I’m going to state my reflections as aphorisms or questions without an answer. I hope that simply offering how I personally feel after visiting SF provides some value. I don’t think convincing you my feelings are justified helps anyone here, although I hope these thoughtbits will soon coagulate into more substantial theses.

I see three real challenges in building a startup, or even living in general:

1. Forming first principles

2. Market selection

3. Acting authentically

On first principles:

• Remember the Mii channel on Nintendo Wii? Visiting SF from Sydney was like being dragged from a Mii Plaza where everyone is calmly walking around, to a different Plaza where there are 10x more Miis and everyone is running at full speed in all different directions, brushing past your shoulders, some in groups, some alone, but running nonetheless.

• There’s a lot of noise in SF because everyone is doing something and has an opinion on everything. My gut tells me a lot of them are LARPing — they’re incentivised to, because faking helps them sell to customers and VCs — and it’s hard to distil the signal from the noise.

• That’s why first principles are important — they help you find signal in the noise. Finding an idea that ticks every single box is like finding a needle in a haystack — it helps to have more hay, but it can be overwhelming if you don’t have any sieves.

• It’s exceptionally easy to get swept in a certain direction by a crowd, and there are so many crowds moving in different directions, so I found myself swept away multiple times.

• First principles help you stay focused so you’re not so readily swayed by trends and unfounded opinions. This is not only important because you want to be right, but also because ‘the returns to attention are exponential’.

  • Why are the returns to attention exponential? I tried to rationalise this by modelling the depth to which people explore a direction as a hazard model, where each person has a constant probability they will stop exploring at each step along the way (assume each step requires a single unit of attention to progress). Then, the distribution of people surviving at each step follows an exponential distribution. But if ‘returns’ is linear with the number of people you’re better than, this implies there are diminishing returns to attention, the exact opposite conclusion.

    

    Which implies returns must be more than exponential w.r.t. the number of people you outcompete. I still need to rationalise how that could be the case — this is an unfinished thought.

  • I feel like my mind has been poisoned with restlessness and undirected intrinsic curiosity. I used to be obsessed with things like economics — writing articles and building models. Now, I feel my attention is more sparse, I multitask more, there’s less deep work, less craft. Was it trading? Did the hyperpragmatism kill the fanatic inside me? If so, that’s what they mean when they say that working in trading kills your soul.

• Intelligence is a prerequisite but far from sufficient. It also has very quickly diminishing returns. Agency, resourcefulness, and focus are more sustainable and effective moats. First principles allow you to have conviction in your opinions, and hence focus.

• My gut tells me the AI space is some bubble and that we are due for an imminent correction. There are standard hallmarks of historical bubbles: high autocorrelation (due to circular deals between firms), herd mentality, cheap capital, high leverage. The system feels fragile to me — a sufficient perturbation just due to variance x time could cause the system to cave in substantially. I need to work on having better first principles to build the conviction to short the AI sector, or buy the spread between the large incumbents and the countless AI startups being founded daily.

On market selection:

• In trading, we consider trading PnL to be the product of 3 components: market opportunity, % of opportunity you capture, % captured opportunity you convert into PnL. The same applies to startups.

• The first component, market selection, provides the highest leverage impact on your PnL. This is because market opportunity is [0, inf), while % capture and % margin are both [0, 100]. Additionally, and perhaps more importantly, market opportunity is the only thing that isn’t zero-sum, and there aren’t forces actively trying to minimise it — in fact, the same firms competing to minimise your % capture and % margin are those trying to increase the market size for you.

• The CEO and CTO of Decagon, for example, toiled for 4 years on startups that didn’t go far, before trying various ideas and landing on Decagon when finding high demand in customer support. Decagon reached $100M ARR just this month. It was founded 3 years ago.

• This is why many successful startups are often simple. Validation and execution in a long and deep market wins. E.g., Decagon, Paraform. Building is extremely cheap and easy now with AI. It is much harder to make that a moat.

On authenticity:

• Being authentic means being honest about your utility function and acting accordingly. I think the honesty part is non-trivial. There is a lot of societal pressure for your utility function to be a certain way — ambitious, altruistic, consistent. It’s very easy to lie to yourself that your utility function is closer to the social ideal than it really is, that you indeed have grand ambitions, care plenty for the wellbeing of others, and have unwavering resolve. But it’s ok to be a bit lazy, to be selfish, to be discovering your own utility function bit by bit just as you are discovering the world.

• I had this thought because I realised how much the US embodies mercenary hustle culture — SF is no exception. It is not a place of imagination or dreams. Most people are there to make money, be mercenaries, not missionaries. Your job is to make money, you must love it, and be willing to work hard for it. The US is the epitome of Darwinian capitalism.

• To be clear, both missionaries and mercenaries exist, just not in the proportions I expected. I had idealistic expectations of SF, that’s on me, not the city.

• Making this realisation made me question whether I wanted to enter the space at all. I left quant because I was uninspired by mercenary culture. I don’t want to join another mercenary-dense environment.

  • Perhaps this is a false dichotomy, and the real takeaway is that I must actively seek those that inspire me. As with anywhere in the world, not everyone has similar utility functions to me, not everyone will be my people. It remains just as important to actively seek my people as it would in any other place in the world. Having a place of emotional safety is important for mental health and especially necessary when 16+ hours of your days are occupied by hard work.

  • Why do I despise mercenary culture so much? What is it about fierce competition that I dislike? I think it’s mostly because I subconsciously assume a negative correlation between being a missionary and being competitive. But also, is it because I’m not confident that I’ll win? Am I a weak man created by good times?

    • I think it’s partly because I don’t perform best in competitive situations. It makes me more results-oriented than I should. It makes me think less and do more for fear of falling behind. That’s not good.

  • Is Darwinian capitalism good for society in the same way Darwinian evolution is biologically good for species? I don’t think so. In Darwinian evolution, genes carry the traits of success so that the next generation inherits them. In Darwinian capitalism, inheritance carries the outcomes of success, i.e. capital, whereas traits are not necessarily inherited at all. Additionally, the feedback loops that entrench the incumbents are far more severe in capitalism than in biology, leading to local optima.

• I’ve met and heard stories of people who have toiled for years without significant progress. There are a lot of hidden ‘failures’ (i.e., lack of success) in SF. It makes it all the more important to find something you are intrinsically motivated by, whether it’s the mission or the money.

• Therefore, I’ve learned to focus less on what I ‘should’ do, and just do more of what I want to do. This is especially helpful since the more scarce quality required to succeed is much less intelligence than it is tenacity.

• I like solving intellectual problems, I like creating intellectual art. I’m not good at making money yet.

On a lighter note, I found my new favourite place for food in the US: Mendocino Farms!

It works brilliantly. The brightest individuals, most confident in their answers, take the exam many times, moving the average towards their more accurate answer. Meanwhile, the least confident choose not to participate for fear of the penalty, wielding no influence over the machine’s output. In the end, the machine’s answer is the most accurate possible, better than any individual human could ever provide.

