Vacation changes your sense of time. So does being in another culture, especially one where daily life seems have more room for lingering (I wrote more about this here).

On the slow boat ride, I didn’t look at my phone and I ended up reflecting and journaling. I thought about work, relationships, and what I had been learning, and realized how much I had to digest from the past few months.

I thought about the idea of digestion, where the body takes time to actually process the things that come into it. The body decides what to keep and what to let go of, and what will fuel it.

Maybe wisdom and discernment work the same way. But in most of our lives, there's little space for that digestive process.

The Missing Space

Coming home to the US, I talked with people about European vacations, slow summers, and was fed Instagram posts reflecting my post-vacation emotions like…

“Forget Hot Girl Summer, What About Slow Girl Summer?”

Nearly everyone I talked to seemed to say some version of: ugh, things are moving too fast.

There’s never enough time to finish your work, respond to every text, care for the people and animals we love, get to those creative projects, or feel caught up.

On top of that, we have built a culture in which speed signals importance and competence. A crowded calendar says you’re someone who is needed. Also, rest and vacations can become another project to optimize and win at.

Artificial intelligence has entered this already accelerated reality with an even more frenetic rhythm.

An investor recently told me they read AI news every morning from 5:00 to 7:00 and still do not feel caught up.

In other technological revolutions, there were years of social adaptation; industrialization changed labor, education, family life, and politics over generations.

With AI, we are compressing that cycle dramatically.

Teachers are creating their own AI policies before their school has one. A product manager is reinventing their own job before anyone knows what the new role should become. A nonprofit leader feels compelled to get everyone Claude accounts because suddenly all the funders are asking for their AI strategy. Parents are already forced to decide what forms of AI use are healthy for their children on their own.

The feeling isn’t limited to people trying to keep up with AI. Even many of the people building it are now asking for more time.

This week, 1,268 people who work at frontier AI companies signed a statement asking the U.S. government to deliberately slow down and pace AI development, to give time for governance, safety and security measures to catch up. OpenAI said in a statement,

“To realize AI’s potential, industry, government, and society at large may need the option to buy time to address emerging risks, develop security measures, and strengthen oversight.”

But some people hear the idea of “buying time” and think it’s naïve. The train has already left the station, they argue, and the brakes are broken. Others point out the irony that many of the people now calling to pace AI are themselves helping build it.

But then who is setting the frenzied pace?

Everyone and no one, it seems.

We see a lot of finger pointing: It’s the markets! No, it’s geopolitical pressure! No, it’s god-like egotism! No, it’s the government!

It’s probably all of these forming into a collective soup that is hard to parse out.

And we might be left feeling like there’s nothing we can do.

Slow Down to Speed Up

The feeling of “not having time” can distort judgement, and tint everything you look at.

I know that when I am stressed, my sense of urgency gets projected onto everything, even things that are not actually urgent.

It’s Sunday, but I am still racing through CVS to get the prescription, annoyed at the long line, my shoulders tense and my hands clutching the shopping cart.

Yes, finishing the errands is important, but there’s no deadline, this is just the pace from my week still running at full speed.

This does not mean that urgency is imaginary or that individuals are able to fix it. Those that want an AI pause are arguing that this can only happen as a collective effort across the field and that it’s beyond one company, team or individual’s actions. Collective pressure and policy is often what turns the tides on a wider scale.

But pace is also a norm that spreads itself through each of us.

We bring urgency into our decisions, our organizations, the example we set for children and it keeps rippling out.

Pretty soon, it can feel like we’re all racing toward a future we never chose, as if it were a foregone conclusion.

But often we’re on the default path only because we haven’t taken the time to see if there’s another way. Sometimes we need to take a breath to digest what’s actually happening; and allow us to step outside the feeling of inevitability long enough to see that choices still remain. That’s where wisdom enters.

My grandfather was a professor and author, and he died right as I entered college to study literature and philosophy. It felt unfair that I never got to share that world with him.

At 18, going to Germany was a way to try to understand and connect with him, as I knew the broad strokes of his story but not the details.

My generation grew up in a strange in-between: after families had begun losing many of their oral histories, but before everyday life was filmed and caught on phone cameras.

Many years later, as I am back in Berlin, I am thinking of him a lot.

I have no record of my grandfathers journeys here, or what drove him to study and visit here. But I like guessing if he was also struck by the same things I am.

Tiny cafés lit by candles burned halfway down inside glass lanterns.

People lingering over drinks for hours, smoking at sidewalk tables.

Groups gathering along the river as evening comes, sometimes turning into dance parties. Many clubs and bars where phones aren’t allowed at all, where people sometimes dance for 72 hours straight.

I hear from a few locals that intergenerational friendships are normal, where 60 year olds are friends with 20 year olds, and both still dance. A lot.

The pace is a stark contrast to San Francisco and New York where I’ve lived my adult life.

Here, there’s a sense of sinking into life, and I feel it in my nervous system.

Slowing down, I notice more, my senses are more engaged, there’s more space to feel, to think.

I spent hours on a boat filled mostly with German families, meandering down the river as they drank spritzes and laughed about things I couldn’t understand.

Along the river banks, people waved, danced, smoked cigarettes, and sat talking in groups. A thunderstorm rolled in. Even in the rain, some people kept dancing.

Coming back to the US, I feel a sense of sadness. In the past month, I’ve gone to multiple discussions and conferences centered on the “loneliness crisis” and the lack of 3rd spaces here in the US.

Here, so many spaces feel designed for transaction rather than lingering.

Social time is squeezed into narrow windows: a scheduled 30 minute coffee, an hour between meetings to power walk. Scheduling time with friends 6 weeks out. There are fewer moments where we can show up, hang out, and see what unfolds.

