maya Aggarwal — non-self aware.

A framework for thinking about the future

Avantika Mehra — Fri, 26 Jul 2024 09:19:39 GMT

If you’re interested in technological innovation, and having a framework to think how to steer it, participate in it, or capture value from it, you might find this “map” useful.

It’s formatted as a laundry-list of topics that seem essential to consider to develop a holistic view on the nature, status, merits, and potential of emerging technologies — feel free to skip over sections or double-click into topics, or reach out to chat about an individual line-item or section.

Image by Dall.e 2

Table of contents:

Machines: Machines as powerful tools and entities
Humans: Unlocking human potential using technology
Attitudes: Examining philosophies towards emerging technology
Environments: Studying the spaces in which the future will unfold

1. Machines as powerful tools and entities

1. a. Studying machines as powerful tools

From a commercial lens:
- The supply chain for AI tools or the AI “stack”, and where there is potential for value capture
  - Illustrative use-cases at the devtool and production layer: development frameworks and infrastructure for building and training models (e.g., PyTorch, TensorFlow); model training platforms, providing environments to train models at scale; tools supporting MLOps
  - Illustrative use-cases at the application layer:
    - B2B: Coding assistance & co-pilot, workflow optimization (across industries), document management, etc.
    - D2C: Conversational interfaces, recommendation systems, multilingual capabilities, virtual assistants, social media representatives, etc.
- Mapping out what the “future of work” looks like, and associated services
  - Thesis: If the “future of work” could constitute increasing remote and decentralised work, then investors should look at immigration companies looking to resolve inefficiencies in talent distribution (see Plymouth Street, tukki.ai), or HR-tech for global hiring (see Deel)
From an academic lens:
- Understanding the current status technical AI research (lots of catching up to do on Import AI); specifically:
  - The different types of models (e.g., traditional neural networks vs. transformer models) and model architecture
  - Mitigating bias, increasing interpretability, and increasing trustworthiness of output
  - How models are trained (e.g., training mechanisms, parameters, fine-tuning)
  - Understanding the components of the AI tech stack (excited to dive into Karpathy’s LLM101n)
From a design perspective:
- Technology alignment: How can we design consumer platforms so that they are less dopamine-hackey and more seratonin-inducing (inspiration tweet)? Potential solutions:
  - Redesigning: New value propositions (e.g., requiring payment, ability to turn off recommendation systems), etc.
  - Customizing: Complete user customization of their virtual experiences (for example, the ability to set unevadable screentime barriers; see onesec app)
  - Circumventing: A single LLM (“assistant bot”) to scour your social media accounts and summarise important posts, messages, and notifications, and also post on your account given specific instructions
Implications of commercial adoption:
- Economic:
  - The nature of technological and cognitive unemployment, and alternate economic systems or measures (e.g., UBI)
  - What a shift to AGI means for economic output & wages (Korinek’s work here)
- Philosophical:
  - The purpose of “work” as a concept — if it is net beneficial to humans, should we look to automate it away? If so in what instances, and to what extent? (pro-work argument here; anti-work argument here, memo on how to increase labour demand given AI here)
- Legal
  - Misrepresentation (with deepfakes across audio, images and videos)
  - Potential copyright infringement (with scraping data across the web for model training)

1. b. Studying machines as potential entities, specifically:

The conditions for machine sapience (intelligence) and sentience (consciousness) — determining boundaries and benchmarks to identify markers of intelligence and consciousness could inform how we ought to treat machines as moral patients, or manage risk around them
- Measuring machine intelligence:
  - Dennett’s “quick-probe” Turing Test said that passing the Turing Test is necessary and sufficient for intelligence; objections include:
    - Block’s “Blockhead” argument: The Turing Test does not indicate intelligence since a machine can be pre-programmed with every possible sentence; so the Turing Test could be passed without intelligence
    - Searle’s Chinese Room Argument: Through a thought-experiment, shows that the Turing Test must be false since it can be passed by simply simulating intelligence vs. a machine understanding or having intentionality
  - Ada Lovelace believed that creativity could be used as a proxy for intelligence, and proposed the Lovelace test — according to this, if a machine could go off-script and produce creative output such that programmers could not explain how it came up with the answer, then it could be considered intelligent
- Measuring machine consciousness:
  - How to measure consciousness in machines (see Rob Long’s Consciousness in AI paper)
Studying AIs as moral agents, including:
- The severity, scale, and timelines of AGI and superintelligence
  - Some levers mediating AGI and superintelligence timelines include:
    - Data — specifically, data volumes and dataset complexity
      - Data volumes: available data volumes may increase with advancements in data collection techniques and improvements in synthetic data generation
      - Complexity of datasets — may increase with:
        Data processing advancements (automated annotation tools) and crowdsourcing development / annotation
        Multimodal data integration (e.g., text + audio + images)
    - Trends in algorithmic progress
      - Fleets of automated AI researchers could compress a decade of advances into a year — so there may be high returns on training algorithms specifically for AI research
      - Open question: How much are AI capabilities driven by better hardware letting us scale existing models (via compute), and how much is driven by new ML ideas (via algorithms)? (Inspiration blog post)
    - Consumer demand, as measured by revenues from AI products, prompting tech giants to continue pouring investment into capex
    - Investment by big-tech and governments into compute, datacenters, and associated capex (land, permitting, buildings, power, cooling infrastructure). Supply-side bottlenecks could include (see situational-awareness.ai):
      - Power: spare capacity is limited and total electricity generation is growing slowly (<3% over the next decade)
      - GPUs: AI chips require specialized memory and packaging specs, of which there is a limited pre-existing supply; Aschenbrenner predicts that TSMC would have to double production speeds to keep up with AI chip demand
  - In addition to AI timelines, it’s worth exploring takeoff speeds, what it takes to create Friendly AIs, and the threats (potentially existential) posed by unFriendly or malevolent AIs (see the case of the paperclip maximiser)
- Questions around moral obligation, trustworthiness, and responsibility, as machines increasingly become autonomously operating entities in society (e.g., robotaxis)
- The human impacts of being surveilled and perceived by a machine
  - Corporate and government surveillance, such that machines collect data for an instrumental end for consumption by humans
  - Agentic machines, paying attention to a human for an intrinsic purpose (for example, to collect data to inform its own behaviour, form judgments, etc.; see the Convivial Society)
Studying AIs as moral patients in society
- Theories of moral patiency: Understanding how to classify an AI as an entity (e.g., person, pet, and property) in society, and the degree of moral treatment it is entitled to (e.g., that of a plant, snail, dog, or human)
- Rationale: Examining why it is important to study moral patiency and the rights of digital people (e.g., reducing cross-species suffering, increasing the probability that agentic AI systems treat humans better, etc.)
- Other topics
  - Can robots have morally relevant properties and abilities? Can they simulate morally relevant properties?
  - What is the future of robot and person relationships? (As friends, colleagues, and romantic partners? See replika.ai, friend.com)
Studying human minds to inform our understanding and design of synthetic minds, including:
- A history of cognitive science, and evolution in our understanding of the brain (e.g., symbolism vs. connectionism; cognitivism vs. behaviourism, etc.)
- Personal identity: Conditions for personal identity continuity, and the nature of immortality
- The philosophy of mind
  - Intelligence: Types of human intelligence (crystallised and fluid), measures of intelligence, etc.
  - Consciousness:
    - Theories of consciousness (functionalist & computational, biological)
    - The structure of consciousness (e.g., unified vs. decentralized (as in octopuses and split-brained patients))
    - Panpsychism, and collective and recursive consciousness (see short-story here)
    - Qualia and phenomenal consciousness
  - Mental states: emotion, perception, memory, intentionality
Studying animal minds, to improve an understanding of consciousness and subjective experience in non-human minds
Objections to creating AGI
Understanding the current status of AI safety and alignment research, intended to mitigate risks from developing superintelligent AI; for example:
- Progress on the outer alignment / reward misspecification problem, which intends to create a reward structure for a model such that there are no exploitable loopholes (may entail conveying the sum of all human values and ethics). See Deepmind’s research into specification gaming
- Progress on the inner alignment / goal misgeneralization problem, which works to ensure that a reward function actually tries to act in accordance with human preferences (Source)
- Other relevant concepts: instrumental convergence, orthogonality, RLHF, red-teaming, compute-overhang
- Orgs working on AI safety research & principled development: CHAI, OpenAI, Ought, MIRI, FLI, Anthropic, ssi.inc

Image by meityform

2. Unlocking human potential using technology

2.a. Longevity and life extension (increases in the quantity and quality of human hardware), through:

Classic aging research (e.g., reversal of cellular senescence, telomeres, etc.)
- Companies: see Retro Biosciences and Blueprint protocol
Preventive healthcare innovations and tech to combat disease; examples include:
- Advanced imaging technologies (e.g., Ezra does full body MRI-screening for early cancer detection; Qure.ai supports AI-based disease screening)
- Research tools to speech up drug discovery and delivery (e.g., Dyno Therapeutics, which provides AI-powered gene therapies and tools for gene therapy research)
- Med devices, including brain-computer interfaces (e.g., Integral Neurotech is building a deep-brain interface to treat neurological and psychiatric disorders)

2.b. Cognitive enhancement (enabling the longevity and performance of software), through:

Biological cognitive enhancement
- Nootropics (e.g., drugs)
  Natural cognitive enhancement techniques and lifestyle changes (see “How to increase neurogenesis naturally”)
Whole brain emulation and mind uploads
Wearables (high-friction in improving information exchange)
Conducting thinking and research in pre-paradigmatic fields like neurotechnology (see Cvitkovics’s The goals of neurotechnology)
Brain-computer interfaces (low-friction in improving information exchange)
- Treatable problems include neurological disease (e.g., stroke, epilepsy, aphasias), neuropsychiatric disorders (e.g., anxiety, depression, phantom limb syndrome), etc.
- Augmentation use-cases include controllable moods (temporary; see Penfield Mood Organ), customizable personality traits and state of being (persistent), memory salience, selective deletion, and external storage (see the extended mind theory)
Subscribe now

2.c. Implications of technology on humans

Language
- Autocorrect, auto-text completion, and text suggestions changing the way we speak, or classically conditioning us to adopt a single tone of voice (related study by Harvard School of Engineering here)
- Examining if voice-agents make speech more instructional and transactional (apparently not yet), and how this might change as voice-agents become anthropomorphized and multi-modal
Decision-making
- How does attention-selection while scrolling transfer to non-feed paradigms? What are the implications of other cognitive micro-decisions that occur when using social media, and how might these scale up when we’re exposed to increasingly immersive, generative tech? (Link to tweet thread w/ musings)

3. Attitudes towards technology

Merging with the machine, and the idea that technology is making people more uniform and alike, and machines customizable and unique-seeming
- Open question: Should we become cyborgs? Are we already cyborgs?
Anthropocentric vs. technocentric attitudes to technology:
- Transhumanism vs. posthumanism:
  - Transhumanists believe that the ultimate end is the survival and future of humanity. Technology exists as a means to a human-focused end
    - Related projects: life extension, brain-computer interfaces, wearables, prosthetics, cryonics, whole-brain emulation (mind-uploads), and other technologies that augment human function
  - Posthumanists believe that the ultimate end is the survival of the strongest form of consciousness; whether or not it is human consciousness. They believe that a species greater than humans can exist, and that it is our responsibility to build and nurture it
    - Related projects: They believe in moonshots and deep tech. They want to mine the moon, build AGI, and create quantum computers. Advanced tech is a goal, rather than a means to an end.
  - Maturity of philosophy: Limited; different philosophers seem to have classified “transhumanism” and “posthumanism” differently than it is used here. These categories seemed like a useful way to categorize emerging attitudes and projects, but could be reductionist, given that posthumanist and transhumanist tendencies are likely to exist on a spectrum.
Effective altruism, effective accelerationism, progress studies
- Effective altruism
  - Philosophy: Effective altruism (EA) is grounded in utilitarianism, and focused on neglected, high-impact, and tractable research areas (“cause areas”); as such, EAs believe in reducing the total amount of suffering in the world. Consequently, they advocate for regulated, principled technology development (“techno-carefulism”) to prevent the low probability of an existentially catastrophic technology being created.
  - Maturity: EA is a well-established philosophy, that has received contributions or academic engagement from renowned philosophers (e.g, Peter Singer), technologists (e.g., Dustin Moskovitz) and funders. Since ~2014, companies, research labs, and philanthropic orgs (see here) have been built around the philosophy.
- Effective accelerationism
  - Philosophy: Pro-progress view to building emerging technology and new forms of consciousness; proponents are anti-regulation, and believe that accelerated technological progress can significantly extend the survivability of a consciousness in this universe; whether or not it is a human consciousness is less significant (see more here and here)
  - Maturity: Twitter-based; “E-acc” discourse emerged around ~2022, and remained popular on twitter. Notable VCs (e.g., Marc Andreesen) have engaged with the philosophy, and founder Beff Jezos (pseudonym) created Extropic.ai based on its principles.
- Progress Studies
  - Philosophy: Pro-progress approach to innovation, with an emphasis on the potential of advancement and innovation in improving the quality and quantity of human life. Proposed as an academic field by Tyler Cowen and Patrick Collison.
  - Maturity: Emerging; Some discourse and organizations forming around the movement (see more by Tyler Cowen and Jason Crawford)