The prediction market is the incarnation of collective human knowledge. Or at least it could be.

Instead, we’ve turned it into a casino. The front pages of Polymarket and Kalshi are choking with sports, trivialities, and dopamine hits on demand — just look at the 5-minute Bitcoin Up or Down markets. You can compete on the leaderboards, then engage in the comment sections, where users call each other “uncs” and “retards”. This is no platform for people seeking truth. This is a platform for people looking to gamble.

And gamble they do. Global prediction market trading volume surged more than 400% from 2024 to 2025, reaching nearly $64 billion, up from under $16 billion the prior year. On Polymarket, roughly 70% of the ~1.7 million trading addresses recorded realised losses, mirroring loss rates seen in CFD/forex markets. On Kalshi, takers lost almost 32% on average while makers had an average loss of about 10%. And just as the odds set at casinos are never the fair value, PolyKalshi's ‘predictions’ exhibit a consistent overestimation of probabilities, the same inefficiencies plaguing retail-heavy options markets.

In some ways, this is unsurprising. The goal was never to seek truth. Financial markets were created to facilitate trade. The theory of efficient markets only arrived nearly four centuries after the world’s first stock exchange in Amsterdam. And exchanges’ incentives perpetuate this — most of their revenue comes from trading fees, which are a function of trading volume, not truth. For example, the CME Group’s clearing and transaction fees accounted for 81% of their revenue in 2024.

We are toying with the most powerful epistemological device ever created. But we should not lock away the key to the universe simply because we can’t be trusted with it. Instead, we need to learn to use it as it was intended.

The machine’s accuracy rests on one assumption: the smartest are the most confident and take the most exams. This relationship simply doesn’t hold in financial markets. In Indian equity derivatives, for example, more than 75% of those who took losses in the two consecutive years prior to FY24 persisted with trading in the same products — 91.7% of those people lost again in FY24. Their flow doesn’t just fail to facilitate truth-seeking; it actively disrupts it, pushing prices away from fair value.

To redesign prediction markets for truth-seeking, not trading volumes, three things must change.

First, the loudest voices must be the most accurate ones. In most markets, this relationship is sound: market makers distinguish informed from uninformed flow, and move prices more when trading against informed flow. However, in retail-heavy markets, such as prediction markets in their current form, uninformed flow outweighs the risk appetite and position limits of market makers, forcing market makers to reflect uninformed opinions.

One way to avoid this could be to enforce rolling performance-based position limits — the more accurate you are, the more you are allowed to exert your influence on the markets. The definition of accuracy must not only incorporate PnL, but also be normalised by trading volume and number of markets traded, to account for the effective sample size of the PnL calculation — we don’t want to give a lucky gambler massive position limits. One robust metric: the probability that an individual could have achieved returns at least as good as what they actually did, by pure chance. The lower it is, the more likely they are genuinely informed.

Second, winners must be made by merit, not manipulation. We must fight the market manipulation that plagues Polymarket and Kalshi. This is not only an ethical point, but a practical one — pricing in a prediction market is only a good predictor when the flow behind that price has truth-seeking intent. The truth-seeking exchange must conduct extensive market surveillance to crack down quickly and punitively on bad actors who game the system rather than express genuine opinions.

Third, incentives must align with truth. For the exchange, the key revenue source must be replaced with a fee correlated with market efficiency, not trading volumes. The exchange must also answer to regulators that serve the individuals and institutions that participate in it, to prevent it from devolving into another game. And for the participants, when there are substantially lower retail volumes, market makers must be compensated another way, potentially by rebates funded by philanthropists, governments, and institutions that benefit from this truth-seeking process as a public good.

With these three structural reforms, the machine is properly reassembled. Now, we must fuel it. Local regulators must find common ground to form a global coalition, enabling markets to be interlisted across borders. More informed players mean more accurate markets.

Finally, we can use the key to the universe as it was intended. Instead of asking questions like ‘Will Clavicular be Iran’s next Supreme Leader by March 31st?’, we could ask questions that actually inform decisions on government spending, regulation, and capital allocation, questions such as:

These are the questions the machine was built to answer. It's time we started asking them.

[W]e mostly just do things. We act from habit; we act from impulse or convenience when primed by the activities in front of us; we remember our goal and choose an action that feels associated with our goal. We do any number of things. But we do not systematically choose the narrow sets of actions that would effectively optimize for our claimed goals, or for any other goals.

Even this I feel undercuts the gravity of the problem, which I think is better reframed as not being ‘meta’. Because once you generalise the core issue into the framework I’m going to sketch out, you realise that the meta arguably accounts for over 50% of the levers affecting your life. That is why it’s a disaster that we all suck at it. After all, I’ve been called ‘too meta’ by the same person on two separate occasions, which I think says less about me than society’s generally low benchmark for the meta and dismissal of the meta as something worth scrutinising.

Let’s start from the least sophisticated human and work up towards the omni-optimising ‘meta-man’.

Nearly all of us choose, to a large extent, what to do with every moment of our lives. The least sophisticated of us make these choices based on habit, impulses, convenience, peer pressure, etc. Most of us, however, make these choices because we consciously believe that the choices we have made will, in one way or another, help us achieve our goals. We choose to pick a chicken salad over a cheeseburger because we believe it will help us achieve our New Year’s resolution of cutting our weight by 15kg. We choose to go to the speed dating event today because we believe it will help us find a girlfriend. We choose to accept the higher-paying job over the lower-paying job because we believe it will help us make more money.1

Many of us continue further down this line of reasoning and link each of these goals to an even greater one — we aim to cut down our weight by 15kg because we believe it will help us live to 100, we aim to get a girlfriend because we believe it will help us raise a family, we aim to make more money because we believe it will help us buy whatever we want. And a few of us reason all the way to the end, linking each goal to our very own utility functions — we aim to live to 100 because we believe our utility increases with our time alive, we aim to raise a family because we believe tight familial bonds will be significantly net utility positive, we aim to be able to buy whatever we want because we believe our utility increases with the amount/quality/prestige of the things we own.

The meta-man not only does all of this, but also:

• Clearly defines each goal and what its completion entails.

• Understands exactly how his utility function works, including every partial derivative w.r.t. each input and also what all the inputs actually are, while also systematically updating his understanding of his utility function as it is dynamic.

• Systematically interrogates his beliefs on why his choices should lead to his goals, and why his goals should lead to utility.

• Scrutinises, based on his beliefs on how his choices lead to his goals and how his goals lead to utility, between different possible choices and different possible goals, systematically seeking the optimal combination of choices to achieve his goals and the optimal combination of goals to achieve utility.

This describes the general framework of life optimisation as a tree where:

• The nodes are different states of the universe. We control the states that we can (our choices) to reach a different state of the universe (our goals) so that we eventually reach the state in the universe where we are ‘happy’ (utility-rich).

• The branches are the dynamics of the universe that allow one state to lead to another, a black box whose inputs and outputs are states of the universe. Our beliefs on how our choices lead to our goals and how our goals lead to utility are merely assumptions about these dynamics of the universe.