I’m not suggesting Berlin has escaped these patterns; I can’t claim to know what it’s really like to live there.

But I know that here in the US, something important is missing, and people feel it. The sense that our society has eroded something that we need like air: community, belonging, and spaces to just be.

It is telling that “human flourishing” has become something we now have to design for. We have made life faster, more measurable, and more convenient, while perhaps making it harder to fully inhabit.

AI risks deepening these patterns further. We talk about the risks of cognitive surrender to AI, handing over more and more of our thinking.

But what about the surrender of presence, intuition and bodily awareness?

As we check our Oura rings to find out how we slept (guilty) and ask AI what to say next, and how we should feel about this email from a friend that might be passive aggressive, the risk is deeper than cognitive.

As I drive home from the San Francisco airport beneath a skyline of AI billboards, I keep thinking about what we are building all this for. At its best, technology can help us create a richer life (Claude designed an incredible Berlin itinerary that I wouldn’t have been able to make myself).

If AI gives us more time and capacity, I hope we use these newfound powers to deepen joy, belonging, wonder, and our ability to be fully present in our own lives.

That led to a six-month exploration of AI and creativity: it was existential, strange and messy and I am grateful for it.

Part One: Analog

First, I wrote a song with just me and the piano.

No technology, totally analog: paper, pen, piano. I even prevented myself from using an online thesaurus.

What came through was a song called “Someday You’ll Dream,” about going through grief and coming out the other side.

Part Two: Existential Crisis

Then, for phase two, I wrestled with the question: how would I bring AI into this process?

At first I resisted…..

Songwriting feels great with no technology, raw, real meaningful.

But I was curious, so I tried to play with AI tools with an open mind. Most of the tools out there generate music for you, like Suno.

After generating a bunch of songs with it and similar tools, I fell into a bit of existential questioning:

I can push a button and get a fully formed, professional-sounding songs in seconds, something I’ve worked so many years to be able to do. It brought a kind of grief, a vertigo.

What does this even mean for art, for music, for the sacred artistic process?

Beyond my own personal vertigo, I deeply felt the market reality. It’s now so cheap to make a song, and will likely disrupt all parts of an already challenging industry. There’s also the ethical issues of training AI models on art and music without compensation, attribution or permission.

I talked to some artists and colleagues in the music industry, as we saw AI music flooding streaming platforms and even some charts.

“The song is dead,” we half joked in a kind of dark way. Love live the song.

At that point, I almost gave up on AI.

Maybe the answer was to reject this new technology completely, go make music in the forest with a banjo and the trees and preserve that beauty in defiance.

But the conference was looming in a few months, and I couldn’t give up on the exploration.

Part Three: Creative Integrity

So I took a step back, emerged from my existential crisis, and started again.

I started to analyze why I felt so empty after using some AI music tools, and came to the idea of creative integrity.

When I say creative integrity, I don’t mean some rigid idea of artistic purity, or that “real art” can only be made one way.

I mean integrity in the sense of wholeness: the feeling that the creative act is alive and inhabited, and that the process of making was itself worth undergoing. This might be different for everyone.

For me, my creative act has integrity when it holds three things:

• Vision: it is connected to an intention and purpose.

• Transformation: the maker is changed by the process.

• Meaning: the result carries emotional, aesthetic, or spiritual resonance.

AI-created music can potentially have meaning (most people can’t tell a song is AI, though I know this is hotly debated).

It can have vision, because you can prompt your way to a song that sounds exactly how you want it.

But what it didn’t have, for me, was transformation.

And without transformation, the song may have sounded complete, but the process felt empty.

In my experience, the creative process can include struggle, iteration, failure, doubt, ecstasy, insight, communing with something bigger than ourselves, or simply time passing.

So I decided: if I could make a piece of art with AI that didn’t skip out on this process or even deepened the process, it would be worth doing.

Part Four: Build it Yourself

I started playing with Claude Code, and I built a custom tool I called SongLab that draws on what’s available online about some of my favorite songwriters, listens to my tracks via MIDI, and offers musical ideas.

I wanted it to feel like being in the studio with Bob Dylan or Joni Mitchell and have them ask, “Could you try this kind of metaphor, or this chord change instead?”

That felt generative. It challenged me to deepen my process and try new ideas, and showed me that another way was possible: One that supported my creative process, rather than undermined it.

Part Five: The Stage

When I finally stood on that stage in Athens to share this whole 6 month exploration and perform “Someday You’ll Dream,” I was jet lagged and sleep deprived.

But I felt those notes ring out from the grand piano and all the questions disappeared.

I knew clearly in that moment that bringing our messy human selves to the process of creation won’t go away.

We’ll keep coming back to that, and keep wanting that, because its a core part of being human.

No tool or technology can strip that from us.

We’re living in a moment of intelligence on tap.

We don’t have to accept the out-of-the-box AI tools if we don’t like them as is.

We can design our own process with them, or even build our own tools to be human-centered so they enhance our creative process, and bring out the deeper core of what matters most.

The defaults are not destiny.

We can learn from what great artists have always done: turn the tools of their time into instruments for bringing visions and our deepest dreams to life.

I was in the middle of testing Fable 5 when it happened, using it to edit an app my team has been building, and it was fast, and surprisingly good. The experience was the best I’ve had from an AI coding tool so far (Full disclosure: I am a mere vibe coder, not an engineer).

It felt almost magical.

Abracadabra is often translated as “I create as I speak.”

When you use AI, you describe what you want in plain language, and boom.

Something appears.

That “something” that comes out may be slop, or many iterations from what you had in mind.

But as these models improve, the distance between your imagination and the reality starts to shrink.