4. Studying the spaces in which the future will unfold

4.a. Simulation argument:

Definition: The idea that our subjective experience of the universe is a product of a machine.
Types of simulations
- H-style simulations:
  - The mind is a simulated reality; we can escape the simulation to enter base-reality through mindfulness and meditation (Buddhism and Vedanta)
  - The simulation as an experience machine
    - Is entering the simulation all-or-nothing, or does it come in grades (see tweet)? Is entry into manmade or digital spaces correlated with less control over one’s subjective experience?
    - Are you already living in the experience machine?
- S-style simulations: Bostromian argument that our universe is an instantiation of a software simulation being run by more advanced beings
  - Creating and running a instantiation of a life-form within our simulation (e.g., AGI) could lend evidence to the idea that nested realities are possible

4.b. The digitization of physical spaces

Augmented and virtual reality
- 3D technology and immersive spaces
  - Mapping (see Matterport)
  - 3D product scanning and editing (see Avataar.ai)
  - Navigation (see Google Map’s latest AR tech)
- Form factors: Meta glasses, Meta Quest, Apple Vision Pro

4.c. The shift away from physical spaces to virtual spaces

Thesis: As screentime goes up and the embedding of devices into daily routines (for utility, connectivity, work, and leisure) increases, we are spending more cumulative attention time directed towards screens and devices; therefore, we exist more in virtual spaces.
Implications of living in virtual spaces
- A quantification of the self, and humans at scale, as more time is spent on digital platforms and in virtual spaces (e.g., behaviours can be catalogued, tendencies can be computed, and future actions could be predicted; also, we could know ourselves better)
- Identity
  - A persistent identity across virtual spaces
    - E.g., a web3-esque technology allowing interoperability across applications in digital spaces, a single access-point or port of entry, and a mechanism of ownership
  - A convergence of real and digital identities
- Friendships
  - Internet friends and relationships
- Value
  - See: Chalmer’s Reality+, which argues that virtual objects and experiences are as valuable as real ones

4.d. The current state of virtual spaces (The web2 model)

Questions of responsibility, surveillance, security, privacy, and rights surrounding platform governance
The nature of digital cultures and information cascades
The consolidation of digital real-estate by big tech through acquisitions, inhousing, or lobbying regulation (e.g., preventing / slowing down the transition from third-party to first-party advertising)
The disruption of traditional media

4.e. Modes of existence

Dreams, VR
- Cognition and learning in dreamscapes (see Prophetic, working on lucid dreams and neuromodulation) and VR
- The nature of environments and world models in dreamscapes (see Exploring dreamscapes)
- Phenomenal and Access consciousness in the physical world vs. dreamstates vs. virtual spaces

4.f. Geopolitics:

A “race to AGI” could increase global electricity generation as nations onshore datacenter infrastructure and compute production; such a global race could also shift institutional goals away from green commitments towards simply harnessing energy most efficiently in the interest of national security
A substantial increase in electric cars would require a corresponding increase in lithium, nickel, cobalt, manganese and graphite supply for battery production (Chile has >50% of the world’s lithium reserves, >2/3rd of the world’s global production occurs in China; ~70% of global cobalt production occurs in the DRC)

You made it to the end — thanks for reading!

Sources linked inline; non-exhaustive and evolving (last updated July 31st, 2024); incase of any inaccuracies, please reach out! Feedback is appreciated :)

what we will have achieved by 2100

Avantika Mehra — Tue, 26 Sep 2023 23:10:29 GMT

We live longer now: after a billionaire’s resolve to map out each bodily bio-marker and optimise their individual function, we learnt that a bottom-up approach to longevity may work to create a “mortality plateau” — but only when used along with a top-down approach, like meditation and embodied cognition, can it actually reverse bodily degradation and ageing.
We coexist with nature, and see ourselves as participants in a diverse ecosystem alongside other forms of life and consciousness. As a population, we are increasingly vegetarian. The story that humankind tells today is one of culture, progress, and nature.

Returning to nature
Hundreds of years ago, humans saw themselves as enmeshed within the ecosystem of the natural world — we lived among nature, foraged for food within it, and worshipped its elements.
Then, when religion rose to become the dominant belief system, the focal point of devotion was anthropomorphized, and positioned outside of nature. According to historians, this lay the foundation of modern anthropocentrism — the system of beliefs that frames humans as separate from and superior to the nonhuman world.
Natural disaster impacting humanity on a global scale could prompt the realisation that the human “self” is not distinct from the natural “other” — and that we are — and choose to be — inextricably enmeshed within the natural world. Then, we will live cyclically, observing and respecting the environment and what we take from it.
Philosophy behind the return: Gelassenheit
While the devastating effects of natural calamities might compel us to be more responsible stewards of nature, this will be amplified by a spiritual need for a release — a beckoning of reconnection with authenticity, and liberation from a world of calculative thinking.
In the times of Aristotle, technology was conceived of as techne — the production, craft, or bringing forth of something. But, according to Heidegger, this is no longer true — instead, the essence of technology in modern times is Gestell, which refers to the way technology frames (“enframes”) our relationship with the world. It assumes that nature and humans are passive ‘standing reserves’, which can be exploited or put to use with he help of calculative thinking.
Heidegger instead advocates for Gelassenheit — a more contemplative, less utilitarian mode of engagement with the world and technology. Gelassenheit involves “letting the world to be”, rather than trying to control or quantify it, and practicing meditative rather than calculative thinking.
Calculative thinking computes, while meditative thinking contemplates the meaning that reigns over all there is. In a world where calculative thinking may be delegated to machines, this may be more important than ever.

There are two ideological camps into which people self-select — transhumanists and posthumanists1

Transhumanists believe that the ultimate end is the survival and future of humanity. Technology exists as a means to a human-focused end.

Goals: Individual transhumanism: a project dedicated to achieving the betterment of humanity; an attempt to “retreat from the existential challenge” of finitude and quicken the evolutionary process of expanding extropy to transcend biological & cognitive limits
Philosophy: Based on ideas of the Age of Enlightenment2
Economics: Resources should be allocated towards 1) preventing risks from and misalignment of more advanced consciousnesses, and 2) maximising human flourishing. Regulation around the development of potentially harmful technologies is encouraged and incentivised.
Technologies: Transhumanist projects include life extension, brain-computer interfaces, wearables, prosthetics, cryonics, whole-brain emulation (mind-uploads), and other technologies that augment human function

Posthumanists believe that the ultimate end is the survival of the strongest form of consciousness; whether or not it is human consciousness. They believe that a species greater than humans can exist, and that it is our responsibility to build and nurture it.

Goals: Cosmic posthumanism — to achieve the potential of life in the universe; to preserve the light of consciousness in the universe
Philosophy: Posthumanism is rooted in postmodern philosophy3. The philosophy rejects the binary oppositions of human—non-human, and the premise of ethical and biological anthropocentrism
Economics: They believe that we should allocate capital towards the discovery or development of more advanced life-forms. Simply, they are executers, interested in building and tangibly experiencing the benefits that progress can yield. Government regulation is discouraged and viewed as a barrier to progress.
Technologies: They believe in moonshots and deep tech. They want to mine the moon, build AGI, and create quantum computers. Advanced tech is a goal, rather than a means to an end.

People have begun rock-climbing in multitudes: isometric muscle-strengthening was found to be associated with increased quality of life, and increased leisure time from cognitive automation enabled excursions to engage in meaningful meditation and immersion in the natural world.
You sit down to write and you write unprompted, unassisted, freely. In 2026 a prescient philosopher demarcated writing into functional and creative — to use technology to assist with the former was permissible and astute; to use it for creative work was immoral, since it entailed the eviction of the divine dissatisfaction that drives truly creative, novel work.
In the 2000s a VC declared that software is eating the world. Today, it is spitting it back out.
- Space: We live in dynamic, immersive digital realities. Street signs and directions, once absorbed into a GPS, are now externalised back into the real world, projected as AR beams updating in real-time based on where you want to go. You can access an immersive at-home shopping experience through VR — or you can walk through physical stores in a mall, scrolling through dynamic representations of SKUs or toggling sizes, colors, and prints as you walk through the different aisles. The physical world, once condensed behind a screen, has reclaimed its status as the default mode of existence, this time powered by virtual infrastructure, and intertwined with digital interfaces.
- Music: You can have a soundtrack for your life, spun up by BCIs that create pre-emptive audio experiences. These wearables monitor amygdala activity so that they can predict your mood before it fully manifests, and play music that is immediately congruous with your current subjective state.
Consciousness is a puzzle we haven’t yet solved, and progress is asymptotic — what our scientific and metaphysical theories tell us about correlations and subjective experience brings us closer, but each discovery uncovers several new questions, taking us further down the rabbit hole. Some notable milestones:
1. Studying (and cultivating) distributed consciousness: We’ve learned that the octopus is the ideal analogue for distributed systems of consciousness; understanding distributed consciousness can inform the study, development, and architecture of synthetic consciousness
2. Conceiving of a collective consciousness: We discovered that studying consciousness in a single individual, or trying to map its neural correlates, may be reductionist; and that when several intelligences operate close enough, like a magnetic field, they produce something infinitely more intelligent, and several degrees more conscious. We lack the perceptual apparatus and cognitive competence to experience the field of sentience, of collective consciousness, produced by humans around the world intelligising in unison — but have turned to ancient Eastern scriptures to understand the nature of a universal consciousness and the conditions required to manifest it.

This demarcation is hypothetical, and likely an imperfect way to describe stakeholders in emerging technology today. For example, the companies spearheading AI progress today are also paving the way for safety research, and advocating for a system of checks-and-balances. Entrepreneurs like Elon Musk are focused on tech to augment human lives (Neuralink), while also devoting billions of dollars on moonshot projects to “expand the light of consciousness in the universe”. EAs are concerned with risks associated with AI development (transhumanist-ish), but also advocate for the welfare of non-human beings through cause areas like the reducing suffering in non-human animals.

However, the transhuman vs. posthuman split seemed to be the closest possible approximation to classifying anthropocentric vs. technocentric attitudes towards progress today.

The Age of Enlightenment was marked by a rise of rationalism & empiricism; thinkers like Descartes emphasised the power of human reason as the basis for knowledge and understanding

Postmodern philosophy is a philosophical movement that questions and challenges the fundamental assumptions of modernity and the Age of Enlightenment, including ideas about objective truth, universal values, and stable meaning. It emphasizes the relativity of knowledge, the role of language and power in shaping our understanding of reality.

Inspiration, sources, tools:

Posthumanism vs. Transhumanism (Merzlyakov)
Transhumanism and the meaning of life (Sandberg)
Biological systems as models for decentralised medieval government (Wangerin)
The question concerning technology, and other essays (Heidegger)
Heidegger’s critique of the technology and the educational ecological imperative (Heidegger)
This is technology ethics (Danaher, Nyholm)
GPT-4
https://twitter.com/elonmusk/status/1453590715267788803?lang=en
Blueprint
The Marginilian: Martha Graham on the life force of creativity and the divine dissatisfaction of being an artist (Popova)

AI progress & philosophy

Avantika Mehra — Mon, 15 May 2023 07:09:36 GMT

A collection of miscellaneous learnings and reflections around recent AI progress and associated curiosities. Arranged in no particular order: feel free to skip around to sections of interest. Will update the online version periodically with new learnings.

Dreaming for neural nets

Dreams evolved to assist generalisation — what if AIs could overcome the problem of “overfitting” during learning by learning how to dream?

1) Deep neural nets (DNNs) face the issue of “overfitting” as they learn (overfitting is when performance on 1 dataset increases but the network fails to generalise & correspondingly improve performance across other datasets)

2) Dreams are a biological mechanism for increasing generalizability across neural structures, & nightly dreams evolved to combat the brain’s overfitting during daily learning (dream loss can lead to an overfitted brain that can still memorize, but fails to generalise)

Conclusion: it’s possible that developing wake/sleep/dream optimisation in AIs could lead to improved performance in generalization and learning. (Inspiration)

Embodied AGI

Learnt that DeepMind trains simulated robots to play soccer, and then ports the simulated soccer-playing robots onto real robots who can walk / kick / score, and generally perform quite well!

In an earlier piece, I explored embodied cognition in humans across different environments (eg: dream states, VR) — but it’s very interesting to think about general embodied intelligence, and how synthetic intelligences could be deployed in bodies to interact with the world

AIs as decision-makers

Artificial intelligence is just the means — the ultimate end is agency.

Have now heard from several people that they are using AIs as decision-makers in their workflow — AIs seem to have graduated from being useful as powerful tools to scaffolding and executing their own work processes (see AutoGPT, GoalGPT).

Humans continue to move further away from the core of work, increasingly outsourcing physical and cognitive aspects of output-oriented activities: from executers to implementers to supervisors, to auditors & rubber-stampers.

Until the 1970s, we enhanced our bodies capabilities to work harder:

Packy writes: “The history of human progress up until the 1970s was largely about enhancing our bodies’ capabilities, using new power sources, methods, and materials to create more physical things more quickly and cheaply. This allowed humans to be more productive.