The meta-man knows exactly how each branch transforms states of the universe into another and makes the choices that exert the highest leverage (i.e., requires the fewest amount of resources such as time, energy, or money) on his utility.

To be ‘meta’, then, is to focus as much on our assumptions about the black boxes of the tree (few people do this) as we do on our inputs into them conditional on our assumptions about them (most people do this).

Examples of meta activities include:

• Regularly journaling to both better understand your utility function (i.e., the black box taking states of the universe as an input and a certain amount of utility as an output) and alter your utility function to provide more utility on average (e.g., through practising gratitude).

• Regularly seeking new experiences to better understand the black box that is your utility function.

• Systematically experimenting with different choices to attain your chosen goals, documenting the results of each experiment to Bayesian update your understanding of the black box that can produce that goal.

• Systematically seeking and engaging with people with similar trees to you to understand the higher level branches that necessarily require a prior judgements because you can only experiment with them at most once or twice in your lifetime (for example, there’s a limit to the amount of careers you can try within the most productive years of your life).

• Systematically reviewing your understanding of your tree and deciding whether your current actions are truly aligned with your tree.

So, why aren’t we automatically meta? A few hypotheses:

• The most fundamental reason: it’s hard! It not only takes significant time and effort, but also emotional pain to confront the truth.

• It’s social anxiety. An overwhelming proportion of the meta activities that broaden your understanding of how each black box works and the possible inputs into each black box involve talking to people you aren’t familiar with. I, for one, am fully aware of this, yet still need to hold my breath before forcing myself to commit to such social activities.

• Ignorance is bliss. People may not want to know the truth, because it’ll hurt too much when they realise they’ve been misaligned for the last 5, or even 20, years. It’s also a grand manifestation of sunk cost fallacy.

• Several people have called me ‘privileged’ to be able to think about meta problems instead of more urgent ones on finances, health, or family that many people grapple with daily. In other words, meta thinking is nearing the top of Maslow’s hierarchy and is simply not a concern until you’ve fulfilled all lower levels first. I don’t fully agree with this. I don’t think meta thinking is about thinking about self-attainment. I think it’s about whether you think about the hierarchy at all.

• Getting tunnel-visioned during execution mode!

• The modal outcome of a meta activity is often 'failure’, or a lack of observed benefit, so we are not rewarded sufficiently frequently for a positive feedback loop. As AnnaSalamon puts it:

[…] the main problem is that most courses of action are extremely ineffective, and that there has been no strong evolutionary or cultural force sufficient to focus us on the very narrow behavior patterns that would actually be effective.

• Even when we do reap rewards from a meta activity, it’s easy to associate the outcome with something far removed from the positive EV fat positive tail decision that actually caused the outcome. A somewhat silly example: I have a kinda narcissistic habit of speaking louder when discussing something potentially interesting/original/controversial with my friends, such that people nearby can overhear the conversation. I’ve done this in many different settings, but the only occasion where it led to a tangible positive outcome was when I was chatting with a friend about technological progress in China, in a Westfield food court, which landed us a savvy internship during our first year of uni. Surprisingly, many people’s conclusion from hearing this is to frequent that particular food court, rather than express opinions confidently, or even better, read books!

• Due to all the reasons above, most people don’t think, even less talk, about the meta that often. That negative peer pressure makes it even harder to be meta yourself.

Finally, here are a few ideas on how we can combat these factors and be more automatically strategic:

• Reduce the time and effort required to be more strategic. This involves finding suitable tools and methods to interpret your life holistically, so you can systematically align your choices to your utility function. Participate in online instead of offline discussions to overcome social anxiety, or bring a friend. Create a crystal clear routine for yourself to reduce decision paralysis.

• Delegate your decision-making. I imagine that if I were able to control myself as a Sims character, I’d be spending my time much more optimally than I am now, since emotional biases like social anxiety don’t really exist in the Sims world. One way to do this would be to commit to something sufficiently far ahead of time such that the emotional biases aren’t strong enough yet. Or, you could commit to share weekly journals with a friend, leveraging the social contract.

• Continuing on the Sims analogy, part of the reason why it’s so easy to optimise a Sim’s life is because the utility function is unambiguously quantified (e.g., the status bars). So, to the extent that we can, we should strive to quantify our utility functions. Use habit trackers, sleep trackers, biomarkers, screen time, etc. Evaluate decisions quantitatively using your framework of choice (e.g., EV calculations, goal factoring, etc.).

While most people optimise by default, those who optimise the optimiser—the meta-man—wield disproportionate influence over their own lives. Being meta isn’t just abstract philosophising. It’s a lever on the lever.

Several weeks ago, I was still in dread mode. Not only was it the first time I had taken several weeks of vacation in a couple of years, but also a good portion of it was spent driving the quiet roads of New Zealand with my family, and let me tell you, we aren’t a talkative bunch.

When you have both the temporal and mental space to sit with your thoughts undistracted, you start to zoom out. If twenty years from now I was to observe where I currently am, what stage of my life would I consider today to be? How does where I am today fit into the billowing tapestry of butterfly effects that I will one day look upon as my life? These were the questions I tortured myself with over the past few months, scared shitless in the face of my own free will as much as a nighttime sailor would shudder upon an uninvited awareness of the infinite abyss below and around him.

You know those random existential crises you have in the middle of the night when you can’t fall asleep? Yeah, it’s like that. You get too much time with your thoughts, so they naturally drift towards questions of your survival.

Glad to say, I think I’m about 90% out of it. Partially because I’ve genuinely managed to untangle the knots in the threads pulling me left and right, partially because returning to working 10-11 hours a day certainly does not provide anything close to temporal and mental space.

But whether it’s the natural next step after figuring 90% of my shit out, or a function of my temporal and mental constraints, I am in execution mode now. And maybe for the better.

For the most part, I’ve boxed away most of my dreading self, and have started executing on the plans my dreading self made. Less thinking, more doing. Forget plans, focus on results. I’m clicking back into a schedule that I believe maximises my lifetime utility. My frustrations are no longer towards my misdirection, but instead towards my mediocrity.

One thing I’ve come to realise while returning to my executing self is something that two of my friends have previously brought to my attention. In his comment on my previous post (sorry I haven’t gotten around to responding to your thoughtful comment, please consider this post as part of my reply), pointed out that the focus shouldn’t be to avoid competition, but rather to find your 10x edge. This is especially the case because, as my other good friend said as he cited Cal Newport’s book, passions are often built on success, so doing something you are good at may begin a virtuous cycle of passion fuelling results fuelling passion. Looking in the mirror again, this is probably more so the case for myself than many other people, considering my constant lament toward my lack of passion. It’s not that I’m not passionate. I’m just not good enough at anything yet.

Well, I’m basically being taught daily how to make money, i.e., influence, which is an incredibly powerful asset, considering how much the causes I care about need it to survive (as I’ve recently learned after attending a neurotech talk on BCIs). This has been my first and only full-time job, so perhaps this is where my 10x niche could be found. Especially considering I feel quite different from all the other traders, I take it as a positive signal that I’m doing something differently and can at least increase the chances I become the 10x, even if it means I succumb to a bimodal distribution of outcomes.