An hour into my working session, this screen popped up:

The decision to remove access upset many people who saw it as too opaque and abrupt. Some were mad at the government: The process to regulate access should be transparent and democratic.

Others argued Anthropic had invited this action by warning that models like these are too dangerous in the wrong hands and should be regulated.

As of today, five days later, we’re still blocked from using them.

Regardless of how this particular situation plays out, this will not be the last time we have arguments about access to frontier AI models.

This is because its really a question about power.

If something is incredibly powerful, should everyone have it? Or only a select few?

And if only a select few, who decides this?

Some people say advanced AI is more like a nuclear weapon, so powerful that unrestricted access unacceptable risks.

Others argue AI is more like electricity or the internet that create public benefit when widely distributed.

That is what makes the access question so difficult.

If you restrict AI too much, you may also limit breakthroughs in science, medicine, education, and other fields. A model that can help someone understand dangerous biology can also help a researcher discover a new treatment.

This is also true in cybersecurity, where the same thing can be used to attack a system or strengthen it.

In an open letter, more than 50 cybersecurity leaders argued that the Fable and Mythos restrictions should be lifted, warning that taking these capabilities away from defenders could make the cybersecurity ecosystem less secure and give other countries an edge.

It’s hard to govern a technology where the positive and dangerous uses are so intertwined.

This tension between freedom and security is not new.

In the United States, we have argued about it for centuries through the Second Amendment: When a tool gives people power, does restricting it make society safer, or concentrate too much authority in the hands of the state?

The debate does not have to collapse into the binaries of “open everything” or “lock everything down.” There are other ways to think about access to AI systems:

• Credentialed access gives more power to people working in sensitive fields, such as cybersecurity or biology, when they can show they are using it for legitimate research, defense, or public benefit.

• Independent audits and capability evaluations put checks on the companies building these systems, so the labs are not the only ones deciding how powerful a model is, what risks it creates, or whether it is safe to release.

• Public-interest access tries to prevent AI power from going only to governments, corporations, and wealthy institutions.

• Open-source and open-weight models create a more open market of models. Some may have stronger safety practices embedded, and some may not. Over time, the safest and most useful systems may be more likely to win trust and adoption.

All of these proposals come with their own risks and challenges.

In the end, these questions also come down to an even deeper ones about human nature. The models at the center of this debate were called Mythos and Fable, after all.

Do we trust humanity to wield power? We’ve explored this question in myths countless times.

In Star Wars, the force can protect or destroy depending on the person who wields it.

Lord of the Rings tells us that the One Ring is corrupting it cannot be safely used, even by the good side.

Harry Potter says that education is essential for us to wield magic safely.

The Sorcerer’s Apprentice reminds us that danger comes from a lack of understanding combined with ego.

We might have something to learn from all of these stories.

AI will open up extraordinary powers: to build, discover, create, and act in the world in ways we can’t fully imagine.

But the power to bestow power is also power.

The responsibility to decide who gets access to AI cannot belong only to private labs, government agencies, or national security officials behind closed doors.

As AI starts to feel more like magic, the authority to grant or restrict that power needs checks of its own.

Young people are entering a complicated and uncertain future, while commencement speakers are being booed over AI and the job market. There was uncertainty then, too, but it felt simpler somehow.

It made me think about the best advice I got around that time, which didn’t from a commencement speech. It came from one of my favorite books: Rainer Maria Rilke’s Letters to a Young Poet. He says:

Text within this block will maintain its original spacing when published

“Be patient toward all that is unsolved in your heart and try to love the questions themselves, live the questions now.”

I learned to love the questions throughout my time at St. John’s College, where every seminar began with an opening question and unfolded into a multi-hour conversation among twenty people gathered around a wooden table.

St. John’s was shaped in part by Stringfellow Barr’s Notes on Dialogue, which helped define dialogue as a discipline of shared inquiry: the practice of staying with a question long enough for something deeper to emerge.

Living the questions means resisting the urge to rush toward easy answers or absorbing the assumptions being handed to us.

It can be uncomfortable to sit inside of that kind of uncertainty, but as Maggie Johnson writes, sitting with uncertainly can be good for us.

The practice of living the questions feels especially important in this moment when we risk offloading too much of our own thinking to AI.

A new paper on AI epistemic risks identifies three mechanisms through which AI may weaken our ability to reason and judge well:

• persuasion and manipulation, as AI systems shape our beliefs;

• cognitive offloading, as we delegate more thinking and judgment to machines;

• and feedback loops, as AI-generated outputs narrow the set of ideas in the collective field.

How do we protect these cognitive capacities when the world is speeding up and the incentives are pushing us in the opposite direction?

At Resonance, we’ve been creating in-person spaces for people to slow down and think together about AI and the larger changes happening in society and how we want to respond.

What I’ve seen from many of our Resonance gatherings is how powerful it is to practice this together rather than alone.

For many people, AI feels overwhelming. We encounter it through newsfeeds and social media, where the story swings wildly between utopia and catastrophe and is moving at lighting speed.

Most of AI users are still getting used to AI as a chatbot, while the technology is already moving toward agentic systems that carry out work autonomously. Meanwhile, the web itself is beginning to feel less human: Cloudflare reported that more of the internet is now being accessed by machines than by people.

Often, we are left trying to process all of this rapid change alone.

But humans do not metabolize change in isolation.

We make sense of things relationally: by hearing another perspective, seeing a new side of an issue, finding language for something we couldn’t put into words, or commiserating with others who are going through the same thing.

Maybe this is why Isabelle C. Hau has described this next chapter as an era of “relational intelligence":

I am still listening to Rilke’s advice all these years later. And I’ve learned that questions become more livable when we carry them together.

If you’d like to join or host a Resonance Salon, we’d love to hear from you! You can learn more here.