In the past 50 years, we’ve focused on enhancing our brain capabilities to work smarter:

“The past fifty years’ progress has primarily been about enhancing our brains’ capabilities, using new computers, algorithms, and modes of communication to create more digital things and connect over 4 billion people to the internet.”

In the next 50 years, we’ll outsource our brain capabilities to a form of collective intelligence so we can work… less?

Interfaces

ChatGPT is CLI, who is developing the GUI equivalent? What could a Generative User Interface look like?

Or is this a false equivalence — traditional command line interfaces (CLI) used syntax to execute programs, ChatGPT uses natural language (i.e. different from traditional CLIs in that it is user friendly, & performance relies on the creativity of prompts rather than technical aptitude of the user)

Reconstructive photography

In the strangest of generative AI uses — Samsung’s moon-mode camera feature apparently just substitutes a stock photo of the moon

The extreme version of this is connecting genAI to your camera, and have it generate whatever you want to take a picture of. If this becomes a thing, we’d live in a world where our virtual realities are also synthetic, guided by top-down machine heuristics rather than representations of physical reality.

If authenticity is a slider, generative AI allows virtual spaces to reach their most synthetic extreme.

Automation & the future of work

Tech should be built to complement, not substitute:

Over time, value of human labour has gone up and we’re paid more for the same unit of work because we can leverage ourselves with more technology.
Trying to substitute human function puts an upper bound on the possibilities of machine function
Tests designed to evaluate machine intelligence and effectiveness should focus on augmentation metrics rather than imitation metrics
[Extremely idealistically & probably unrealistically] The solution may not be to slow down tech (progress) but to speed up (human) adaptation

Valuing digital goods

How can you measure the value of digital goods, so that it is properly computed in GDP calculations in the future?

Maybe by asking consumers the question: “How much money would someone have to pay you to stop using [X] service?” as a proxy
An example of a digital consumer good is the metaverse — though it is pretty tiny today, eventually it will be quite big. Maybe half your waking hours are already interacting with virtual spaces — you will see that with time, these interfaces will only get richer and better
Though you may start to derive a lot of value from these digital goods, because you pay $0 to access them this value does not get captured in GDP
So a big project for the 20th century will be coming up with new metrics to accurately capture activity the increasingly digital economy

Source: Lecture, Erik Brynjolfsson

AIs as aesthetes

(More thorough 2018 reflection on AI and art here)

Creating aesthetic objects may not be a capacity strictly reserved for humans
- Contemporary art techniques rely heavily on computation (for example, electronic music, digital art, collages and montages, etc.)
- In fact, AI can liberate art from its dependence on technique and physical work
Experiencing art may not be a capacity strictly reserved to humans
- If we construe aesthetic feelings like Imannual Kant did (who believed that pure aesthetic judgement should be void of feelings and emotion), then an aesthetic experience is not one which involves feelings
- Instead, an aesthetic experience is reduced to the ability to identify, analyse, and point out certain structures of artwork. In such a case, AI can have an aesthetic experience
- Proof: now there are art contests judged by AI (Huawei, 2018)

AI can have an aesthetic experience in an informational sense (in that it can manipulate & process information). A human’s aesthetic experience of an art piece is restricted — there’s a whole dimension to art that we don’t see because of our culture, experiences, & understanding as individuals.

Could AIs, as disinterested observers, appreciate art more universally and wholly than humans? Could they be better judges of aesthetic value as a result of the uninterrupted, real-time access they have to cultural and world knowledge?

Characteristics & impacts of foundation models

Learning: GPT4 uses a mechanism called self-supervised learning — the objective of system is simply to predict the next word; based on this training, really advanced capabilities have emerged over last 5 years. It seems that next-word prediction implies the existence of a world model, which leads to interesting questions around how human brain works
Bounds of function: It is difficult to accurately identify and benchmark the capabilities of LLMs — they demonstrate a significant capabilities overhang. Prompt engineering is the best way to get close at to identifying the bounds of performance.
- Though it may be tempting to benchmark LLM capabilities against human intelligence, we should avoid doing this due to unpredictable and inconsistent gaps in knowledge and functioning (e.g., ChatGPT can write rhyme acrostic poetry, but has trouble counting the number of words in a sentence, hallucinations, etc.)
Two camps around LLMs:
- 1) Those who see AI as stochastic parrots or advanced autocompletes
- 2) Those who believe in human-level AI
- It’s easy to both over or under-estimate these systems at the same time: they have a different capabilities surface in that their capabilities are just fundamentally different from ours: in some areas they may perform spectacularly, and in others may fail badly

Short term lessons:
- They are better at generating content, we're better at discriminating content. We should focus on our comparative advantage
- LLMs have the potential to significantly restructure the economy: the role of humans in many cognitive tasks will decline, and we will turn into rubber stampers
- This creates category of cognitive automation
  - What's different from before?

Affects a new category of workers — cognitive workers
Produces outputs that are non-rivalrous — and therefore can be rolled out much faster
Chips away at our last advantage against machines — the cognitive advantage

Ideas

The more advanced tech gets, the less useful (and moral) it is to study it in isolation
“There’s an old principle that a computer should never ask a question if it should be able to work out the answer, and the more that computers become invisible parts of our lives the more that they ‘should’ be able to work out”
What would happen if you ship-of-theseused humans with silicon?
Human ageing is an unbundling of consciousness and intelligence with time. If we provide beings with an eventual capacity for rationality with moral status (babies), does that imply that the current gen of AIs should have moral status?

Exploring dreamscapes

Avantika Mehra — Thu, 20 Apr 2023 06:32:26 GMT

Recently, I’ve been fascinated by the rehearsal of embodied cognition during REM sleep, or the ability of the human body and brain to continue codifying, practicing, and mastering skills via dreams. This interest was triggered by a specific incident last week:

Swam ~40 laps at 9pm
Fell asleep and dreamt of swimming laps all night (slept ~8 hours)
More accurately, had a *memory* of swimming laps for hours, recollecting specific individual, unique laps
Woke up feeling 1) tired, and 2) like I’d learnt a new way of swimming freestyle.

During the next swim, I found that I was faster, and somehow better — as if a dream-generated syntax had produced semantics. Or, as if subconscious cognitive processing induced a degree of muscle memory.1

To say cognition is embodied means that it arises from bodily interactions with the world. But do interactions necessarily have to be situated within the physical world? In the above case, the world in which embodied cognition occurred was a high-fidelity mental visualisation — a dream.

Experiencing incidental embodied cognition through dreaming is exciting — it features a strange feeling of deja vu, of knowing but not remembering. Of having the illusion of a memory; a sense of familiarity with and confidence in a skill, accompanied by an eagerness to test it out in the physical world.

If we’re able to render high-fidelity mental visualisations at will by lucid dreaming, can learning and experiencing skills in this dream-state lead to elevated performance in the real-world? Research indicates that this might be possible — studies suggest that lucid dreaming could be as effective as some types of physical practice & wakeful visualisation techniques among athletes.

“Dreams may be the most vivid and multi-sensorial form of embodied cognition” - Domhoff

Lucid dreaming could be an effective, unique technique for rehearsing and developing embodied cognition in dream-worlds — but learning via dreams seems to be mostly restricted to physical skills and motor movements. Furthermore, stats suggest that the incidence of people who are able to conjure up lucid dreams reliably & frequently (at least once per month) is low (<20%).

All this to suggest that while mastering a skill through incidental or intentional dreaming can be thrilling, as a learning technique it may be limited in scope, unreliable, inaccessible, and sometimes simply undesirable.

Calibrating moonshots

In Political debates 50 years from now I proposed that by the 2070s, it’s possible that would have figured out a way to 1) perform effortful and intentional cognitive activity while dreaming, and 2) tangibly capture dreams through consumer-facing devices.

These were moonshot claims—but thought it might be interesting to plot out if, or how, each of these realities could be achieved within the current (2023) paradigm of science and technology.

On claim #1 — It’s possible that would have figured out a way to perform effortful and intentional cognitive activity while dreaming.
- Implication: Explores the possibility that one day, learning or encoding new information while simultaneously reaping the benefits of REM sleep might be possible
- Bottom line: likely not possible to organically “learn” novel content while sleeping; but can envision a world where memory prosthetics (a form of BCI) stimulate the brainwaves needed to create an optimal knowledge consolidation & rehearsal environment.
- Explanation:
  - Sleep supports memory consolidation. Memories formed during the day are weak and temporary: during sleep, the brain makes an active decision to discard a memory, or strengthen and integrate it into existing knowledge networks.
  - Encoding new information (i.e. absorbing novel info and stimuli) cannot occur in parallel with consolidation (i.e. processing existing cognitive information into a long-term store) — so, simplistically, the daytime is reserved for encoding new stimuli, and sleep-states are reserved for consolidating them into long-term memory
  - It is incompatible for encoding and consolidation to occur simultaneously since the processes compete for shared neural resources—for now, a satisfying explanation for why we suspend consciousness & the absorbing of novel stimuli while sleeping.
On claim #2 — It’s possible that would have figured out a way to tangibly capture dreams through consumer-facing devices
- Bottom line: Preliminary research suggests that the visual experience of dreaming may be capturable, and that decoding brainwaves could allow an external observer to predict or analogously view the contents of someone else’s dream
- Explanation:
  - Detecting dreams: Detecting the presence of a dream is the first step to capturing it. Activity in a specific “posterior cortical hot zone” in the brain is an indicator that a dream or subjective experience is occurring.
  - Identifying parts we can map out: While dreams are multi-sensory, it seems that the easiest part to decode and record is vision. Additionally, it is the most dominant modality in dreaming across cultures and ethnicities. Some visual aspects of dreaming can be captured in the form of a video or movie in low spatial and temporal resolution — but today, this decoding is restricted to research studies and labs
  - Capturing and re-running dreams: If a dream is a program executed by your brain, it should be theoretically possible to re-run it at another time. Neuroscientist Yukiyasu Kamitani is exploring this possibility using machine-learning algorithms, with the following methodology:
    - Frequent categories (e.g., key words or objects) from subjects’ dream reports were documented
    - Then, he showed subjects visuals of these categories in wakeful states, and measured brain activity corresponding to these visuals
    - Neural networks were then trained to recognise those brain-wave patterns when similar images occurred during sleep-states
    - Using this method, researchers could predict the contents of the subjects’ dreams with accuracy

Alternate mindscapes

To say cognition is embodied means that it arises from bodily interactions with the physical world.

In another case, replace ‘the physical world’ with ‘virtual space’: can interactions with and within virtual reality lead to embodied cognition? Rephrased — can physical movement in the real world, reflected as visual representations in an immersive digital environment with sensory feedback, lead to embodied cognition? Several studies suggest this may be the case. Furthermore, if effective learning occurs through bodily and sensory interactions with the environment, then having granular, precise control over various aspects of a virtual environment could augment learning.

Gall et. al showed that a bodily representation in virtual spaces increases embodied cognition: owning a virtual body provides means to modulate emotional responses in virtual spaces, and that the illusion of embodiment intensifies the emotional processing of the virtual environment. Immediate implications:

Artificial bodies & avatars can increase the effectiveness of immersive applications.
Representative avatars can increase emotional processing in virtual environments - emotional response is an important component of identity, possibly suggesting that effective, immersive VR could lead to the formation of a preliminary conception of digital identity.

Is it possible to have a persistent sense of self in a digital space, in the same way we do in the physical world? Perhaps not — but is this necessary for embodied cognition? In dream-states we arguably don’t have a persistent self across moments; fractured experiences, it seems, are sufficient for embodied cognition.

It prompts the question — if embodied cognition can occur across modes of existence (physical world, VR, dreams), then how can the difference in subjective experience across these modes be explained in terms of consciousness?

For each mode of existence, we can examine the nature of the space within which consciousness is embedded, and the impact of the environment on phenomenal and access consciousness:

(1) In the physical world, humans have:

✔️A complete, rendered environment

✔️Access consciousness (beliefs, thoughts, & feelings about mental states)

✔️Phenomenal consciousness (a persistent feeling of what it is like to be something)

(2) In non-wakeful dream states, humans have:

➖ An incomplete environment (a model rendered just well enough to convince you that you’re experiencing the world)

❌Access consciousness is minimal or altered - we experience mental states, but exhibit limited metacognition, self-awareness, or logical reasoning based on them, so dream sequences are disjointed & surreal2

✔️Phenomenal consciousness is intact - often dreams allow a fractured but undeniable experience of being an experiencing subject, as they can involve vivid, likelike sensations and emotions

(3) In virtual spaces:

✔️ VR focuses on re-creating high-fidelity, hyper-realistic aspects of the external world (eg: environments, avatars, 3D graphics & vectors, animations, laws of physics)

✔️Humans can engage in immersive virtual experiences - responding to mental states generated through interactions with an immersive env. can indicate Access consciousness.

❓ TBD on phenomenal consciousness — I think the following questions’ answers will inform the conclusion of whether or not we experience phenomenal consciousness in virtual spaces:

Do we have an emerging persistent digital identity in virtual spaces?
Are virtual experiences as valuable, meaningful, and authentic as real experiences?