In the end, I’m happy with how the last few months have panned out in my mind. This seesawing between dread mode and execution mode has helped me take some well-reasoned and bigger-than-usual steps in my SGD. So if you miss my dreading self, don’t worry. He’ll probably be back in another few months.

Part I: Your life in a nutshell

I think all of life is just stochastic gradient descent.

Crippled by imperfect information, cognitive biases, and blindness to the future, we nevertheless stumble on, armed only by an unwavering will to live a life worth living.

Each step is a decision with some prior probabilistic distribution (a ‘decision distribution’). We don’t know for sure whether this decision will improve or worsen our outcomes, and we will only find out after the decision has been made. Generally, we’ll only make decisions that have a positive mean: one that gets us closer to that minimum we’re descending towards.

Your overall outcome in life (I like to measure it in lifetime cumulative utility, but the choice is yours) is where you end up on this optimisation problem after taking all the steps you can till death. Your life outcome is the sum of all the decisions you make. Therefore, it similarly has a ‘life distribution’.

The more time you have left in your life, the more decisions you can make, and therefore the wider your life distribution becomes. Additionally, assuming your decisions are on average positive EV, the mean of your life distribution is higher the longer you have left to live.

For my fellow options traders, you can almost think of your life distribution analogously to that of a stock: variance increases with time and volatility (variance of decisions), while its mean increases with time and the interest rate (quality of decisions).

Part II: Tirades against myself

To motivate this section, I’m going to share something that’s previously gotten me relentlessly berated for “betraying my identity” and “forgetting my culture”.

As much as I hate to admit it, a small part of me resents my Asian upbringing.

When I migrated to Australia at 4 years old, my birth certificate had my lunar calendar birthday, which made me just old enough to start kindergarten as soon as I arrived in Sydney. I worked my way to an OC class, after which I attended a selective high school. Outside school, I went to piano classes and coaching college. I started uni studying law, albeit by my own volition, but undeniably a corollary of the values I was raised with, just like my choice to start a career in options trading. A glaring exception was my attending one of those ‘elite’ private schools in Sydney, where I was one of the 6 Asians in a cohort of over 250, but that mostly just made me even more self-conscious about my Asian heritage (read ‘internalised racism’).

In my mind, this pattern of stepping through a series of well-trodden and socially laudable milestones that is so typical of an Asian upbringing does significant harm to your life distribution. A culture of scarcity driven by decades of rampant poverty has amplified our parents’ risk aversion, while thousands of years of social status being locked behind the doors of government — you can knock on those doors with flawless grades — has made the desirable decision distribution one that is (m)ean-(in)creasing but (var)iance-(d)ecreasing.

Hence follows a narrow life distribution, where tail positives rarely eventuate. Look around — how many founders and CEOs of global companies, top scientists at the frontier of their field, or renowned mathematicians were ones raised by MINVARD choices?

As a matter of fact, I think it’s highly possible that MINVARD choices lead to MEDVARD life distributions, at least relative to MINVIN choices, for several reasons:

1. MIN: Instantaneously mean-increasing steps commits you to a particular local minimum. The only way to find an even better minimum is by taking some mean-decreasing steps first. And even if you fail to find a better minimum, you can always follow your discovered path back to the local minimum you’ve already found. Besides, there are too many minima in the universe to probabilistically believe that the few you’ve found contains the global minimum.

   

2. VARD: Most MINVARD choices derive their low variance from the comparitively large amount of information you have on the nature of that choice — familiar case studies (friends/family who’ve walked that path), formal processes (hierarchies, applications, ladders), and clearly defined outcomes (compensation structures). There are very few known unknowns here, let alone unknown unknowns that could conjure those highly positive tail events. Again, you’re edging closer towards a local minimum, the speed of your descent diminishing as much as your potential is.

3. VARD: Dramatic actions, or big steps, can help you avoid local minima. Low variance small steps only get you towards the minimum you can see. The path to the global minimum cannot be seen, but it lies, waiting for the bold.

Our tendency to mistake MINVARD choices as optimal is simply the symptom of ingrained congitive biases we all possess:

• Risk aversion not only discourages MINVIN choices relative to MINVARD ones, it also leads to sacrificing high EV high variance plays for low EV low variance plays.

  • Again on the topic of the archetypal Asian upbrining, tiger parenting (fortunately I was not tiger-parented) destroys self-confidence, which instills risk aversion onto the child.

• Myopia, or tunnel-vision, causes us to fixate on the local minimum immediately before us, blinding ourselves from the countless better minima in existence.

  • On well-trodden paths exists competition, and competition is for losers. Sometimes we get so stuck in the adrenaline of this competition that we end up mistaking the battle for the war.

• Risk aversion and myopia are symbiotic to each other, making them doubly parasitic to us. Since we can most clearly see the immediate future, risk aversion encourages us to only optimise for short-term gains. The most significant opportunities exist in the extreme long-term where outcomes are highly uncertain.

Part III: Onwards!

I think the most powerful antidote to cognitive biases is to experience first-hand the benefits you can receive by opposing them.

The best part about many of the actions that counter these cognitive biases is that they have very limited downside and very large potential upsides.

Cheap call options, if you will.

Against risk aversion: Being authentic.

Authenticity means unforgivingly acting in accordance to your utility function, broadcasting your utility function, and making every effort to refine your understanding of your utility function.

The reflexive effect of authenticity — you act according to your interests, you get better at your interests, you love and pursue your interests more — skews your outcomes in a direction that will provide you maximal utility.

Some authentic actions:

• Self-publicising: Tell the world who you truly are! Several of my friends have on occasion said there is only downside to sharing a negative opinion or a sensitive truth (you know who you are), but I disagree. There is great value in revealing your values, your gripes, your intentions. You could be proven wrong and become a better person for it. You could earn yourself a hater, even someone who could do some damage. Or you could find someone you truly resonate with, someone who will amplify you towards the best version of yourself.
  
  I would gladly accept a hundred haters for a true friend like this, every single time. This is a big reason why I started this blog, and why I believe you should too. Since I’ve started writing, I’ve learned much about the people around me, either from their response to my writing or the lack thereof. I appreciate every discussion and thought; thank you to those who read and care.
  
  Don’t produce content. Content is designed to be seen. We want to be understood. The more visceral your writing, the stronger reponses you receive, the clearer you see the people around you. Write for yourself, whisper your deepest secrets onto the canvas before you.

• Sit in silence: On your next commute, your next drive, or your next bathroom break, don’t do anything except sit, drive, or shit. Free yourself from all distractions so you can create a personal echo chamber that amplifies your innermost thoughts. The deep thoughts that take effort, time, and space to reveal themselves are often the ones that bring the most clarity as to what your utility function truly looks like. So many of my personal dilemmas have been solved by just understanding myself a little better.
  