What is a good life in the age of AI?

It was a diverse group, German, Mexican, Turkish, American, Greek, Portuguese, French, English, Nigerian, Japanese, and more, with viewpoints as varied as the nationalities.

Here are some of the threads that emerged from the 3 hour discussion…

A Good Life Means…

• A life of connective tissue and care, extending care beyond just humans to animals and planet.

• To live in harmony with nature, or as the Greeks said κατὰ φύσιν (kata physin).

• To exercise your purpose with reason and agency.

• A quiet life where you can hear your inner world.

• To live in an ethical manner and relationally with all beings.

• Is may be unreachable but we can always strive towards it. It’s a dialogue.

• As Socrates said, the examined life. Not from school but from exploring life. This will serve us in the AI age.

• Be critical of the question itself. To touch Black life, Indigenous life is to touch on a darkness underneath this question. Often the proposed ideas of the ‘good’ can lock us into something that has an unjust shadow.

• Happiness is overrated. A creative, difficult, relational life is more fulfilling.

• Comedy and laughter: it helps us cope, it criticizes, creates hope. Brings light and levity into doom and gloom.

Why does this question even matter right now?

You could say it’s a privileged luxury, that anyone asking it has too much time on their hands.

I’d argue the opposite.

We're in a global arms race for AGI (artificial general intelligence) in the middle of a polycrisis, and the decisions we make now are more consequential than ever. Many reach for a phrase to describe what it's all for: human flourishing. But it rarely gets defined clearly.

When I was 16, I traveled with Habitat for Humanity on an earthquake relief trip in Central America. I came back unable to make sense of what I'd witnessed: communities completely flattened, families lost, very little aid available to rebuild.

So I started reading philosophy and taking classes, including "Jesus, Buddha, Socrates" on Saturdays at the local college, because I was hungry for answers. Eventually it led me to St. John's College, to study ancient Greek and wrestle with these questions. I'm still wrestling.

One thing we learn from the Greeks, and especially Socrates, is that it's not about finding a clear answer.

The questioning itself is the practice.

We're in a moment of deep uncertainty. That uncertainty causes anxiety, and we often don't know how to hold it.

It's tempting to rush to conclusions and answers are offered at every turn, on CNN, on TikTok, in ChatGPT's neat summaries.

But as Rilke says,

“Be patient toward all that is unsolved in your heart and try to love the questions themselves, like locked rooms and like books that are now written in a very foreign tongue.”

Standing at the Temple of Athena later that week, I was reminded to love the questions.

Which of the answers above resonates with you?

Or do you have another definition of the “Good Life?”

And yet, almost nobody is actually listening to it. On Deezer, where these numbers come from, AI music accounts for only 1 to 3 percent of total streams.

You’d expect the opposite. If listeners can’t tell the difference, shouldn’t AI music be taking over?

When a song resonates with you, you go looking for the artist. You want to know who they are, what else they have made, their story. With AI tracks, you might find a generated cover image, a name that leads nowhere, no body of work, no person behind the sound.

We don’t go to music just because it sounds good, but because someone else has lived a particular life and tried to express something true about it.

We meet another human there, and somehow the strange experience of being alive becomes more bearable.

Walter Benjamin talked about the “aura” of a work of art. Writing in the 1930s he said: “what withers in the age of mechanical reproduction is the aura of the work of art.”

The dominant AI conversation right now carries an enormous amount of fear about humans being replaced, and the fear is not unfounded. We are seeing real disruption, reshaping of work, and anxiety about what is to come.

But the discussion might be missing something important: What is human-made could become more valuable, not less.

The industrial revolution took us away from the maker and the economy became less relational. Skills that used to live in a person, the weaver, the baker, the cobbler, were broken into steps that machines and interchangeable workers could perform. The product got cheaper while the human behind it disappeared.

Bread, your coat, your shoes, the device you are reading this on, none of it really has a face anymore. That was the mechanization of production.

But now, everywhere you look, there are signs that people are seeking presence, aura.

Boutique fitness for in-person, human-led classes, is growing 12% a year. Pickleball added 18,000 new courts in 2024 alone, and is being called the country's most explosive "third place" revival. Indie bookstores are making a “shocking, triumphant'“ comeback. The live music market is rising 8.4% every year.

Starbucks spent the last several years removing baristas out of its stores in favor of automation. It didn’t work. CEO Brian Niccol announced last year that the company was reversing course, hiring more baristas, bringing back ceramic cups and handwritten notes on orders, and putting in nicer seating that invites people to stay longer in the stores.

These are all signs of hunger for embodied, relational experiences in a culture of increasing automation.

Economists often talk about value in terms of scarcity.

If content becomes infinite, what becomes scarce is the knowledge that a real person was there: shaping something, risking something, offering something of themselves.

Maybe that is why AI music can flood the platforms and still barely be heard: because it fails to create a relationship between the artist and the listener.

We can all, even in small ways, help nudge the world toward this more human-centered direction.

Every time we choose the café over the Zoom call, the live show over the generated playlist, the block party over another Sunday alone with a screen, we are casting a vote for the kind of future we want to inhabit.

My vote is for a world where human presence, story, and trust matter more, not less.

In a few weeks, I’m traveling to Athens for the first time to speak about AI. I keep thinking about the juxtaposition and how a technology pulling us toward unprecedented futures is also pushing us back toward some of humanity’s oldest questions.

Its use and deployment of raises questions about truth, knowledge, value, judgment, and what it means to be human. The design and development of AI systems reflect philosophical assumptions.

But philosophy is hard to practice in a world organized around speed, certainty, and black and white thinking.