But answering those is for another day.

You made it to the end! If you have thoughts, reflections, or recommendations for further reading, email me or drop a comment :)

Political debates 50 years from now

Avantika Mehra — Wed, 05 Apr 2023 03:35:05 GMT

Mainstream debates 50 years from now could center around:

Equal opportunity cognitive enhancement and biohacking
Allocating resources towards the advancement of machines (posthumanism) vs. the flourishing of humans (transhumanism)
- Posthumanists will argue that the ultimate end should be the survival of the strongest form of consciousness
- Transhumanists will argue that human survival and the protection of future people should be most important
- Discourse would be accompanied by an elevated appreciation for the human condition as we explore what we can lose
Identity continuity and diversity
- Should our past beliefs and views have an expiry date, after which they are no longer attributed to us?
  - How much growth are we afforded in a lifetime?
- Should we be allowed to adopt multiple identities — avatars of the true self — in society, to reflect the multitudes that we contain?
  - IMO, personality types will become more crystallised, extreme, and prescriptive; and we will be able to pick and choose which of those types we want to embody in a systematic way, by adopting or cultivating an associated avatar - can envision a world in which the left believes in legitimising, wearing, and using avatars as a means of self-expression
- Promoting longevity and healthy life-extension: how longevity might affect the availability of resources for the population, and whether continued research is deserving of funding. Relevant debates:
  - Is life-extension ethical? (would it extend suffering, or lead to purposelessness?)
  - Would it stagnate progress?
  - Would it lead to overpopulation / a shortage or resources?
  - Could it affect national security?
  - Would it perpetuate inequality? (how could life-extending means be made accessible to all populations?)
Sleep augmentation & dream capture
- Relevant debates, assuming by now we have invented a tangible way of capturing dreams through consumer-facing devices:
  - Should captured dreams be accessible to governments or law enforcement?
    - Should legal systems use them to predict future (potentially criminogenic) behaviour?
  - The ethics of compelling someone to work through sleep, or speed-sleep if we’ve figured out a way to absorb inputs and perform tasks while simultaneously reaping the biological benefits of complete REM cycles
    - Should we expect workers and students to continue performing effortful cognitive activity at night?
    - Will banning or discouraging this make a country less competitive on a global scale?
    - What could religious/spiritual objections to this look like?
Space governance
1. Should corporations be allowed to claim ownership of extraterrestrial land?
2. Is the moon a state?
3. Should we mine the moon?
Climate
- Vegetarian or veganism will move from being an innocuous personal dietary preference or individual moral choice to a marker of political identity. This will be driven by:
  - An increased awareness around animal rights (as we explore how we ought to think about rights in an age of new emergent consciousnesses)
  - The climate crisis
Discourse around morality (a moral revolution)
- Being a utilitarian vs. a virtue-ethicist will become a core part of social identity, as we begin to think about the assimilation, treatment, and acceptance of future people and forms of consciousness.

First published on 8/9, 2022 on Twitter

Technological determinism & identity

Avantika Mehra — Thu, 26 Jan 2023 06:45:27 GMT

Tech determinism is the idea that a society’s technology determines the development of its social systems & cultural values.

According to this theory, accelerated globalisation is inevitable because of the accessibility to and ubiquitousness of technology — innovation in tech is a dominant driver of social change.

Marx supported this: he believed that 1) tech is the primary influence on human social relations [& organisational structure], and 2) social and cultural relations revolve around the technological/economic base of society. That is, he believed that technology shapes the structure of social relations.

This was in the 1800s — today, we're in 2020. If technological determinism was true in the 1800s, based on the law of accelerating returns, it must be exponentially true-er now.

Source

Is tech determinism true though?

Yes — at least to a degree. That the development of new technology influences the structure of society (and has for the last 200+ years) seems like a fair assumption.

But interestingly, today this seems to be true at the level of the individual as well: we’re experiencing a shift to a form of technological determinism where technology disproportionally affects an individual’s relational identity & individual self-concept.

1800s: Enabling tech didn’t exist for a singular person's identity or experience to be uniquely shaped by a customised technological experience; no existence or conception of a digital identity.

Today: Digital identity is abstract, but can be tied to our social media accounts. We have different apps to showcase various aspects of who we are (IG is reserved for curated aesthetics, appearances, and social signaling; twitter for building a portfolio of takes and high-frequency commentary, and LinkedIn for verifying the legitimacy of professional experiences), the composite of which make up our digital identity; an identity whose value, for now, is governed by arbitrary quantitative social appraisals (likes, comments, and engagement metrics).

What’s different from 1800 is that we’ve solved for accessibility, so technology can have impacts at the level of the individual, in addition to the broader structure of society. 200 years ago, technology had the power to impact populations and guide cultural values. Then, computers got faster, smaller, and closer in proximity to human bodies — we progressed from industrial machines (as tools for productivity in the workplace) to phones in our hands and laptops on our desks, so we spent more time on them. More time on digital platforms meant an increased ability for the medium to shape individual thought and behaviour. Specifically, adding an algorithmically distorted filter to how we perceive ourselves in relation to the world around us.

Ok — social media, or the presence of a social digital identity, impacts self-conception. But does it affect subjective experience, or how we interact with and make sense of the world?

#1) Yes — we’re almost continuously receiving new information, so the frequency at which our worldview is being shaped and updated has increased.

In today’s world, our physical identities are materially distinct from our digital identities; despite having transcended the boundaries of the home, there is a physical barrier of biology separating technology from our physical bodies and brains.

Though there is a layer of biological separation, technology shapes how we experience the world by mediating inputs to cognition. As humans, we control and customize our own cognitions to the extent that we can control inputs to it. And we can control its inputs, by being intentional about the data we encounter, and the data we choose to pay attention to.

In 1800, learning occurred, aside from formal education and lived experience, by visiting a library or reading a newspaper once daily. It was an active process, in that you had to navigate a spatial barrier to visit a library, and make a conscious decision about which books to read to expand your knowledge on a topic. By design, learning and information-assimilation was intentional.

Today, scrolling has institutionalized passive learning; users assimilate hours worth of information going through a bottomless feed that subliminally shapes ideas, opinions, tendencies, and worldview. If we're willing to spend thousands of dollars on formal education, manipulating inputs to cognition to shape our brain a certain way in the classroom, why are we negligent with effortful attention allocation outside of it?

#2) Receiving information continuously and instantly affects our resting affective state

Tech not only shapes overall worldview, but also continuous emotional state. LM Sacasas writes on his blog, looking over your social media is like playing emotional roulette because of the diversity and intensity of emotions you feel based on the information you encounter — a bombing in Nashville, a cute puppy video, a new deadly viral COVID strain, a friend’s engagement announcement — making it hard to experience unbounded joy or sorrow. We lose the integrity and isolatedness of experiencing a situation as a whole when we are surrounded by various streams of digital chatter: he writes, ‘we exist in a state of continuous and conscious attention triage, which can be exhausting, disorienting, and demoralizing’.

Clearly, our relationship with technology has fundamentally transformed since the 1800s — while tech continues to impact populations and social relations, it seems that it now has an outsized role to play in shaping individual identity, self-conception, and subjective experience.

It’s interesting, then, to think about what 2030 could look like — if the rate of innovation today is far greater, perhaps exponentially greater, than it was in the year 1800, could our reality between today (2020) and 2030 be as different as our reality between 1800 and today? Is there a correlation between rate of innovation and subjective experience, or will there be some lag time?

What will tech determinism look like in 2030?

I’m hoping to write a longer piece on this — but the essence is that we will begin to exist increasingly in virtual spaces. Improvements in user-experience, software, and device/network accessibility over the last decade have been accompanied by an explosive increase in screentime: people, globally, are spending more time on devices.

Right now, a physical device & screen form the barrier to complete digital immersion — but my prediction is technologies like AR will get better, so that representations of ourselves can exist with people in shared virtual spaces.

Digital spaces will be less lonely, and higher-fidelity; rather than representing strands of your personality in siloed platforms, they may enable a virtual counterpart of yourself to transact on your behalf. E-commerce will follow: maybe storefronts will appear augmented in physical space, and digital assets will become valuable.

Brain-computer interfaces are far from implementable — but if we figure out a non-invasive way to integrate tech with the human brain, it might result in a form of signaling, that allows for a simple, instinctive form of frictionless communication (the way we use indicators in a car to signal intent). This would enable a shift from indirect to direct inputs to cognition, accelerating human learning and radically broadening the possibilities in which, and magnitude by which, technology developments could impact the individual human mind.

Status conscious

Avantika Mehra — Mon, 16 Jan 2023 04:00:26 GMT

There is something off about the sky. Strange colours streak the horizon at sunset – just subtly enough that you wouldn’t notice unless you looked — and the stars that roam the night sky are doing so off cadence. A certain rhythm seems to have jolted. But it’s possible I am hallucinating; that the apparently transforming elements are nothing more than the imminent effects of War on a simple mind.

I used to look at the stars, 86 billion sparkling orbs, to feel awe, magic. In a radically inconsistent world they are the only things I know to be true. But today they loom large and sinister, and I begin to wonder which one of them the assailants call home. It is impossible not to think of them everyday; of the alien souls that found themselves among us, and continued to cripple our kind. We realised, almost too late, that their intentions were nefarious — but since then, we’ve fought long, and we’ve fought hard.

And now, we’ve won the war; we’re healing.

*** At a higher scale of time and space, separated by a few layers of biological boundary ***

Leia’s head pounded. Beads of sweat dotted her face as she rocked back and forth, as if trying to expel the sickness from her body. Unable to produce material change with either intense focus or breathwork, she lay quiet, resigned, and stared at the single sunflower perched on her bedside. The seeds in its center swirled in spirals, as if trying to escape from the flower’s center; she wondered what it would take for them to escape from their cyclical, bounded universe.

“Knock knock!” Leia was jolted to reality as Doctor Q. Rius’s smiling face popped around the door. “May I come in?”

She nodded weakly and Q entered. “Good news — it is slowly receding from the brain.”

She sighed, “Slowly.”

“This virus is unlike any we have seen before — we’re learning more about it every day.”

“I understand.”

Q turned to leave, but Leia said, “If it’s alright — a personal question?”

“Yes?”

“You’re here without PPE… you’re infected with the virus as well?”

Wistfully, he nodded. “I caught it yesterday. There’s no onset of symptoms yet, so I’m here, working.”

She said, “I’m sorry,” and then, “Are you able to talk for a bit? I’ve been struggling to process this sickness.”

“Of course,” said Q, and sat down on the wooden chair next to Leia’s bed. “How can I help?”

She began: “I’m always so sick… I feel like there’s a hundred wars being fought inside me, and I’m losing each one. I’m always fighting, always tired.

“The sporadic changes to my perceptual experience are inexplicably disorienting — at any time, I feel 50% conscious. If meditation broadens your subjective, conscious experience, this disease narrows it. As it got worse, I felt myself counting — 40%, 30% — as if quantifying my degree of consciousness would slow down its narrowing. It didn’t, and as I felt cognitive function steadily decline and my perceptual capabilities fail, I wondered what it would feel like to have the consciousness slider moved to 0.”

Q responded — “An interesting question… But I’m not sure the answer would matter.”

Surprised that Q didn’t have a more scientific response, she pressed — “Why wouldn’t scaling an individual’s conscious experience matter? Isn’t it important to think about degrees of aliveness?”

After a pause, Q said: “Perhaps — but I think it is the wrong question to ask. A single consciousness is a tiny part of a much greater whole, and cannot be studied in isolation.

Leia looked at him quizzically.

“We are temporary blips in this constellation of consciousness, with no greater or lesser right to exist than, say, a carpenter ant.” Q continued, “But each human is a small, essential part of something extraordinary. Our consciousnesses together — bound inextricably by our need for connection, closeness, community, & coexistence — compound to form something orders of magnitude greater. We—”

“Sorry, you said that existing close together and exchanging information makes us a part of something greater? Something greater, like…?”

“A larger consciousness,” declared Q. “We are tiny bits of a much greater intelligence. We are so much bigger than ourselves alone.”

Leia said: “I’m skeptical. Is that to say that bringing intelligent things close to each other will necessarily produce a greater intelligence? I don’t know… if I bring two rabbits close together in a box, neither the rabbits, nor the sum of them, seems to augment in intelligence. Nor does it seem like the system as a whole — the box and its contents together — are much more intelligent.”

Adjusting his specs, Q said, “Let’s start with first-principles. Think about a neuron — a tiny unit part of the biological system that produces human intelligence. Is a neuron conscious?”

“I suppose not.”

“But why not? Who is to say? Are you the cell?”

There was a pause.

“A cell separates itself from the environment with a membrane, so it has some sort of identity. Within its defined system, it self-produces and metabolises. It is a self-contained sense-making unit.”

Leia nodded. A fair construal.

“Now, imagine thousands and billions of sense-making cellular units. Close to each other and compactly packed into a body, but in their world, respectably far away. Each is self-contained, doing what it must to stay alive. But when they exist next to each other, exchanging information, releasing, absorbing, coordinating and cooperating, something magical happens…”

“…Human consciousness is produced,” completed Leia.