  Self-publicising is a corollary of this. Be an open book, literally. No idea is ever complete until it is put into words — the same applies for ideas about yourself.

  
  Also, meditate.

• 别管江湖龙虎斗 / Don’t interrupt the dragon-tiger fight: To your left and right are countless fierce battlegrounds — for money, sex, or fame — where, supposedly, glory is won, honour is achieved, and selves are actualised. Why are you competing? Are you so stupid as to need another person or institution to tell you what you want? These battlegrounds are as much a crude standardisation of utility functions as public education is of the path to wisdom.

  
  Understand yourself. Trudge deliberately and persistently towards the beauty your heart truly desires. Remain focused. If you happen to pass through a few battlegrounds to get where you want, then sobeit. But more often than not, on your journey towards your magnum opus, you’ll likely find yourself in the wildnerness, far away from the distant rumbles and war cries.
  
  I cannot stress this enough. Competition is for losers.

Against risk aversion: Seek the unknowns.

Progress lies in the unknowns of the landscape on which we’re descending. While we have many more steps to take — and we will only have increasingly more steps to take — expanding the number of minima in our field of view remains a highly positive EV bet.

Crucially, people always talk about unknown unknowns. People less often discuss unknown knowns. I think unknown knowns are equally important as unknown unknowns, if not more (especially if you don’t spend much time self-reflecting). This is because, arguably, known knowns are more useful to us than known unknowns. And once discovered, an unknown known immediately becomes a known known, while an unknown unknown becomes a known unknown (what a mouthful!). Therefore, finding new minima in an authentic way is one of the best ways to finetune your search, as those truths we’ve always secretly known are only discovered when we view ourselves from a different angle.

On discovering new minima:

• Meet new people: In hindsight, close to none of my ideas can be considered original (e.g.). Every single one is a Frankenstein’s Monster wading out the pool of knowledge and ideas I’ve learned from others. So meet new people to expand your pool, making unknowns known. And talk to them about things that occupy your mind! Rubberducking in this way will help you rediscover what you’ve always known to be true.

• Read! Books are the cheapest call options I know. Be pragmatic and authentic when reading. Read what you’re interested in — no more, no less. I find books are the best way to discover unknown knowns, since the sentences that strike the deepest chords are often those that perfectly describe an array of your past experiences into the neatest little mental model. I’ll make a post on how I read and my favourite books soon.

• Take a break: Basically the same explanation as that of sitting in silence. I’ve never really understood burnout and how taking a vacation could help treat it. But I think it manifests in different forms. For me, it manifested in the form of tunnel vision: I get too stuck into my work and all I can think about is how to optimise my current direction. Taking my first multi-week break in two years helped me distance myself from business a bit and gain a higher-level perspective on where my life is at right now.

Against myopia: Mastering time.

The core issue with cognitive biases is that what’s stopping you from making optimal decisions is not a lack of incentives, but rather psychological bugs that drive you towards suboptimal decisions or blind you from your own utility function.

Indeed, we are already heavily incentivised to consider our future accumulation of utility. But nonetheless, we’re driven towards self-destructive behaviours (I, for one, love my sugar and carbs), either by flippance, or temporary ignorance.

I think the best ways to tackle myopia is to further increase our incentives to pay attention to the future, and increase the amplitude of our utility function over time:

• Don’t die (lifespan): Live longer and you’ll be given more time to confront the consequences of your actions today.

• Be healthy (healthspan): Live healthier and you’ll have more capacity to do what you want in the future.

• Pay attention (sensespan): Experience more and you’ll multiply the effect of your actions on your utility. Touch grass. Dopamine is the biggest perpetuator of myopia — avoid it at all costs.

Part IV: A life worth living

What’s all of this for? To improve your life distribution so you can achieve a life outcome worth having lived for, one where:

• Your pursuit of the unknown creates rich and diverse life experiences.

• Your mastery of time makes your passage through time utility-dense and unforgettable.

• Your loud and radiant authenticity allows you to find (1) your true community, and (2) true love, someone who is an extension of yourself, someone who makes the pains and joys of existence worthwhile.

• You have found and are fulfilling your purpose.

We’re all in one of those massive mining holes. Each of us, the team members, is an excavator. We’re all a little different. Some of us work faster in soil, others in rock. Some of us have a small shovel that can dig up small quantities of gold fast, others have large ones that take a while to dig anything up but it’s usually something big. Some of us may just be built better than the rest — faster, tougher, bigger. But crucially, each of us can be upgraded using the very gold that we each mine.

We mine gold the intuitive way: just digging our shovels into the ground and bringing soil, rock, and bits of gold back up. But, to cash in the gold we mine, we have to put the gold into a magic bucket each of us has.

Our team’s goal is to maximise the sum of the gold in all of the team member’s buckets.

Each bucket has 2 openings:

1. Input opening: Gold can only be put into the bucket through this opening.

2. Output opening: Gold can only be taken out of the bucket through this opening.

Now here’s the magic of this bucket. Once gold has been put into a bucket through its input opening, it can indefinitely clone itself such that if anyone tries to take out a piece of gold from a bucket, they will receive an exact copy so that there are now two of those pieces of gold in existence. This cloning effect applies to unique pieces of gold, such that you can’t get endless copies of a single piece of gold by passing it through the same bucket over and over again.

Why this is important is because if you can take a piece of gold out an output opening, not only can that one piece of gold in your bucket be used to make you, the excavator, even better, but also it can be put through the input opening of other team members’ buckets. Therefore, there is a compounding effect — one piece of gold can turn into multiple, which can in turn improve everyone’s excavating skills, allowing more gold to be mined.

The trouble is that the input and output openings of some buckets may be a bit tricky to get into. While your bucket might more closely resemble a large pan, my bucket might be extremely complex and can only be accessed through a series of winding tubes. Nonetheless, these buckets can be shaped over time to be more easily accessible.

Ok, so that’s the model. I think you know what I’m getting at here.

First, I think output is largely synonymous with experience. This is especially the case in very technical and innovative teams — the amount a team member can contribute is heavily correlated with their experience, and each time they produce something they acquire more experience. Hence why gold mined can be used to upgrade excavators — individuals become more performant with additional experience.

Second, this experience is replaceably transferrable between individuals through collaboration and discussion, hence the magic cloning buckets (our brains!).

Third, the ability of an individual to transfer their experience to another is dependent on both the ability of the individual to teach, hence the various output openings, and the other to learn, hence the various input openings.

So, in this model, output is not just the sum of what each team member can produce. It is also a function of how well team members can communicate with each other.

Overall, I’d express it as

where N is the number of team members, P is the output of an individual team member within a certain timeframe, alpha is the ability of a team member to teach, and beta is the ability of a team member to learn.

This expression derives from the fact that each unit of production (P) by one team member will definitely collect in their bucket, and to transfer that unit onto each other team member, it must both be taught by the team member (alpha) and be learned by the others (beta).