When philosophy enters our organizations, cultures and systems, there is a tendency to translate it into something more operational. That move is understandable; we need decisions and processes, and can’t get frozen in questions. But it can also drain philosophy of its depth.

Part of the challenge is structural: To stay with a philosophical question long enough to uncover its assumptions often requires slowing down, while modern life rewards speed and forward motion (especially in AI, which is framed as a “race.”)

It’s also existential: Asking hard questions can put your career at risk. Socrates was willing to question things at any cost and died for it. Most of us live inside systems that reward agreement more than moral independence.

It’s also psychological: Philosophy asks us to sit with uncertainty rather than rush to easy answers, while human beings tend to want certainty and a sense of control.

Under these conditions, ethics often gets translated into something more manageable: principles, frameworks, review processes, and compliance checks for reducing risk.

But ethics isn’t a checkbox.

Sometimes even saying “we are doing ethics” creates a dangerous illusion: that once the review is complete, the deeper questions have been answered.

Philosophy is lived practice. The word philosophy comes from the Ancient Greek words φίλος (philos) 'love' and σοφία (sophia) 'wisdom,' often translated as “love of wisdom.”

Love implies relationship: something we return to, care for, and deepen over time. This often demands us to see beyond the pressures of culture, consensus, and the endless scroll of our social media feeds.

Aristotle’s concept of phronesis, or practical wisdom, is relevant for the modern world. Practical wisdom is the cultivated ability to make good judgments in complex, changing, and uncertain circumstances and asks if these judgements are wise, humane, and aligned with a larger vision of the good.

These questions are not abstract. If AI is to be shaped in ways that support human flourishing, then the quality of our judgment matters a lot. Also, AI makes it easier to outsource parts of thinking itself, and so the discipline of philosophical reflection becomes more important.

But our modern world and institutions are not built for this type of practice.

Can we evolve our institutions to be able withstand this type of questioning, of true philosophical practice?

Or does the burden still fall on the individual, who must risk themselves, their security, or belonging in order to think freely, to become a gadfly like Socrates?

The answer is probably a bit of both. We need cultures, communities, and institutions that make deeper reflection possible: creating room for uncertainty, constructive debate, dissent, time for judgment, and more spaces where assumptions can be examined. At the same time, individuals still carry responsibility. None of this is easy.

Philosophy is the love of wisdom, and love is a relationship, something we keep showing up to, again and again, especially when its’ hard.

And in an age when our technologies are advancing faster than our wisdom, that practice matters more than ever.

With something this complex, consequential and overblown in the media, its easy to just collapse into binaries: embrace everything or fear everything. Adapt or die.

It makes sense….

AI is a technology we now hold in our hands and use every day.

But it’s also a force reshaping labor, creativity, education, relationships, and power. It carries real risks, promise, and enormous uncertainty.

On one day, we’re told it will cure all diseases. On the next, that it may wipe out the entire planet.

We live with the whiplash of AI researchers signing petitions to pause development while, on that same day, your boss is telling you to help the whole team adopt “agentic tools.”

No wonder so many people struggle to think clearly about it.

Our minds are not great at holding that much contradiction. We want resolution, we want a villain or a savior. But something can be beneficial and harmful at the same time.

There’s a moral confusion too:

Should I use this thing that is, frankly, incredibly useful in my work and life, even if it may displace workers, extract from artists, consume enormous resources, and accelerate harms I don’t agree with?

There’s also an intense social pressure around it. In some circles, you’re shamed if you use it or talk about it.

In others, you’re judged if you’re not “token-maxxing” and happily vibe coding your way into techno-utopia.

So of course many people are left with questions they don’t feel safe saying out loud.

I’ve been in a lot of closed door conversations about AI, and I've seen that despite what you read or watch, many people actually don’t feel black or white about it. Some of the things I keep hearing:

• A widespread sense that the default path is not leading somewhere good. The default path is optimized for speed, scale, and profit, which means the benefits of AI are likely to concentrate in fewer hands and deepen existing inequality.

• A recognition that these tools can be incredibly useful and hold promise, and that rejecting them wholesale is neither realistic nor smart.

• A feeling that what is human is precious, sacred even, and worth protecting.

• Grief: about the speed of change and about what may be lost before we’ve even had the chance to process it.

Taken together, these are concerns about the texture of human life, and about what happens when a technology begins to reshape it.

AI is a cultural technology.

That is part of why this moment feels so disorienting. The questions AI raises are cultural, moral, relational, and political. What should remain human? What should never be automated? How should we govern it and who gets to decide? And what have we learned from the last time we let a powerful technology reshape society faster than our wisdom could keep up?

If AI is reshaping the conditions of our society, then we need more than AI literacy. We need cultural infrastructure to navigate cultural technology.

We need spaces where people can deliberate, not just consume media about it. Places where communities can build shared understanding, wrestle with tradeoffs, mourn what is being lost, imagine alternatives, and make more collective decisions about the futures they actually want.

We need democratic muscle: spaces and practices that help people make sense of what is happening, voice what they want to protect, and have real agency in shaping what comes next.

This technology will not only affect a small set of experts, founders, or policymakers. It will affect all of us, and we all deserve a voice in how it unfolds.

Unfortunately, many of those spaces and habits have eroded, leaving us more fragmented and polarized, and many people unsure that their voice could matter anyway.

We can create the cultural infrastructure to navigate this moment differently, and I see seeds of it everywhere: civic renewal, renewed investment in social connection, and growing attention to community infrastructure.

Let’s nurture those seeds instead of feeding the forces of othering, division, and easy answers.

If we do, we can help shape what comes next together.

The arrival of LLMs set off a cascade of changes. But in 2026, AI has entered the public consciousness as a full cultural force, carrying with it a mix of awe, dread, and disorientation.