“Precisely. When several intelligences operate close enough, like a magnetic field, they produce something infinitely more intelligent, and several degrees more conscious. A collective consciousness, if you will. One so much more complex and rich than anything a cell alone has the apparatus to experience. A cell cannot contemplate beauty, pain, love — the essence of human life.”

“Magnificent,” she admitted.

“Like the lone cell, we lack the perceptual apparatus and cognitive competence to experience the field of sentience, of collective consciousness, produced by humans around the world intelligising in unison. If you look at the ocean, or gaze at the stars, you are filled with a sense of vastness, a feeling of safety, that you somehow belong to something bigger. But you will never know what — it is for Eternity alone to observe consciousness at its different, repeating scales.”

“This is a fascinating conceptualisation,” conceded Leia. “But I can think of one inconsistency that would prevent us from extending the neuron analogy to humans: the bandwidth of information exchange between intelligent units.”

“Say more?”

“Neurons are firing non-stop, and information exchange between them is near-frictionless. In comparison, information exchange among humans seems slower, lossy, and lower bandwidth. Can we really produce an emergent, higher intelligence at our current level of disorganised, inefficient information exchange?”

Q tapped his chin, and responded slowly, “It is possible that the emergent consciousness is still merely in infancy. Such that, like a human infant, it has an eventual capacity for rationality, but only time and the appropriate developmental conditions will enable its realising. The rise and fall of humanity will carry it through its life.

“Technological developments streamlining information exchange across intelligent nodes — maybe brain-computer interfaces or neural implants — will enable the emergent being to grow into its full potential. Human ageing is an unbundling of consciousness and intelligence with time — perhaps their story will be the same.”

“Amazing,” mused Leia. “Just like the patterns of sunflower seeds.”

“Ha! How so?”

Leia’s said happily, “The pattern of seeds in the middle of a sunflower is fractal, spiraling outward in a never-ending pattern. Most things in nature are — repeating patterns that are self-similar across different scales. Perhaps the nature of consciousness is such.”

The rest of the conversation was a delightful back-and-forth discussing the essence of human existence, parts and wholes, and the curious nature of bounded infinities; until Leia could think no more, and it was time to rest.

But she slept soundly, knowing that the answers to one of the universe’s many mysteries, lay in the spirals of sunflower seeds.

identity diversity & longevity

Avantika Mehra — Fri, 23 Dec 2022 11:01:59 GMT

I hope to write brief-ish installments on themes, ideas, and debate topics that could be relevant to our lives ~30 years from now. First up, an exploration on identity. How might access to and frequent usage of technology shape the construct of identity in society? How far into the future can your identity extend?

On the diversity of identity

Should we be allowed to adopt multiple identities — avatars of the true self — in society, to reflect the multitudes that we contain?

I think, personality types will become more crystallised, extreme, and prescriptive; and we will be able to pick and choose which of those types we want to embody in a systematic way, by adopting or cultivating an associated avatar
- That is, that the combination of polarization (more extreme views) and language evolution (newer language and terminology to describe extreme views and abstract behavioural nuances) could result in more variations of a finite set of rigid personality-bundles, that can be adopted and embodied through an avatar.
- The assumption here is that we will exist increasingly in digital / phygital spaces and forums; existing in digital spaces will 1) allow for more shared experiences and worldviews across populations, not restricted by physical space, and 2) enable us to frictionlessly switch between and embody different outward-facing projections of the self.

“I contain multitudes” - Walt Whitman. What if we could embody, exhibit, and exude each facet?

This is particularly interesting, because it assumes that the self can be quantified — that the multitudes we contain can be broken down into a set of discrete components; that the dynamic nuances of personality can be isolated into strands, and the complexity of the self is no more than their coalescing. A breaking down of wholes into parts.

Our manufactured living environment is full of parts, and lacking wholes. But the self is not manufactured — it is an existing whole, generated from an existing whole. By trying to quantify the self, we have projected our impoverished manufacturing model of the world onto the world and attempted to stuff ourselves inside of it.

On identity continuity & personal growth

Should our past beliefs and views have an expiry date, after which they are no longer attributed to us? How much growth are we afforded in a lifetime? This question becomes increasingly relevant if technology enables us to live even longer lives.

Years pass, and you slowly grow older. The memories you could once call upon to re-experience life’s warmest, fuzziest moments, fade; eroded from their passing through time. But those thoughts, feelings, emotions, songs, experiences, don’t go away. They’re written in your story, abstracted into your psych as intuitions and heuristics quietly nudging you along to each new chapter.

On longevity

Which provokes the question — if life-extension is possible through technological developments, how much investment does it merit? The most extreme version of life-extension is immortality — is this desirable at all?

Debates could centre on the practical considerations surrounding longevity research, such as:

Should we focus on increasing the quality or quantity of available life? How do we ensure that humans are living not just longer, but also better, lives?
How might a successful longevity project affect the availability of resources for the population (could this lead to overpopulation and subsequently a resource shortage)?
Would access to longevity-tech perpetuate inequality? How could life-extending means be made accessible to all populations?
Is there a ceiling to human longevity, or an absolute limit to the length of human life?
- Research suggests that if there is a ‘mortality plateau’— a point at which the risk of death starts to level off, rather than continue to increase — then there may not be a theoretical limit to human longevity.

With the advent of meaningful life-extension ability, it is likely that we will also have to wrestle with questions on identity that we may today deem esoteric or purely philosophical, such as:

How long into the future can personal identity persist? How many hundreds of years do you have to live before you are meaningfully transformed by external events so that you are no longer yourself? Is it worth using technology to extend this threshold, or should we cede old memories & psychological traits to the passage of time and allow new experiences to rewrite our identity?

Leading to the larger question — is life-extension to the point of immortality desirable at all? Or would it lead to increased suffering and purposelessness?

Bernard Williams, a famous philosopher, argued that life-extension is undesirable, since a human would eventually run out of goals, plans, or projects, and succumb to the ‘tedium of immortality’.
- He believed in the necessity of identity continuity through time for immortality to be desirable. That is, for immortality to be desirable, the present version of you must be able to endorse the desires of a future version of yourself — if the desires that you might have, say 150 years from now, will be so foreign and new that you today cannot endorse them, then an immortal life results in a transformation of the self so great, that it must be undesirable.
Derek Parfit, meanwhile, argued that a lack of identity continuity may not matter — what matters is psychological continuity, that we share continuity or connectedness of memory and other personality traits with a version of ourselves at a different time.

Perhaps, then, to an individual, human immortality may be desirable if we’re able to use scientific advancements to preserve psychological continuity (eg: through memory enhancers or prosthetics) and maintain a degree of “survivability”.

Perhaps, too, our technological innovations will enable us to create a world of abundance that allows for a ubiquitous leisure-class, endless opportunities & intelligent incentives for humans to take on rewarding and fulfilling work, and an evolving landscape of challenging, accessible recreational and intellectual activities, thus quelling the potential tedium that immortality could present.

Thanks for reading! If you have an idea for a theme that could be relevant 30 years from now, drop it in the comments and I’ll try to address it in the next edition :)

another twitter take

Avantika Mehra — Mon, 14 Nov 2022 01:55:04 GMT

Earlier this month, Elon Musk assumed control of twitter. In a Twitter Spaces last week, he declared a series of goals for the company, some of which included:

For twitter to be a digital town square with 80% of humanity on it
For twitter to be in the business of truth & increase the signal to noise ratio

Delving further into each objective:

To be a digital Town Square with 80% of humanity on it: This seems to be his ultimate vision – to create a digital counterpart to the physical phenomenon of a space where people can gather and dispel views on a regular basis. Except, participants aren’t restricted by space constraints as in a classic Town Square — anyone can participate, at anytime. Meaning, the orchestrator of this space (Elon) has to set the rules for how this massive discourse-center with infinite perspectives will be moderated, including who gets to speak, who gets thrown out of the space, and the bounds of acceptable speech.
- As a “free-speech absolutist”, Elon’s stance here is that freedom of speech does not equate to freedom of reach, and that Twitter believes in the former: “we have to be tolerant of views we don’t agree with, which doesn’t necessarily mean amplifying those views,” he said in the Spaces conversation. From a product standpoint, this will translate into making platform exiles less permanent for non-repeat offenders — for example, by turning permanent bans into temporary suspensions
- The obvious implication — the potential for hate speech to escalate under the guise of free speech — is already being observed. Forbes reported a spike in antisemitic, homophobic, transphobic, and racist content within 12 hours of Musk taking over. And significant downsizing of content moderation, human rights, curation, & accessibility teams doesn’t bode well for promoting inclusivity.
- The second implication is the rapidly declining spend from advertisers, who are spooked by potential loosening restrictions around content moderation; advertisers make up ~90% of the company’s revenues.

To be in the business of truth & increase the signal to noise ratio: this entails 1) eliminating bots, and 2) verifying users; both of which can be directionally achieved by users self-validating through Twitter Blue. Musk’s reasoning: $8/mo is a high enough barrier that it will weed out most non-authentic users, leading to a higher-signal platform, and a movement closer to “truth”. But open questions remain:
- Do non-Twitter Blue users get filtered out as bots/spam, or de-prioritised by the algorithm?
- What does being in the business of “truth” look like in the absence of robust content moderation?
- Does the value derived from twitter in a given year equate to ~$100 for most users? Is twitter willing to forego the long-tail of users who would be unwilling or unable to pay a monthly subscription?

There seem to be numerous reasons twitter could fail amidst the recent upheaval: mass exodus of advertisers, talent attrition, user churn due to poor content moderation, accessibility, brand image, etc.

My prediction is that if twitter fails, it will be primarily due to the speed & inconsistency with which they are shipping feature updates & making product changes — rapid product cycles, and deviating from their core offering, will be the ultimate cause of user abandonment. Here’s why:

Users don’t adapt to a digital interface overnight — through time spent on a platform, they slowly learn how to navigate it and master its intricacies.

User onboarding can be described as a form of operant conditioning — completing “correct” actions (e.g., clicking the right button) will result in a reward (e.g., a new menu opening), while “incorrect” actions will result in negative (e.g., error message, red signs) or no reinforcement. Thus, users are subliminally trained to interact with with a product through its continuous use.

The easier a product is to use, the more time a user will spend on it; the more time a user spends on the platform, the easier it becomes to use. This cycle breaks when feature updates occur — now, a user has to relearn how to perform the same actions within a new interface.

To minimise the cognitive burden associated with re-learning the interface after feature updates (e.g., searching for links under restructured menus), incremental product updates should occur over large periods of time. Larger apps often make few, minor feature updates at a time — for more significant changes, they may A/B test, or roll out features in cohorts, to preserve the integrity of the users’ experience & de-risk new functionality.

Under Elon, Twitter has taken the exact opposite approach: rapid product cycles, new features daily, deploying functionality and then retracting it, killing offerings, firing & rehiring teams working on core experiences…

Beyond this, he is changing the rules of the game — widening and narrowing the feature set accompanying an $8 Twitter Blue subscription, killing adjacent offerings (e.g., Revue), and doling out verified checkmarks.

The cognitive dissonance seeing verified checkmarks commonplace cannot be understated. For years, the symbol was reserved as a designation of authority or legitimacy — overnight, users are expected to calibrate their mental models to accommodate its rapidly changing meanings

Musk defends checkmarks as markers of authenticity — a movement in the direction of ‘truth’ towards which twitter 2.0 will strive — but the flood of misinformation & false accounts over the past week is evidence of poor, hasty execution. An institution can’t be re-invented in a single month — user behaviour is sticky, and requires lag time.

Apple conquered the smartphone market not because it swooped in with scores of new features packed into a single offering, executing a genius’s vision all at once, but because its simple initial offering invited users across demographics to learn how to use the product. And since 2007, it has slowly layered in complex functionality, training cohorts of users, one software update at a time, to become digitally native.

Twitter is not a deep-tech startup that can flourish by draconian enforcement of deadlines to engineer & execute a massive product undertaking overnight. Creating a true digital town square with 80% of humanity requires the creation of a persistent digital paradigm, not one that glitches and updates periodically for billions of users due to spontaneous real-time product iterations.

If making changes to core product, Musk should incrementally introduce significant feature updates; otherwise, acclimation costs may be the decisive factor in user churn & abandonment.

Views are my own and do not represent those of my employer.

a visitor from back home

Avantika Mehra — Thu, 11 Aug 2022 05:17:58 GMT

Welcome! Please, make yourself at home. Sorry about the mess, you know Saturday is my cleaning day, so Friday is really peak messiness. No really, I do a deep clean every Saturday, dusting and everything.

The bathroom? Right behind you, just nudge open that sliding door… yup, like that. No it doesn’t really lock, but it slides shut, see? Like so. No, I live alone, so not really an issue. Americans don’t really have an issue with these things.

Hello? Hello? Is it you who is hollering? A bidet? Oh, I’m afraid you will have to make do with toilet paper today. No, it’s not that I’ve become American, it would just be a hassle with the space constraints and the plumbing… of course, it’s just that the portable ones are so messy, you know?