This framework helps determine what is the highest marginal value improvement in the team. I think mediocre teams generally focus too much on P, often because they underestimate or are even unaware of the impact of alpha and beta. If alpha and beta are both zero for all team members, then each unit of individual productivity increase translates strictly to one additional unit of output.

But in a team of ten, even if the alpha and beta of all team members increases from 0 to only 0.1 (that is, each person only shares 10% of their accumulated experience to the others), then the total output is doubled!

I’ll leave the derivation of the exact partial derivatives of output w.r.t. alpha, beta, and P as an exercise to the reader.

The point is this. If you’re a team manager, ask yourself: is my team P-deficient or alpha-beta-deficient? Don’t just reward your high P team members, reward those high alpha-beta team members that lift up those around them. Be extremely sensitive to those silent killers: high P, low (even negative) alpha-beta team members.

And if you’re a team member, don’t neglect your alpha and beta.

That’s what I’ll be referring to when I use the term ‘decisive immortality’. Here’s why I think decisive immortality should be considered one of the top priorities for the next century, and death itself should start to be viewed as a serious existential risk rather than a mere fact of life. Because it isn’t.

The most basic reason is that death is, more often than not, the loss of utility. When day after day lives are cut short before marginal utility w.r.t time reaches zero, that represents a massacre of potential utility at biblical proportions. And it does seem to be the case that most people would like to live longer than they generally do.

But we all know that it’s not as simple as that. Even if we were 50% of the way towards decisive immortality, the world would be drastically different in ways that would affect the extent of people’s awareness of their utility functions, which would affect their decisions, and so consequently other people’s ability to achieve utility.

So let’s consider that world.

Many of the concerns towards increasing longevity are already addressed here, so I’ll save some breath repeating these. Although, I will add that I believe we should, in principle, assume most of these concerns will be addressed in practice by the many brilliant minds in the world.

The main contribution I want to make today is to further expound on one crucial argument for decisive immortality.

I believe mortality itself is a demotivator rather than a motivator. At least, the likelihood of this being true for a particular person increases with their expected contribution to society. Societal contribution is correlated with awareness of one’s utility function, which is correlated with having grand long-term life plans and goals, which is correlated with wanting more time in their life.

Today, I see capable people with goals lose motivation to either achieve and/or pivot due to an awareness of their mortality. Why invest more heavily in myself when I only have a limited amount of time to enjoy what I have now? Why start a new endeavour when I very well may not achieve a meaningful level of mastery before I die, and even if I do, reap the rewards of my pivot?

In a time when meaningful career contributions require immense specialisation, our awareness that we only have a very limited bankroll of time incentivises us to take variance-reducing bets with our educational and career choices. More often than not, we observe other people’s decisions and use that to proxy what the highest Sharpe bet is. Often the relevant optimisation function is wealth, so maximising this doesn’t necessarily, and likely doesn’t, even maximise individual utility.

This phenomenon leads to the loss of utility on a grand societal level for three reasons. First, those individual outcomes that have large positive externalities need to come from very skewed distributions. Money is nothing more than influence, and I think the influence money brings comes with each order of magnitude change in wealth, rather than in a linear fashion. If you currently earn $200k a year, your influence point is probably nowhere beyond family and friends, and this is unlikely to change even if you get a massive pay bump to $500k a year. It’s only when you start earning $2m+ that your influence point might actually start pushing more towards community, thereby providing utility for a larger scale of people. Once you start earning $20m or even $200m+, then you can start making serious humanity-level impact.1

Now imagine the vast majority of the most competent people in the world — those that are most likely to be able to achieve those skewed outcomes — choosing choices with a distribution so tight that it clips off their maximum potential. I think this tremendous loss of human capital should be considered a humanitarian crisis.

The second reason why risk aversion in life choices is extinguishing collective utility is because decisions have synergistic effects. We currently already have maybe 2-3 shots at a good career, and when we reach significant life extension, this number will likely reach 7-8, not to mention the general cultural shift towards job-hopping. When your experience in one career enhances your ability to perform in the next — doctors applying domain expertise in tech, physicists applying arcane methods in finance — the chance of success of reaching those tail outcomes that could hugely benefit society becomes greater with each unit of accumulated experience. By continuing to choose low-variance outcomes, we simply forgo that potential. And this effect means that sometimes those ‘highest EV’ bets we end up choosing may even be lower EV than those skewed decisions.

The third reason is that, especially for those people who don’t derive much marginal utility from experiencing multiple domains, decisive immortality enables individuals to pursue more depth in the domains they choose to pursue. This is because one’s ability to reach depth in a domain is inextricably linked to how much time they have to go deeper in that domain, as each level of depth can only be reached by first conquering the previous levels. This closely links to the first point, because much of those skewed outcomes that provide humanity-level benefits come from achievements at extreme depths within a domain, at the depths of science, technology, engineering, etc.

More time will motivate more change, more risk, and more depth, creating more human capital.

On the other side of the coin, more time will motivate more action and prevent the loss of human capital. I think meaningless hedonism, and especially those self-destructive acts (binge drinking, drugs, bedrotting), is motivated by mortality. We hear too often, whether in pop culture or among our peers, that ‘life is short’ so we should say ‘fuck it’ and ‘do what we want’ immediately. And to some extent, this line of reasoning resonates with the above: why try so hard now to achieve better outcomes when I’ll either be dead or too old and frail to enjoy it?

When faced with decisive immortality, the individual is held further accountable for their actions. More years to suffer or enjoy the consequences of your actions likely means more deliberation is put into your choices. To some extent, I think a large portion of those who practice self-destructive hedonism do so not because they will not in fact have the time to face the consequences of their actions, but rather because they are not aware or are simply choosing to ignore the long term effects on their utility. This is why I also think we need a general shift in our vernacular towards aging and death. Aging is a disease, death is simply a symptom, and the causes are treatable. Once we start taking that as a given, so will the fact that we are individually accountable for our long-term decisions.

Individuals will also be held accountable for long-term societal outcomes. As utility functions begin to overlap across generations, we will become more compelled to face the greatest existential risks threatening humanity today. Some of the tragedy of the commons that appear on a temporal dimension will begin to be acted upon, the most obvious one being climate change.

One counterargument against the premise that decisive immortality is a great motivator is Parkinsons’s law. It is certainly not unusual to see procrastination being motivated by the abundance of time before deadlines in our everyday life. Won’t this also be the case with the ultimate deadline, death, which would invalidate all these potential benefits of achieving decisive immortality?

I think this would only mostly apply to a certain subset of people, the subset of people that even if given enough time, would not be able to achieve much of substance. But those who are less vulnerable to Parkinson’s Law — those who are passionate and intrinsically motivated — those are the people that decisive immortality will find more time for, to get them closer towards achieving maximum individual utility and contributing maximum societal utility. So on net, I think to the extent that the premise is valid, the benefits would still be undeniable.

Before we even reach decisive immortality, much of these positive outcomes can already be achieved if we first convince ourselves that decisive immortality will be here soon. And I think that’s undeniable. On the biological immortality front, we already have someone (admittedly rich) who’s slowed their speed of aging to only 57% of their chronological pace, and there’s been significant headway made towards both understanding the causes of aging and tackling each cause. On the mind uploading and cyborgism front, brain-computer interfaces are coming into fruition alongside real-life cyborgs and some seriously cool shit.