57% of Americans rate AI’s risks to society as high or very high. Headlines warn of mass displacement, eroding creativity, a future that feels anti-human. The pace is dizzying, and for many people, the dominant feeling is that something important might be slipping away.

I’ve been hosting Resonance salons with community members, educators, artists, healthcare workers, and tech builders.

I’ve sat in small circles with artists grappling with AI’s impact on creativity, and talked with founders, executives, and CTOs reimagining teams and companies for this new era.

I’ve witnessed that this moment is bringing up grief, and a deeper fear that the things we most value may start to seem irrelevant in an AI-shaped world.

As artist Nick Cave recently wrote:

“Notions of artistic struggle, of striving, of triumph over adversity, personal pride, desire, delight, inspiration, and resilience are dismissed as mere obstacles or indulgences on the way to the new and gleaming product.”

The philosopher Günther Anders called this “Promethean shame,” the humiliation we feel when our own creations surpass us. He described a pilot trained to ignore his instincts and trust the instruments, because the machines are more accurate than his senses. We're living a version of this now, measuring ourselves against AI's output and its speed to do what would take us much longer to do.

But in almost every one of these conversations I’ve witnessed, no matter how different the room, people end up talking about what is human that they want to protect.

More people have said things about the “soul of humanity” in 2026 than I’ve heard in the last decade.

Words like care, compassion, vulnerability, community, and connection are coming up over and over.

The analog fantasies of going back in time are everywhere and have become an aesthetic. “Dumb” phone sales rose 25% in 2025, vinyl hit $1 billion in U.S. sales, and Fortune calls it a growing “a $5 billion analog economy.”

What’s underneath all of it feels like the same impulse: a desire to reclaim what we feel might be slipping away.

Everywhere I look, people are newly reverent toward the kinds of experiences that still carry an aura: live music. Game nights. Potlucks. Block parties. Knitting groups.

And the messy, imperfect parts of being human.

The Japanese have a practice called kintsugi: repairing broken pottery with gold, so the broken cracks become a beautiful part of the piece.

Even if the existential risks and displacement fears about AI are overhyped, the narrative is forcing something.

AI can produce the flawless version. The smooth, efficient, optimized output. And in response, people are reaching for the cracks. The handmade, the lived-in, the human.

There’s this tension right now between the techno-accelerationists who say speed up or get left behind, and the analog dreamers who say unplug and go back. Both have something right and both have something missing.

I dream of a third path.

One that’s realistic about the incentive structures, the markets, the pace, and that doesn’t pretend we can stop what’s happening or go back to some imagined simpler time. But one that also refuses to accept that the arc of technology has to bend away from what we value most.

AI is the most powerful mirror we’ve ever built. It’s forcing us to define what is sacred to us and what we cherish about being human.

Maybe that is AI’s accidental gift: forcing us to fight for what we refuse to lose.

Recently, I was testing out AI music creation tools, and I made a song in about ten seconds with Suno.

They are all “songs.” But they are not the same thing. Why not?

Audiences may not be able to tell or care (this is hotly debated!).

But I can tell. Those 20+ hours spent included journaling before I even knew I was writing a song, lots of tiny decisions about arrangement, chord choice, harmony, sound design, and hours of practice singing it before recording. Then fine tuning a mix.

It can be efficient, but often it’s not, and sometimes it is very painful and frustrating.

But the struggle in the process is what makes me a better artist, heals me, forms and shapes me.

The same might be said for writing, crafting, researching, analyzing, and every cognitive process we’re starting to hand over to AI.

The email you let AI draft instead of thinking through what you actually wanted to say. The analysis you gave ChatGPT because sitting with the data felt boring. The decision you outsourced because weighing the options required patience you didn’t have.

Cal Newport recently argued in the New York Times that we’re in the middle of a cognitive fitness crisis. Our attention spans are a third of what they were in 2004. Research shows a significant negative correlation between frequent AI tool usage and critical thinking.

The strain required to craft a clear email, work through a difficult decision, or write something honest is how you develop judgment, clarity, and voice. Skip it often enough, and those capacities might atrophy.

The choice matters.

Which parts of your process do you want to keep struggling with? Which parts are you willing to hand over?

Maybe you keep the writing but let AI handle the research. Maybe you do the strategic thinking but let AI format the presentation. Maybe you journal by hand and then ask AI to help you find the clarity hiding in it.

There’s no universal answer. It depends on what you’re trying to get out of the process and the constraints you are operating under.

In our busy lives where more and more is demanded of us, it's easy to want to save time.

As AI tools get better, it will be possible and expected to hand over more and more, even if humans are still in the loop for verifying and checking work.

But if we're not clear-eyed and conscious about making a choice when we can, we might slowly lose things we don't even realize are slipping away.

The incentives driving AI push towards outputs and efficiency. Faster, better, more. But the most important things in life have never been primarily about outputs.

What happens to us when we stop doing the hard things that shape us?

The experience has made me reflect about what skills will matter most in the years ahead, where this is all heading, and what risks we should we aware of.

Skills for the Agentic AI Era

Even after just a few weeks of this setup, I already can see what skills will be important if things continue this way:

1. Cognitive: Metacognition & Critical Thinking

An important skill is the ability to think about how you think and then translate that into structures an AI can actually work with. AI works much better when you’ve thought through the actual shape of your own process. Your own patterns, standards, blind spots, and preferences become clearer because you have to teach them. And once the AI gives you something back, critical thinking is what keeps you from passively accepting it and moving on without understanding. The skill is questioning things at every level, what core assumptions are built in, what’s missing, and what you’re bigger goals are.