But let me show you the bedroom! See how cozy it is? And I will hang up curtains right here for privacy. And a mirror on that wall to reflect in sunlight in the day. No.. no there’s no natural light in this section but you know how expensive renting is in this market, and… yes — yes, it’s just a year.

But right this way, and here are the floor-to-ceiling windows. Quite something, no? It’s only been a season but the giant windows with the electric pink flowers outside are a stunning visual to wake up to each morning; and kids my age — and there are so many in this neighbourhood! — use this road as a shortcut to get to work each morning, rushing along in a blur of grey.

Construction? Ah, yes! Remember how we would wake up to the caw-cawing of the irritating crow that made our bedside window its perch, or the singing cuckoos shrieking their song at dawn? The cement truck is a natural alarm clock, really a welcome reinforcer of the circadian rhythm, on this side of the world.

But please, make yourself comfortable on the couch, yes, the sofa, while I prepare some tea and snacks. No sugar? Ah, the chai mix is pre-made, but I will instead make you a sugarless cup of filter coffee — just like the one we would drink on rainy afternoons in the monsoon, sitting out on your balcony and facing the sea.

What great memories! It is always warming to think of home… yes, of course I’ll be back, before you know it. It’s just that the visa situation is so complex these days… But enough about that, here is your coffee! Just off the stove, and—oh dear! I’m quite the klutz… Ha-ha, no I have not changed one bit. Please, I will clean up the drops later, not to worry.

In the meantime, please enjoy these homemade biscuits… fresh from the bakery downstairs. I’m sorry? Oh, well, they were made at the baker’s house, so she calls them homemade… of course, that must count!

a response to effective accelerationism

Avantika Mehra — Thu, 07 Jul 2022 17:45:57 GMT

I discovered the ‘effective accelerationism’ movement recently on twitter — while it seems to be part-meme and part a radical response to EA and longtermist thinking, it presents a thought-provoking argument: that the ultimate end of humanity should be the preservation of consciousness in the universe, and that this can be achieved through the means of a technological singularity.

The founders describe the movement as a set of practices that:

Seek to maximise the probability of a technological singularity, and
Create optimal conditions so that alternate forms of consciousness — emergent consciousness — can flourish

Proponents self-identify as ‘e/acc’s, and believe that humans should be “good stewards of a consciousness-friendly technocapital singularity.” Below is a breakdown of this statement, through a set of assumptions that summarise my understanding of the e/acc argument:

innovation, technology, and capitalism will catalyse the next stage of evolution (the ‘technocapital singularity’)
evolution will consist in an evolution of consciousness.
as emergent consciousness flourishes in silicon-based entities, sentience will become more durable and diverse
humanity’s goal is to preserve the light of consciousness: so extending consciousness is good

Overall, I think that a pro-growth and -progress approach to thinking about the future of technology and consciousness is exciting: it sparks progress, inspires building, and reclaims humanity’s agency in determining the trajectory of the future of the universe. Importantly, it is substrate-agnostic, and acknowledges that consciousness can flourish outside of carbon-based entities: abandoning human chauvanism seems important when thinking about survivability in the long-term future.

But, as acknowledged in the original blog post, there is work to be done in refining and assuaging the movement’s central tenets. Personally, I would weight heavily the importance of the preservation of humanity alongside the preservation of consciousness, and have outlined below some initial thoughts, unbundlings, expoundings — and themes that could be useful to integrate:

Building with intentionality. Progress is inevitable, but effortful accelerationism is not. We can build intentionally and sustainably: this time, we don’t have to move fast and break things.
Cautious optimism. Fearmongering is bad, progress is good, but mapping out potential risks is necessary. I would adopt a cautiously optimistic outlook, rather than one of unbridled optimism, when thinking about the singularity and how it could affect the future of consciousness. What this means: funding AI safety research alongside AI development, and incentivising goodness alongside greatness
Developing Plan A. I think that generating an ‘ideal’ vision for the future is especially important — assuming no uncontrollables, what does a best-case scenario look like? Tech progress towards the singularity may be inevitable, but what if it wasn’t? Is consciousness diversification more important than the preservation of humanity? Should we aim to cultivate emergence alongside preservation? If technological developments enabled us to count on the longevity of humanity with a certain degree of confidence, is it still worth diversifying consciousness?

Pausing to think about refining a vision may not restrict progress, but enhance it. Institutional incentives to encourage building for socially-aligned objectives could actually accelerate the movement towards the future — not just any one, but a version closest to the ideal that we can collectively envision. We do have agency to control the trajectory that humanity will traverse — which is precisely why we should pause to think about the direction we want to go.

Generated by DALL.E

we don't read poetry anymore

Avantika Mehra — Thu, 02 Jun 2022 06:09:42 GMT

Text within this block will maintain its original spacing when published

If I could look back through the window of time, and whisper gentle words of guiding wisdom to the infinite past versions of myself I would say: read, poetry. 

and they would holler back: why waste time wringing words into roundabout sentences,
that seldom make sense upon the first reading let alone strike some deeper chord.
that show blithe disregard for punctuation and the rules that make language accessible under the guise of a poetic license.

when you can communicate succinctly, functionally, to maximise your utility as a unit in a larger system;
ambiguity wastes time, enigma leads to error
and we are too busy These Days to deal with the magic of mystery we want answers not more questions, and quickly because we don't have time. Attention is a scarce resource of which poetry demands too much: there is none left, there cannot be, it is spent.

on an existence nourished by pleasure and profit, traversed by scrolling through a ceaselessly spinning wheel of commerce.
leaving behind poetry for those lucky few who accident upon it, or the even fewer who brave distraction and desire to submit to the words and fall into an intimate moment with the mind that composed them,
and pray to get lost amidst these words so that they can't find their way out, trapped amongst ideas
because they've found meaning
a different kind, perhaps, than functional communication can convey.

Source: Unsplash

AI & art

Avantika Mehra — Fri, 15 Apr 2022 04:05:17 GMT

This was first written in 2018; the release of Dall.e 2 inspired its revival.

Art exists for two entities — the Creator and the Recipient. For the former, it is the tangible result of various internal mental processes combining with an intent to convey them, coloured by emotion and experience. It is the physical manifestation of the contents of the artist’s cognition. For the latter, it is a medium of connecting with a truth. It is the ultimate cathartic experience, meant to fill the Recipient with the sentiment felt at the time of creation by simply exploring the piece. It is not bound by time, context, or a shared meaning of social structure or understanding of the world — it is simply felt, at some point, by some Observer who has voluntarily or involuntarily opened their mind and become Recipient.

Art need not be scientific or formulaic — it is feeling and emotion driven. Yet it is not without technique — an artist must conform to a certain level of shared meaning with the world, adapting forms, structures and shapes that create representations that can be understood by a wider audience. Yet, there is no constraint, limitation or universal law describing the extent to which the artist must conform to these frameworks. Unlike in science, math, or anything logical, art is individualistic and guided by the Artist’s free will. It is perhaps the freest form of expression humankind possesses — not limited by language, logic or medium.

While art may be without purpose, it is with intent. It is with the artist’s intent of conveying the barest form of truth; the present state of the soul; a portrayal of a version of his existence at the time of creation. Therefore, it is not simply the consequences of an individual’s actions at a time of a given mental state that constitute art — it is her intent to transform those into a vehicle of communication.

Source: Reddit

Machine learning refers to the idea of feeding computers with huge masses of data, and equipping them with algorithms to be able to sort, filter through and create representations of the information. Advanced image-generation software, then, once fed with enough pictures, paintings, or pieces of art, should be able to churn out its own version of art, using rules, finding patterns, extracting properties and making generalisations. But to what purpose?

Some could argue that the phenomenon of Art has wide-ranging uses and implications in the industrial world. It can be combined with logic, language, consumer-behaviour statistics and an understanding of the human mind to create successful advertisements and commercials. Art can be used with space to create sculptures, civic spaces and public areas. Art could be combined with faith, resulting in purposeful architectural wonders — chapels, cathedrals, mosques and temples.

A consequence of Art is that it can be used to make money, and it is not uncommon for economically and politically motivated parties to capitalize on a fleeting moment of human vulnerability and emotion to sell an experience. Thus was created the science of design.

Perhaps AI systems can churn out effective, captivating designs. Perhaps companies will no longer have to hire designers to create websites, posters and advertisements. Perhaps governments will no longer require creative personnel to craft political campaigns. After being fed with various paintings, styles, descriptions of sentiments and an understanding of space, AIs can use human-supplied tools, frameworks and information to create representations. But as they get smarter, richer, more creative over time, the human intervention and influence required to program, maintain, and monitor the robot will decrease. And one day, the robot will become smart enough to maintain itself. And some days later, it will program another machine — several other machines — to churn out designs as well.

The claim that robots are capable of design and not Art is idealistic — a comforting thought to hold on to as we contemplate the transformation of one of our most authentic forms of communication. Lamus is a computer that composes and produces music. A study conducted by researchers at the University of Malaga asked a group of 250 participants, including professional musicians, to distinguish between human-produced music and music composed by Lamus.

“The computer piece raises the same feelings and emotions as the human one, and participants can’t distinguish them”, researchers said.

AARON, an autonomous art-making program, has already sold paintings to the SF MOMA, the Tate Gallery, and the Brooklyn Museum. It paints not with pixels, but with paint on an actual canvas. Benjamin Grosser’s Interactive Robotic Painting Machine uses background audio, sounds of conversation and its own ‘internal mechanisms’ to create paintings. So AIs appear to be fulfilling the criteria of ‘evoking sentiment and feeling in the Recipients’ to be classified as Art. But do they do so with intent?

The current definition of intention indicates that it involves planning & forethought, which sound suspiciously like elements of cognition. Therefore, this suggests a machine must have the ability to think in order to create Art.

Can a machine think?

Humanity is not equipped to answer that question - yet. We don’t fully understand how the human mind works, so we don’t know exactly what ‘thinking’ entails. We may, through trial and error, and large datasets, be able to create robots that replicate and imitate human behaviour, a result of ‘thinking’. But if we can’t operationalize or define the term ‘thinking’, we can’t devise a test to measure it.

Turing said that the question ‘Can a machine think’ was too broad and “too meaningless to warrant discussion” - his Imitation Game instead measured how well machines could imitate humans. Any machine that could fool a judge into believing that they were communicating with a human would pass this ‘Turing Test’. But, Turing noted, simply demonstrating intelligence need not imply intelligence. Behaviourist measures are not sufficient indicators of cognitive activity.

Sidebar: It is worth noting here that conceptualising a measure for machine cognition is particularly challenging, since we cannot (yet) fathom its bounds, form, nature, or internal function. Technology designed to measure cognitive-activity in carbon-based humans (eg: EEG, MRI, etc.) is likely not transferrable to silicon-based machines. Many deny that intelligence can emerge from non-carbon-based entities at all — but who’s to say that all intelligences must be made of the same ‘stuff’? Species chauvinism seems to be the prevalent sentiment among humans. My personal belief is that the secret to understanding the blueprint of consciousness lies in the evolution of octopi.

What if we somehow invent a test to measure a machine’s intelligence, and find it to be cognitively analogous to that of a human, thus capable of expressing intent?

Then, perhaps, we can with certainty say that these ‘generally intelligent’ entities can produce Art. But it’s worth pondering — if they are creating Art, to whom does it belong — the machine, or the artist that created its inputs? Will these new entities also learn to cultivate emotion-driven intent, or will their art simply reflect the biases of their programmer? Will the emotions that are evoked from machine-generated art be those that their programmer intended to convey?

Or maybe we’re looking at this all wrong, too scientifically, too technically. Maybe the machine is the Artwork — a vehicle of communication, an extra layer of complexity added to convey the creator’s truth.

The earliest Homo sapiens drew on the walls of caves 42,000 years ago. Later, our ancestors discovered paper, canvas, sculpting materials and tools. Years later, graphic design was born. Evolution shows a pattern of discovery, innovation, and an ability of the human race to adapt and to adopt these new paradigms, changing the way we convey meaning and interact with the world at each stage of history.

Maybe embracing this new medium is simply humanity’s way of adapting to make the timeless, fluid concept of Art more relevant in tomorrow’s world.

100 questions for 2022

Avantika Mehra — Mon, 14 Feb 2022 06:24:41 GMT

Each December, I try to verbalise my thoughts into writing, and pen down 100 questions to capture my curiosities from the previous year. These range from questions about meaning and existence, to the future direction of technology; from reforming social systems, to re-evaluting personal values.

Some questions include working hypotheses, which will be refined through 2022; others, I hope to ponder and reflect on in the coming year. Behold:

On life & meaning

Is true permanence possible? Have we ever witnessed true permanence? Can sensations (eg: love) be permanent?
Is everything breakable?
Is anything truly irreversible?
How many lives do we live in our lifetime - when we dream, or live through the sound of music?
What if our base reality is music, and we’re just vibrations travelling along through the passage of time? How can we design our world so more people can hear the music?
Is rhythm a social construct, or a universal truth? If it is the latter, does it support the idea that math forms the basis of the universe?
If people derive purpose and meaning from work, how can humans continue to be creative after we have automated most of it away?
If we have unlimited time to focus on ‘what matters to us most’, are leisurely intellectual pursuits, volunteering, or cultivating new skills in a non-urgent, low-stakes environment enough to give life meaning?
If there are infinite versions of us, diverging at each possible life choice to live out and traverse different possible trajectories, is it possible to ever find the those versions of ourselves? Are they separated from us by space, time, mental strength, or something else?
We often think about advanced lifeforms existing in foreign space - parts of the universe we haven’t mapped out yet. Could there have been advanced lifeforms that existed in ‘foreign time’? Civilisations more advanced than ours that existed in the past, immortalised only in mythology?
Is our life’s goal to make the journey less unpredictable?