The time is coming, be ready.

Grades, money, and pace. These are the 3 horsemen of Sisyphean pursuits.

These metrics are misguided because, socially, they almost exist independently of utility functions. It’s far too often taken as a given that good grades, income, or pace is a measure of success. But from an individual standpoint, almost by definition, success is the maximisation of personal utility. Utility and success cannot exist separately.

Additionally, there’s no clear limit to any of these measures. You can always compare yourself against an even higher standard, which often comes after you realise achieving your initial aim didn’t quite give you the dopamine hit you were looking for.

So what measures of success are not misguided? I’ve already sketched out a framework that describes what the utility function of any individual might look like. It comes down to those 3 dimensions: domains, depth, and time.

It’s no surprise then that grades, money, and pace have become default benchmarks to strive for, because each of these pursuits are a rough proxy for some aspect of an individual’s utility function. Grades proxy expected income in the future. Money then allows an individual to experience a variety of domains and some level of depth. Depending on the nature of your relationship with money, it could buy you time as well. Pace is an interesting one, because pursuing how much breadth of domains and depth you can achieve by an arbitrary deadline implies the existence of an ultimate deadline in the first place: death. So why not try extend the deadline? Does it really matter if someone else achieved something 10 years earlier than you if you end up living 30 years more than them?

In the end, grades, money, and pace are the one-size-fits-all prescriptions of utility functions. Precisely examining your own utility function and setting your goals and optimising carefully chosen metrics accordingly is the personalised treatment that will serve you more happiness than any proxy society could offer you.

I consider money a subtype of influence. Influence itself exists on at least two dimensions. The first is the scope of your influence. This may range from only a small portion of your own life, all the way up to the whole of humanity and its future. The second is the degree of impact you exert on this scope of influence.

Any given sum of money can be used to create some degree of impact onto a particular scope. You can use $100 to either treat yourself to a fancy dinner or you can donate it all to charity. The larger the scope, the less impact that sum of money makes, and vice versa. Furthermore, you can have an arbitrarily small impact on an arbitrarily large scope, and vice versa.

There exists a frontier on this 2-dimensional plane that represents all the possible combinations of scope and impact that a particular sum of money can provide. I call this an ‘influence curve’. A person’s influence curve is thus all the possible combinations of scope and impact that the person can provide using their aggregate wealth.

The wealthier the person is, the further out their influence curve lies. That is, for a given scope, a wealthier person can create more impact, and vice versa. Due to diminishing returns to capital and the laws of physics however, this shift does not increase linearly with wealth, but instead at a decreasing rate, until the influence curve approaches some supremum influence curve.

So what does my influence curve look like? I think something like this.

I’ve added my best estimate of Elon Musk’s influence curve and the average person in the world’s influence curve for reference.

Considering the average person’s influence curve, it’s no surprise that this framework of thinking about money is so foreign. For the average individual, non-negligible impacts can only be made at a very small scope, so it’s a good enough proxy to only consider what money can do for your own life. That is, most people are effectively unaware of the scope dimension.

Every individual has implicitly employed their wealth at a singular point on this influence curve, which is the average combination of scope and impact across all their purchases and investments. I call this point the person’s ‘influence point’. You could consider the scope of a person’s influence point as where their actions lie on the spectrum between hedonism and altruism.

This is where I think my influence point is at right now.

I don’t consider myself hedonistic — my favourite t-shirt is from Target and my favourite pants from Amazon, I don’t drink nor smoke nor go to clubs, my Mazda 3 is nothing fancy, and luxury goods are out of the question. But I do tend to pay hefty premiums on conveniences and optimisations where I think it counts — spending a tiny bit more on accommodation and flights during vacations, pulling the trigger on an Eight Sleep Pod and a Sleeping Duck mattress paired with a nice pillow, paying for a PT and a preventative health program, always buying paperback over ebooks.

The extent to which I act altruistically, apart from spending money on my family and friends, mostly comes from books, courses, and resources I buy to support those of my goals with huge positive externalities. Admittedly, I don’t donate much to charity, mostly because I’m not convinced it’d be the best altruistic use of my money. I have donated to Wikipedia though, because I use it, love it, and support free accessible information for all.

There also exists some mapping from this two-dimensional space to utility. I think this mapping defines the way and degree to which someone is internally altruistic, which in turn motivates the altruism of their actions, their influence point.

This is what my mapping looks like.

My mapping w.r.t. impact, I think, is almost defined by the scale itself (almost logarithmic) — of course I’m going to derive a shit tonne of utility from an existence-revolutionising change to my life. But I think it’s still worth formulating explicitly, because I feel people often forget about how drastically things could be better for humanity.

I don’t really know how to explain my weird mapping w.r.t. scope logically. If I were to psychoanalyse myself, I would say that the perfectionistic and competitive attitudes instilled within me during my childhood, complemented by a generous dose of not-enough-hugs, has made me more individualistic than I’d like to be and derive personal utility disproportionately much more from grandiosity (helping humanity) rather than mediocrity (helping community). I actively fight against the selfishness-inducing attitudes still lingering within me, not to change my utility function (although that might happen), but somewhat paradoxically to help me achieve higher utility within my current function.

Overall, I use this money-influence framework to help align my money-related decisions to my utility function. It’s very easy to chase money for the sake of chasing money — after all, it’s such a satisfying and explicit score to try optimise. But what I’ve realised is that given my utility function, and especially the strange trough in my mapping w.r.t. scope, it’s really not worth my time trying to optimise for money unless there’s a fat positive tail.

Hence my pursuit of imbalance.

There are 3 main inputs to my utility function:

1. Domains

2. Depth

3. Time

Domains: categorisations of the facets or areas of life that could be experienced (e.g., family, entrepreneurship, cooking, scientific discovery, sports). You can almost think of these as the skill bars over my Sims character.

Depth: the intensity at which I’m able to experience a certain domain. How many children am I raising, the scale at which my company is operating, how fine my culinary skills are, that sort of thing.

Time: the amount of time I have experienced each domain. Because the time must be ‘experienced’, this differentiates from chronological time. Instead, this is a function of the quality of my senses, the amount of alertness I have to actively experience life, the time I spend asleep (where my senses are dampened and chronological time moves faster relative to experienced time), the amount of distractions (whether external or things like physical pain), among other things. I consider healthspan a closely related measure.

So, my utility function is a mapping:

Overall, you can imagine many sheets of paper. Each sheet of paper is a given domain, and the length and width of the paper are the time and depth to which I have experienced that domain. How do these sheets of paper map to the cumulative utility I receive? Spoiler alert: it’s not simply the combined area of the sheets of paper.

Some almost obvious features of my utility function and how it maps the combined space to a utility value:

• Monotonically increases with respect to each dimension.

• Increases at a decreasing rate with respect to time.