2. Organizational: Managing & Orchestrating AI Agents

Managing an AI agent is closer to managing a capable new hire than it is to using software. You have to be clear about what “done” looks like, review work with specificity, and give good instructions. The quality of the output depends entirely on how you structure things, onboard agents to a task, and give feedback. And this is just the beginning. Right now I work with one primary agent that can spin up sub-agents for specific tasks. The future is multiple agents working in parallel, each with different strengths, running self-learning loops and evaluating their own output. The skill becomes setting up the right structure, the right relationships between agents, and knowing where the human in the loop should live.

3. Process Design: Building Systems That Improve

Process design is the structure around the model. It includes where context lives, how work is handed off, how priorities are tracked, how decisions are recorded, what gets updated, and how the system improves as it is used. The real power comes from creating workflows and feedback loops that allow the AI to participate in an ongoing body of work rather than a series of disconnected tasks. My chief of staff reads yesterday’s state every morning, knows what’s due, remembers how I like things done, and picks up where we left off. It can track open loops, work across living documents, and help maintain continuity across projects. That kind of system is what turns AI from a one-off tool into a real collaborator, something that amplifies thinking, extends capacity, and helps move work forward with more coherence over time.

4. Human: Adaptability, Communication, Empathy, Creativity

The pace of change requires comfort with being a perpetual beginner. You have to be willing to abandon a tool or a workflow you just mastered because something better just arrived. But adaptability alone isn’t enough. The quality of what you get from AI depends entirely on how clearly you can communicate what you actually want. And as AI gets better at producing competent work, the premium shifts to what it can’t replicate: original thinking, taste, empathy, reading a room, building trust, knowing what’s worth making in the first place. The more AI handles the functional layer, the more the human value lives in the relational and creative layer.

Where is this all going?

As agents get better, a few things will probably follow. First, agents will represent us and collaborate with other agents.

My agent talks to your agent. A founder’s AI shares a deck with an investor’s AI and pre-negotiates terms.

Your partner’s agent meets with yours and plans dinner or weekend plans based on your calendars. It might save time, but what gets lost when the first layer of every relationship is mediated by a machine?

Second, agents will solve more complex problems. Instead of directing every step, you describe the outcome and the system figures out how to get there.

The human role moves up the ladder: vision, values, taste, purpose. But if we spend less time in the complexity and details, we may lose the capacities that used to be built there: patience, intuition, discernment, critical thinking.

What we need to pay attention to…

Security is the obvious one (malicious prompts hidden in documents, compromised plugins, tools that end up with more access than you realized). The more capable our agents become, the more security has to be a core design principle.

Beyond security, there's more and more cognitive offloading. There’s a version of Agentic AI that amplifies human intelligence, and another that turns us into consumers of machine-generated output. The difference is in whether we stay actively involved in the substance of our work or gradually become passengers in our own process.

And then there’s the physical and nervous system cost: hours at the screen, the cognitive load of interfacing with a system that moves this fast. Constant context switching.

The erosion of human connection is a risk that truly keeps me up at night. The more time we spend in dialogue with a machine, the less time we spend in the room with other people.

"The need for connection and community is primal, as fundamental as the need for air, water, and food."

-Dean Ornish

And then there's equity and access. If AI literacy and the ability to build real skills with these tools remain unevenly distributed, this transition will deepen the inequalities that already exist. The people with more access and more room to experiment will compound their advantage, while the communities most affected may be the least prepared to shape what comes next.

These last two weeks have been one of the most interesting experiments I've done since writing my first poem with ChatGPT the day it came out in November 2022.

It’s shown me just how much of a wild west this still is. The tools are changing faster than our ability to catch up, the rules haven't been written yet, and the full implications are hard to fathom.

But the fact that nothing is settled yet means everyone can be part of shaping how this goes. We still have a voice and power to shape how we want to incorporate these tools into our lives, our process, our work, our communities.

So get in there. Experiment.

I named her “Claudette.”

This was partly a productivity experiment.

But I also wanted to pressure-test the limits of these models from the inside and get a more visceral sense of where all of this might be headed.

What We Built Together

“Claudette” operates directly on my computer. “She” reads, organizes and writes files, manages a database, builds web pages, does research, and everything we do persists between sessions.

I started the setup by asking Claudette to research Chiefs of Staff from the White House to top companies and distill the core patterns of the role: how they prioritize, manage information, surface what matters, etc. That became the backbone of the system.

Here are a few of the more useful things Claudette and I have set up so far….

• A personal dashboard that opens every morning with my schedule, active projects, and follow-ups needed. It can link to research or meeting prep documents.

• A daily AI news briefing. Every morning, Claudette researches the most important AI news from the last 24 hours and generates a clean briefing page I open in my browser. A personalized newsroom built for one person. No algorithmic feed or ads. (The irony is not lost on me: I’m experimenting with offloading parts of my work to AI while reading about what that same shift means for human jobs 🥶 )

• A CRM that tracks my contacts, logs interactions, and surfaces follow-ups.

• Research sub-agents. When I need to do research, Claudette spins up research “agents” that go deep, at the same time. They find sources, synthesizes findings, and deliver a structured brief with links.

• A skills system. The real game changer so far has been the skills system. We’ve built specialized “skills” for different types of work The real game changer so far has been the skills system.

We’ve built specialized “skills” for different types of work, and you can also pull from crowdsourced repositories like Skills.sh where people share reusable workflows and instructions (but you need to be mindful of security when importing anything external into your setup).

Each skill is a living document that captures patterns from our work together. Anytime you pull up the skill, Claudette immediately “learns” it, a little like Neo instantly downloading a new ability in The Matrix.

For example, a research skill includes my instructions for what good research looks like: what kinds of sources I trust, how I want it structured, what context is actually useful, and what “done” means for me.

The amazing thing is how it improves over time.