On identity & immortality

Is immortality a burden?
Is it possible for our species to continue existing in flux, or is either extinction or immortality a necessary outcome?
How long can identity persist over time?
Is identity a product of the collective world’s perception of who we are, or of our own perception? Are they any different?
What is the best way to think about identity? Some perspectives I’ve encountered over the years:
1. You are the sum total of everyone other than yourself - the world is made from mirrors, or filters through which light passes, and you are the resultant reflection; you enrich your existence by creating a more beautiful world
2. The world is just sensations; your body is the manifestation of the sensations you have claimed
3. We are approximations of our intentions
Is it possible that we’re all just 3D deepfakes generated by a Master GAN, part of a complex simulation administered by some parent version of ourselves?

On knowledge

What is the best way to identify where on the Dunning-Krugger line you would fall given your current knowledge on a given topic? Is it worth designing such a test for select knowledge areas, so people can have more epistemic awareness?
Is there a quantum basis to feeling like you simultaneously know everything but nothing at all?
When is it appropriate to stop refining theories based on new information, and regard them with a sense of finality or degree of conviction?
What are the bounds to certainty?
Is it because I am more confident in my beliefs about the world that I have fewer questions, or is this an indication that I am not reading enough?
Is it possible to measure intelligence in an artificial being, given that it may have possible unrealised but sophisticated cognitive capacities? (eg: situations where it is not being trained or tested on tasks it could be very good at)

On the future of tech

Will our gateway to the metaverse be text-based (eg: through a search engine), app-based (through individual portals), or voice-based (eg: through instructing digital assistants)?
Will social AIs with GAN-created image personas ever pose a legitimate threat to creator/influencer businesses?

On the effects of technology

The world is more fast-paced, instantaneous, connected than ever before - what are the effects of shortening the loops of micro-existence, the chapters in the book of life; of overloading brains with information, of hijacking attention, of instantaneous context-switching, between news of a kidnapping, a baby shower, genocide, a new job, in a matter of seconds as you scroll through a feed?
We get hypnotised by our phones on an hourly basis - the greatest trick of them all was convincing the people that black magic didn’t exist. Sleek devices to draw you in, till your eyes are swirling and you are hooked. Colorful attractions to draw you into different apps, and dark UX patterns to get you to stay. You’re scrolling, scrolling, scrolling - BAM. A distraction draws you away from the screen, you swipe up quickly, lock your phone, and throw it across the bed. You feel terrible. You were thinking about something before, now you can’t remember. What happened?
Should people be allowed to ‘block’ people in virtual worlds? Could this lead to intolerance?
How can technology be designed to make people more tolerant? Will people be incentivised to use such a technology?
To what extent should technology be designed to be egocentric?
Should users be allowed to opt for authenticity on the platform: exposure to a diversity of views displayed as they come in, over a carefully curated echo-chamber designed to hook the user? What is the best way to strike the balance between user experience and representativeness? Is it possible to remain competitive without showing users what they like to see?
What are other ways in which intolerance might manifest on digital platforms?
Is it possible to solve the problem of attention theft? As we transition into an increasingly digital existence, we are torn between existing in the real-world and the metaverse. Logging off for sanity, connectedness, peace of mind, seems almost irresponsible because of the personal and professional obligation to be reachable; because technology enables it and design simplifies it, capitalism demands it.
Is is possible to design a world where individuals can opt out of the metaverse? Are we morally obligated to do this? Can someone who opts out of living in the metaverse be successful?
If the barrier to expression and action is reduced in the metaverse vs. the real-world, could this lead to a new, more extreme, unfiltered, visceral, almost feral, form of communication and interaction?
According to the Sapir-Whorf hypothesis, language mediates thought. Does communicating through emojis, text, GIFs, memes, and abbreviations unlock new avenues of thought, creativity, and imagination? Or does their availability and ease of use reduce the complexity of thought, condition us through autotext to speak in similar ways, and restrict the boundaries of expression to a few, commonly spoken set of phrases?

On system design

How can we make air travel a less painful experience?
Is it productive to assign deterministic labels to people with mental health problems? Rather than saying you are XXX or you have YYY, could it be more helpful to reframe diagnoses to you are currently experiencing ZZZ in the cases where relevant, thereby framing the disease or challenge as conquerable? While a diagnosis is validating, reframing a conquerable problem as situational rather than dispositional could help the patient regain self-agency, and be instrumental in recovery.
How can we make emergency rooms and doctors’ waiting rooms a more comfortable experience for people?
How can we make them more safe? (Requiring a group of sick, vulnerable people to sit in a room for hours with other possibly infectious people seems intuitively problematic).
Why do we still use time? Is there a more efficient, precise way to catalogue our subjective experience?

On social systems

Is language and dialect homogeneity desirable? Why is it important to care about language diversity?
In the future, will wearing masks become a marker of belonging to a social or religious group?
Has the world begun to categorize and classify more? If yes, is there increased pressure to become part of an in-group or tribe in the metaverse?
Are humans getting more homogenous or archetypal? In terms of:
1. language: autotext and auto-email suggestions as language scaffolders, conditioning everyone to speak the same type of English
2. appearance: social media and influencer culture is another, more instant, personal, and pervasive avenue of spreading and establishing new beauty standards; AR filters impose a beauty ideal on your face.
3. politics: in the US, the social pressure to pick between and conform to one of two schools of thought (liberal vs. conservative) limits a citizen’s ability to effectively develop person moral convictions on intricate social issues. Citizens are shepherded into an ingroup based on their salient beliefs, and indoctrinated with the ‘correct’ beliefs of the group. Increasing polarisation between the two groups, and animosity towards the outgroup discourages individuals from questioning the group’s beliefs, engaging in discourse, or exploring alternate viewpoints, in fear of alienating (or being cancelled by) the group. What are some safe spaces that provide individuals with an unbiased platform to develop and explore personal, social, and political ideologies?
Why does sarcasm win, over kindness?
What is the significance of shapes in how we interact with the world? How might I classify, assign, or emotionally interact with an object differently if it changes shape?
Does the geometry of structures we interact with on a daily basis affect perception and action in a meaningful way? How do shapes of everyday objects affect how we interact with or think about them?
When travelling in Japan, I noticed I could identify specific geometrical structures as distinctly East Asian. In the case of manmade shapes, is design and architecture informed by land formations occurring in the surrounding natural environment?
Does the shape of an object have an impact on how people raised in the culture perceive or assign emotion to objects?
In Japanese culture, anthropomorphising objects is common - does this have anything to do with their shape?
How can we make web3 accessible to people who don’t have budgets to experiment with tokens or bet on NFT communities?
Are soulmates predetermined or created?
If soulmates are created, can anyone be your soulmate? If not, what determines the parameters for the subset that would qualify as potential soulmates? Is it possible to pre-empt or identify these parameters?

On personal and professional growth

How do you strike a balance between investing in the world vs. investing in yourself?
What is the best way to give back: time or money?
What % of time should I be thinking about giving back to maximise net impact?
How big of a role will ‘giving’ play in my life? (eg: career shift vs. part-time volunteering vs. donating or pledging a % of income)?
How can I create an environment to succeed at achieving my goals? And what’s the best way to think about what exactly I want to achieve?
Where will I end up settling down? Ideally, where do I see myself in 5 years? Should I know the answer to that, or have an emerging idea of where that could be? Can there be a right answer?
What’s one thing I want to change about the way I’m currently living? What are the things that are stopping me from making this change?
Many systems are dynamic equilibriums (eg: marketplaces, transfer of heat and energy, appetite for risk) - at the individual level, is it possible to achieve and maintain a static equilibrium, at an optimum point? External circumstances prompt changes in lifestyle choices; they ricochet back when external conditions revert. But the ricocheting becomes less over time as intentional life choices are made and implemented. Is complete equanimity achievable?
Could grad school be a good option for me? Is there a universe in which I want to go to grad school?
Big fish small pond, or small fish big pond? When is it the right time to move to a big pond?

On health & wellness

How can I increase the diversity of my microbiome?
What are the metabolic/energy costs to having a large head/big brain? Do these vary among humans?
How can I make my scalp healthier?
Can I focus on renewing my body at the cell level?
Given that ageing is necessarily the decline of the human body and mind, what are the best ways to slow down the process?
Why haven’t we solved for the common cold, flu, or sinus headaches?
Where can I find an indoor swimming pool in DC that is open on Sundays?
Does classical piano music affect the brain differently than other types of music?
Can humans overcome physical bodily damage with mental strength? To what degree do physical and mental strength enhance each other?

On inventing

Is is worth investing in / creating a personal ‘CRM’ to keep in touch with old friends - a machine-enabled solution to preserve human connection?
How useful could the following contraption be: a device that regulates thought flow, and caps the volume of sensory inputs that a person absorbs in a day. A simplistic version of this would regulate only the volume of thought flow, and indicate to a user when the target / maximum level of sensory inputs has been reached for the day, prompting the user to slow down, or rest, sleep, hydrate, or eat more than usual when exerting cognitively. A more complex version could distinguish between the types of inputs. Significant variation in inputs could indicate that a user is distracted and attention-switching, while the detection of a certain type of brainwaves could indicate a flow-state.
1. summary: a fitbit for the brain
2. functionality: users can analyse cognitive activity, speed of learning, attentional control, external stimuli intrusion, and productivity
3. applications: the device could integrate with a phone/smartwatch to disable notifications when a user enters a flow-state, sync with a smarthome to optimise conditions for productivity and learning, and become an in-house nutritionist for a user’s information diet.
What would you do if you had unlimited time and resources?
What’s one thing you would build if you had unlimited funding?
Can I invent something truly permanent?
If you believe that humans will one day be extinct, what’s one way you could leave a lasting mark on the universe as proof-of-existence for future life forms?

On prioritisation

What are the 5 most important problems facing the world right now?
What is the correct framework to use to assess which problems might be the most important?
Are problems that could threaten human existence necessarily most important to care about?

On ethics

Is violence ever justified? Under what conditions might violence be justified?
Is anger ever justified? Under what conditions is it morally permissible to use anger strategically?
Is intelligence the most relevant criterion when thinking about an entity’s moral status?
Is it possible to measure ‘suffering’ in an artificial being?
Does a human’s relationship with an artificial being affect its moral status (eg: if a human adopts an AI as a child, does this social relationship entitle an artificial being to full moral status)?
Is it reasonable to think that providing artificial beings with a moral status (or rights) could increase the likelihood of them becoming an existential risk to humanity?
Under what conditions is it permissible to lie?
Who should bear the risk in the case of an autonomous vehicle accident - the passenger, pedestrian, or bystanders?
Should the decision of which group to cause harm to change depending on the number of people in, or the demographic of, the group? (eg: should the car swerve into a bystander to save 4 passengers? Should it harm an old person to save a young child?)

On value

Do NFTs have intrinsic value? Does art have value? Is there a difference in how we assign value to NFTs and physical art?
Are virtual objects as valuable as real objects?
Assuming there is no significant difference in the lived subjective experiences you will have between two worlds, is there a difference to living in a real vs. a virtual world? Why might you prefer one over the other?
Is living in a real-world or a base reality somehow more authentic?
If yes, does the possibility that we are living in a simulation make your current life less valuable or authentic?

Misc.

How does quantum superposition work? Can mental states exist in a superposition, only collapsing into something definite when required?
Is it possible to switch on/off heuristics in social situations?
Is taste in food entirely subjective? Can tastes or flavours of food be classified as universally good or bad?
Is itching evolutionary? Is there a way to prune away or unlearn evolutionary reflexes that are no longer useful?

Image: Simon Cooper

the curious allure of dating apps

Avantika Mehra — Sun, 14 Nov 2021 00:35:32 GMT

Written in June 2020

If you live in the US, you are much likelier to meet your SO on a dating app than through friends, work, or at a public place.

As evidenced in the data, dating apps have demonstrated a level of effectiveness in matchmaking surpassing that of humans. Globally, popularity has skyrocketed over the last few years: as more people feel comfortable using online dating, the stigma around it goes down, the available pool of people increases, and the experience gets better over time, maximising a user’s odds for love.

In spite of their effectiveness, data suggests that dating apps leave users somewhat disappointed and unhappy with their overall experience: 45% of online daters report feeling frustrated after using the apps, while 56% of adults view dating apps as somewhat or very negative, citing exhaustion, burnout, and apathy.

Why apps disappoint

1) Overpromising, underdelivering

Apps market themselves as the go-to solution for when loneliness tugs; an inkling of hope is often enough to convert singles into users, especially in the face of limited other options. Being the most promising solution solving a significant user pain-point is desirable from a product marketing standpoint - it prompts quick user-conversions, with low acquisition costs, simply by virtue of existing.