• Monotonically increases with the quantity of non-zero elements in the space of domains, i.e., I derive some utility from variety in domains, beyond simply the sum of the utility gained within the domains themselves.

Some more unique features and differences from that of what I estimate to be the average person:

1. The domains with the highest gradients w.r.t. depth and time include: entrepreneurship, technology, scientific theory. I have a near-constant sensitivity to time (albeit not high enough of a sensitivity to incentivise me to spend more time in it) in my favourite ‘non-productive’ domains, such as computer games and board games.

2. Locally, my gradient w.r.t depth is increasing for these high grad domains. But I suspect that if we zoom out a lot, it’d look much more like a sigmoid function, although I can’t reasonably estimate where the inflection point is, nor do I suspect it to be close enough for me to care.

3. I think the utility I derive from variety of non-zero domains is slightly less than the average person my age.

4. I think my gradient w.r.t. depth is higher than the average person my age in my highest sensitivity domains, and vice versa for my lowest sensitivity domains.

5. My gradient w.r.t. time increases with depth in that domain.

And now how these characteristics motivate my actions and goals:

• 1, 2, 4: Options trading is a great intersection between entrepreneurship, technology, and science, my highest grad domains. I really can’t think of a job where progression is any faster, which satisfies my need for depth and effectively speeds up my experienced time.

• 2, 5: Any job doesn’t get deep fast enough to counteract the decreasing depth grad over time, nor does it ever get deep enough in the entrepreneurship domain for my liking.

• 5: My pursuit of depth in certain domains helps counteract the diminishing marginal value of time, making extra time worth pursuing. I would one day like to contribute to the field of life extension, and to the extent that I can contribute in a way that satisfies depth and time in my highest grad domains, I would not be surprised to see this becoming my magnum opus.

And what do I see in our future now? I see pitchforks.

Nick Hanauer
On a TED Talk in New York, August 2014

One of my closest friends recoils at the thought that I’d like to contribute to life extension.

“Things like this always disproportionately benefit the rich. You ought to question whether you should.”

This possibility is undeniable, but I am in principle against using this line of reasoning to conclude that life extension is not worth pursuing. The essence of the argument is this. Given the current state of governance and distributive economics, the existence of a significant advancement in life extension will widen the disparity between the wealthy and poor to such an extent that the final outcome will be so negative overall that we end up worse off than we are now. Sure, if society evolves to effectively alleviate inequality across both dimensions of wealth and lifespan, life extension could be a revolutionary net positive. But that’s not where the world is, so let’s take off those rose-tinted glasses and steer away from such fantasies.

Unfortunately, both my friend and I studied economics, which means we’ve been repeating the same misguided mantra for years: ceteris paribus.

Technological innovations do not occur within a vacuum, and we must not assume that everything the new technology does not immediately impact — our institutions, our cultural principles — will remain constant after the advent of that technology.

Indeed, I’m optimistic that once a technology as revolutionary as significant life extension is introduced, one that could make or break society, humanity will take a long hard look at itself and decide it would rather evolve than crumble.

What we can do to further increase the chance that the world will move in the right direction is to continue shining spotlights on the problems that could be catastrophic tomorrow. It’s especially motivating to see individuals who could be considered one of the perpetrators, such as Nick Hanauer, encouraging change among their peers. And if we can’t, the pitchforks will come and the world will show us. Even if it wasn’t Luigi Mangione, it would’ve been someone else.

Most importantly, we need to stop being so fucking pessimistic. Stop thinking in terms of EV. Start thinking in terms of the maximum.

We must rewire the entire way we think. We must stop assuming that the state of entity X at time T+1 will be the state it is in now. We must instead assume entity X at T+1 will be the expected state assuming everything went right.

Because trust is often self-fulfilling, especially when it comes to people. Trust that humanity will do the right thing, and more often than not humanity will.

Doing so means generalising this principle from life extension to any major technological advancement, especially those that could be extremely positive given the right societal conditions. Some other examples I can think of are AGI, human genetic engineering, and space settlement. We need to start seeking what we want, and stop trying to avoid what we don’t.

The future is bright, we just have to create it.

Studying maths and economics at university implicitly taught me that the value of everything should be evaluated by calculating its EV. The optimal decision for the rational agent is that which provides the highest expected utility. The value of a game is the EV of its outcomes.

So has trading options, but weirdly enough, it's also revealed to me a framework of thinking that I think will prove much more effective outside the theoretical world of Black Scholes and utility functions.

American options provide the owner the ability to exercise the option at any given time before expiry. To price American options then, one must consider at each point in time whether the EV of holding or early exercising the option is higher. The option's value is that of the EV-maximising path. In other words, when an option provides you with an infinite number of choices within a certain timeframe, the value of the option is the maximum possible EV it could ever provide you.

Stop thinking in terms of EV. Start thinking in terms of maxima.

Because using EV forces you to frame decisions as one-period games whose outcomes can only be influenced at the time of choice, but no further. After that, it’s all up to chance.

But decisions in life are rarely the be all and end all. Think choosing a university degree is a big decision? What about which other students you will hang around? What extracurriculars you will explore? Which stranger in class you'll say hi to and will potentially change the course of your whole life? There are always countless additional decision points contingent on your original decision that could dramatically change its final outcome. We are compelled to assume we will make optimal decisions. We are compelled to realise the best case. The decisions you seek should be those with the highest maximum.

There is also a practical argument. EV is incredibly hard to predict. You not only need to estimate all the possible outcomes, but also their associated probabilities. And when factoring in the self-fulfilling nature of perceived probabilities of outcomes, this exercise not only becomes difficult, but borderline self-sabotaging.

The maximum, on the other hand, reduces the problem to a single scenario, and in particular, the perceived best-case scenario. Focussing on this one outcome helps actualise it. And the unknown unknowns of the maximum can only strictly be better than the prior.

If you still make major decisions based on its EV, it signifies one of two things. Either you don't trust yourself to even attempt to make optimal decisions that drive you closer to the best case. Or you don't know what you want, so you lack a clear enough utility function to know what the maximum looks like, let alone an optimal decision.

More often than not, you'll end up taking variance-reducing paths, often ending up in institutions (university, big companies) that place incentives (marks, money) around you to continue following a comforting system (academia, the corporate ladder) that promises happiness but instead murders the best version of yourself by making decisions for you one after the other.

Think in terms of the maximum. Look around you. What are the subtle yet near-immovable forces that disincentivise you from making your own decisions? A well-paying job that you wouldn't do for free? An 'education' that provides nothing but numbers that fuel your ego? Get rid of them. Can't do it? Trust yourself. You don't need those things to make 'optimal' decisions for you. Now, imagine the absolute best-case scenario for the next hundred years of your life. Hold that picture, hold it tight. Then, compare that against where you are right now. What can you do right now to maximise the chances of you realising that scenario? No, forget about the EV of that decision. Think in terms of the maximum. Now go do it.

Why think about EV? None of your life so far was 'expected'. You are not 'expected'. You have free will for a reason. Go use it.