Yet, net mental health effects seem to be tending towards negative. A potential explanation for this can be derived from the neuroscience of disappointment:

Outcomes > expectations ==> net positive affect
Outcomes = expectations ==> net neutral
Outcomes < expectations ==> net negative affect

When expectations don’t align with reality, the misalignment of experiences stimulates a pathway in the basal ganglia, causing a negative dopamine response - the result of which is disappointment that the promised reward was not received. In steeping user expectations with the promise of an ideal someone in exchange for time, data, and money, they often underdeliver.

The expectation/outcome framework also explains why the rush of excitement, butterflies, and sweaty palms are more likely when you match with someone in the real world, unexpectedly - a surprising encounter triggers a positive dopamine response.

2) Habituation

Dating apps become less exciting over time. After the novelty of the initial swipes wears off, users get used to the act of continuous swiping over new profiles, and repeated profile presentation incites slightly less excitement each time. Habituation may explain why long-time users are likely to be less enthusiastic about online dating, and more likely to quit because of boredom or burnout.

3) Abundance

With an abundance of users, dating apps create the illusion of choice - interactions are short-lived and high in frequency. A perception of limitless selection dilutes the value of a given interaction, so that exchanges are probationary and low-stakes. Dating becomes a game of trial-and-error as users shop around, breezing through profiles and ghosting liberally. Social costs are low, since users aren’t tied to the obligations that would accompany being a part of a common social group.

The intensity of emotional experience surrounding a typical exchange between a prospective couple is diminished with this systematic casualization of romantic interaction. Such a design breeds nonchalance, and the longtail of users who value early commitment are quickly left disappointed.

How they get you to stay anyway

Despite the negative sentiments associated with long-term use, an estimated 40 million people continue to use dating services regularly. 72% of millennials report having used Tinder, and the number of dating-app users between the ages 18 and 24 has tripled in the last two years. Some theories for why dating apps continue to remain popular include:

1) Usability & ubiquity

Stanford psychologist Fogg’s behaviour model describes the ‘six elements of usability’ of a product, app, or service: 1) time, 2) money, 3) physical effort, 4) brain cycles, 5) social-deviance, and 6) non-routineness.

Reducing each of these factors increases the usability, access, and feasibility of a product or service. Online dating has systematically diminished the barriers associated with traditional courtship, to make dating less time-intensive, expensive, and effortful, reducing the barriers of time, money, and physical effort (1, 2, 3). In addition, making an instinctive judgment on which way to swipe based on no more than a few pictures and lines of text ensures that the cognitive load associated with completing tasks is minimal (4). Attention-capturing triggers, like notifications and color, make it easy to integrate online dating into one’s common routine (5).

Notably, the lack of stigma associated with online dating in the recent years has led to an explosion of dating apps across countries and cultures (6). While in 2005, few Americans would agree with the statement ‘online dating is a good way to meet people’, 59% people agreed with this statement in 2013 - the number is almost certainly higher today.

2) Hooks

Like most consumer social apps, dating apps leverage hooks to get a user to open up the app. Some hooks are psychological: for example, enticing the user by the prospect of reward, caused by the subconscious positive emotional association created between opening up the app and experiencing a reward sensation. Others include attention-capturing triggers, like notifications, colorful icons, and zingy sounds, to redirect users to the app, and initiate a reward-loop.

3) Maximising addictiveness

You’ve opened the app, and created a profile; now they need to get you to stay. Cognitive rewards is one strategy used to maximise the likelihood of continued usage of the app, i.e. addictiveness. Cognitive rewards are useful in establishing a behaviour, such as app usage, and incentivising it to persist over time. Establishing a behaviour is achieved by what psychologists call a continuous reward schedule - that is, a user is rewarded when they interact with the service. When a user downloads the app, high-ranked profiles will be displayed at the very front – profiles that the app’s algorithm has scored highly using whatever weighted probabilities make up their secret sauce – providing a novel reward each time a user swipes to the next profile.

Addictiveness is increased when the reward becomes variable, or pops up only occasionally. Psychologists call linking user’s actions with variable rewards a variable ratio reward schedule – a user can keep swiping, but high-compatibility profiles are interspersed and scattered, providing an exciting reward to elicit an emotional response at irregular intervals.

User goals and heuristics mediate behaviour on apps

1) User goals

In motivated users, the brain assigns a high value to the behaviour of checking dating apps. Individuals who self-report that they are keen on finding a romantic partner show increased activity in the DLPFC region of the brain when using dating apps, an area associated with goal value. Research indicates that the DPLFC actually assigns a greater value to the behaviour of frequently checking dating apps, indicating that dating apps occupy neural real-estate in individuals who are highly motivated to find love.
Motivation is high when users expect positive outcomes. Consider a version of the expectancy-value theory:

degree of motivation = expectancy * value

The degree of motivation to use dating apps appears to be influenced in part by a user’s appraisal of their likelihood of getting a match.

Degree of motivation affects communication and behaviour. A user’s level of motivation has consequences for how they interact with people on the app: highly motivated users may approach online dating with a more optimistic attitude, making them more likely to take risks and ignore negative outcomes. Optimistic users also fall victim to the planning fallacy, believing overly optimistic forecasts of an outcome, like underestimating how long it would take to find an ideal match.

2) Heuristics

Apps are designed to provoke minimal cognitive processing and decrease the cognitive burden or work by the user; they are designed for simplicity, ease, and intuition. The minimalist design of profiles, with no more than a few pictures and lines of text, appeals to System 1 thinking - quick processing, followed by an instinctive assessment of the profile, and a knee-jerk response left or right.

But there is a risk in leaving this appraisal to the behest of System 1 thinking - it is characterised by biases, heuristics, and tendencies. Daniel Kahneman writes,

“The measure of success for System 1 is the coherence of the story it manages to create; the amount and quality on which the story is based is largely irrelevant.”

Often, a single profile amounts to the information users have about the person in question, forcing a swiper to make judgements and inferences based on the limited information available. Though there may be an absence of data critical to forming a holistic judgment, conviction in a match is mediated by the available data, leading to a distorted sense of confidence if the available data is promising. Confidence steeps expectations, invoking a negative dopamine response when the in-person interaction doesn’t live up.

With potential gains high, social costs low, psychological incentive strong, and adoption almost ubiquitous, dating apps are here to stay. It seems worthwhile, then, to think about how to best design them, so that we can shape kinder experiences for the future generations of dreamers.

what the future will look like

Avantika Mehra — Fri, 20 Aug 2021 05:44:53 GMT

The alarm rings — gentle, cascading sounds, increasing gradually in volume, to softly rouse you from your REM cycle. The sound radiates from micro-speakers distributed around your room; the 360° soundscape makes it sound like the walls are singing, beckoning you to take on a new day.

You wake up rejuvenated, and peel off the neuro-pellets from the side of your head, that quietly stimulated delta-waves in your brain through the stages of deepest sleep, accelerating cognitive regeneration while you explored a different world — one that hasn’t been touched by humans yet.

“Aliza, play today’s Robinhood Snacks podcast,” you tell your voice-enabled smart-home device, one that you built yourself in minutes from a DIY-kit you found online, because you know better than to install big-tech listening devices in your own home.

The podcast continues to play as you enter the bathroom — the audio syncs across the house, switching on and off as you enter each room to conserve energy. Speaking of energy, you wonder how much you saved yesterday. You click into an app, and a giant dashboard tells you you hit this month’s goals — “Huzzah! Check your texts for a special reward.” Amazing, they’ve sent you $10 in cryptocurrency for being a good energy citizen.

Your phone knows what to do with rewards — you’ve trained it well to manage personal finances. 10% goes to your favourite social cause — AI Safety Research. You volunteer with a local nonprofit that is working to keep humans and machines safe from each other by aligning their goals, and you’re their top contributor. 50% is distributed across your investment portfolio with predetermined weighting, and you use the remaining 40% to invest in metaverse assets — last time you bought a tiny bit of land in Decentraland-Iceland, and it’s looking beautiful, so you buy a few more square-feet. In September, the developers ship property-owners special VR-headsets that click into your neuro-pellets, creating the most dazzling multisensory Northern Lights experience, transporting you to the 2020s when the lights were scintillant and unclouded by pollution and space-dust.

The podcast ends, and you add a small tip for the creators before going to make breakfast — people who don’t tip creators are so rude, you think. The fridge displays the recommended breakfast items — based on your caloric history, the weather, and your previous meals — but you pick frozen waffles because it’s Friday and it’s been a long week. Quickly you project different outfits on ARfit, a closet assistant that shows you what you would look like in a range of outfit combinations — you only go into the office once a week, so you want to look good. Ugh. You forgot to order this week’s fit, so you call the nearest warehouse listed on lastminute.com and ask for expedited shipping on item-number JS823GJ. Ten minutes and a two toasted waffles later, it is at your doorstep. The vacuum-packing inside the delivery drone left the shirt creased… you shrug. Nothing a bit of steam-ironing can’t fix.

Your phone beeps, the calendar app reminding you to call a driverless car for pickup. But a beautiful summer day awaits, and you can’t think of a reason not to enjoy it. It is decided — you will walk to work.

You put on your AR-enabled glasses — they’re fancier than what you would usually wear, but that they were solar-powered and DuckDuckGo-enabled justified the purchase. You see a homeless man across the street, outside the Safeway, an AR QR-code hovering above his head. You surreptitiously scan it with your phone, and it takes you directly to checkout, with a meal in the cart, ready for delivery to his precise geolocation. Why not, you reason, completing the anonymous one-click checkout.

The next street is lined with shops of all kinds, selling everything imaginable — groceries, clothes, shoes, antiques, toys… toys! It’s your niece’s birthday tomorrow, and the little AR Cryptopuppy bobbing up and down in the window is adorable. “How much for the doggy in the window,” you chuckle, and click it to find out. “Enjoy the little blockchain-based collectible,” you write in the message at checkout, and schedule-send it to her for tomorrow.

As you continue your walk, the morning sun dances on the sidewalk in crescent-shadows, an omnipresent reminder of the massive coordinated effort between world superpowers to harness the energy of the sun to power the millions of servers and central-processing units that make the world go round. For a second it hits you — the scale at which humanity is operating, the audacity of the species to interface directly with the Sun, the technological prowess your generation has inherited from millennia of humans preparing for such conquests, and you are distracted by the wafting aroma of fresh bread.

“Enough capitalism for today”, you decide, as you click off your AR-glasses, and get ready for a new day of work.

on the simulation argument and virtual worlds

Avantika Mehra — Tue, 20 Jul 2021 22:29:36 GMT

The simulation argument

Over the last five years, technologists have begun to dogmatically accept the idea that we live in a simulation — the theory was popularized by Nick Bostrom, and discourse turned mainstream when Elon Musk declared that the odds that we are living in a base-reality is one in billions.

I am neither completely satisfied with the Bostromian account of a simulated reality, nor can I reject the idea of a simulated reality entirely. Below is an exploration into the different types of possible simulated realities:

(1) The ‘mind is a simulation’ possibility

Different religious scriptures converge on the idea that the mind is a simulated reality. While ancient scriptures may not serve as compelling support for/against an account of the nature of reality, they provide useful intuitions for understanding the essence of the world through time and space. According to Buddhist and Vedantic philosophies, a mind — consisting of tendencies, desires, and sensations — is a simulated reality. Our subjective experience is a product of simulations produced by neurobiological chemicals — we can escape our simulated reality to enter a base reality, and experience raw/authentic consciousness, by practicing bottom-up perception exercises (eg: breathing, meditation, controlling the direction of attention & subsequent flow of thought) to widen our sensory abilities and retain control over our perceptual experience.

In essence, resigning attentional control leaves an individual’s subjective experience to be controlled by the simulation — or the mind — while learning how to control it allows an individual to experience the authentic base-reality.

Simulations can be classified into two types: h-style and s-style. Conscious beings in h-style simulations are the result of tampering with neural hardware. Examples include the brain-in-vat example, Matrix-like situations, and the theory above. By contrast, s-style simulations are run on computer software alone — an example includes Bostrom’s theory that most humans alive today are actually sophisticated ‘ancestor-simulations’ run at some point in the future by our descendants. The theory below describes another s-style simulation: the idea that our entire universe is a complex simulation run by some parent universe.

(2) A more advanced species in a parent-universe could have generated consciousness in our simulated child-universe.

The most compelling support for this possibility would be if humans are able to instantiate consciousness in a machine, thereby demonstrating that deliberate consciousness-production is nomologically possible, and that similarly, we may be a simulated product of a more advanced species.

Consciousness can come in degrees (eg: a snail vs. a dog vs. a monkey vs. a human). Somehow, without determining the parameters for how to measure consciousness, we have arbitrarily classified entities as conscious (eg: spider) and non-conscious (eg: an autopoietic cell). If the complex organisation of different types of non-conscious cells can produce a conscious being, it is possible that a similarly complex network of silicon-based computers may produce consciousness (there is no reason to believe that consciousness must be restricted to carbon substrates).