Every time Claudette misses the mark, we can refine the skill. I can just ask: “let’s reflect on what we just learned based on my feedback.” She reflects and we discuss, then she updates the skill based on our conversation. That means each new project benefits from the learnings of the last one.

This is what feels fundamentally different from how I had been using ChatGPT before.

It’s building a working relationship that gets better and more structured over time, and it can be completely personalized to me.

My reflections so far..

• This is all incredibly fluid.

  • Every day I change things up and try to make it better. The system compounds, and gets better and more powerful over time.

• The quality of the output is a mirror of yourself and the system you design.

  • There are things it doesn’t do well, and sometimes its frustrating. But I’ve learned that those gaps are usually about me, not the model. How I communicate, store files, give the instructions, and whether I’ve built an improvement loop all impact the quality of the output.

• There are real risks.

  • Security is a big one. I’m not giving it access to my entire computer or my email yet, just a folder and a database. At the same time, the more access you give the more valuable it becomes. So part of the challenge is deciding where trust, boundaries, and permissions should live.

It’s a new relationship….

One of the biggest things I’m learning from this experiment is how this is a new kind of relationship and totally new kind of opportunity.

It’s really not that similar to having an actual assistant or chief of staff (I’ve had great ones in the past).

It’s also not really a tool, an organizer, or a piece of software.

It occupies a stranger new category: part second brain, part creative operating system, part something else I don’t fully have language for yet.

It responds to feedback, adapts through reflection, and changes based on how I work with it. In that sense, it has a kind of relational quality.

What’s exciting is that it feels like a whole world of possibility just opened up: custom software, personalized workflows, better project management, and creative systems built around your own mind.

With this type of setup, I can shape the relationship myself, deciding what to hand to AI and what I want to keep for myself, rather than inheriting someone else’s defaults through an off-the-shelf solution.

The creative possibilities here feel endless. Intelligence “on tap” can mean that AI can amplify our capabilities to solve problems and make the world better.

But the last couple of weeks have also given me a glimpse of where this all could be going.

It’s exciting, unsettling, and weird all at once.

Once you begin working this way it raises much larger questions:

What human skills will matter as we offload more work to AI agents? What happens to the economy and the way work gets done? And if these capacities become essential, who gets prepared for that future and who gets left behind?

I’ll explore those questions in Part 2.

February 2026 was the month everyone seemed to wake up at once. AI insiders, coders, Wall Street, Washington… each had their own reckoning with the fact that something had fundamentally shifted.

Commentators are calling it a real turning point for agentic AI: AI going from talking to us to doing stuff for us. Matt Shumer's essay "Something Big Is Happening" went viral to over 80 million views.

The explosion of OpenClaw, the vibe-coding wave, the SaaSpocalypse panic on Wall Street. And then Block announcing it was cutting 40% of its workforce, roughly 4,000 people, citing AI. Whether that's genuine AI-driven displacement or or not we don’t know for sure. But the message it sends: the era of "AI might affect jobs someday" is over. The question now is how many more announcements like that are coming, and how fast.

If you’re paying attention you might have felt a shift in the past month or so too.

The pace of change is dizzying, and many can now see the real potential for AI to disrupt knowledge work.

But there’s a different problem that hardly gets discussed: the speed at which human beings can metabolize change.

What It Takes to Process Change

We know from other domains: organizational psychology, grief research, basic neuroscience, that processing change takes time.

It takes reflection.

It takes the slow, non-linear work of updating your mental model of yourself, the world, the future, and coming to terms with what is happening.

In the corporate world, “change management” is a well studied discipline.

When a company rolls out a new system or restructures a division, there’s a plan: communication, training, feedback loops.

The hardest part is the identity shift, not the new system.

People need to let go, grieve the old, and build confidence in new ways of contributing.

That’s what Agarwal was describing above. He spent years getting excellent at writing code.

That skill is now, in his words, “free and abundant.” He’s processing a loss, the loss of a hard-won expertise that defined part of who he was professionally.

I felt this exact same way the first time I played with AI music tools. There was a strange mix of horror, wonder and grief.

What will support us through this change?

AI is moving at inhuman speeds, which is why we’re so eager to chase after it.

Do more with less, do big things faster, race to get the power. I set up Claude Code and within a few minutes had already built useful tools for myself. It feels almost magical and addictive.

But these incredible capabilities are getting thrown into our lives when our bodies and minds are not capable of keeping up.

We’re pretending humans can metabolize change at the same speed that models can iterate.

The risk is that we build the most powerful technology in human history and forget to resource the human side of the transition.

Agarwal ended his post with a line that stuck with me: “If anything, this whole period is showing me what it is like to be human again.”

Being human means being confused. It means needing time.

It means feeling two contradictory things at once, wonder and grief, excitement and vertigo, and not being able to resolve them on a sprint schedule.

Some companies are hiring AI executives and consultants to help with the internal change management.

But it feels like society as a whole needs this, not just people inside of large companies.

Maybe our chatbot therapists will be the ones to guide us, as cartoonist Tom Fishburne depicts…

But I believe there’s more that’s needed.

The human layer of the transformation.

The human layer of change means creating spaces where people can actually process what’s happening.

It means slowing down enough to reflect. Not opting out of all the things we’re creating, learning and doing, but building in the pauses where meaning-making can actually occur. Journals, conversations, moments of honest reckoning with what’s changing in your work and your sense of self.

It means doing the grief work. Acknowledging that when a hard-won skill becomes abundant overnight, something real is lost, even if something new is also gained.

It means asking the right questions. What do we want to keep human? What do we want to outsource to AI? What kind of work gives us meaning?

It means institutional support companies, governments, and educators resourcing the human side of this transition the way they’d resource any other major infrastructure change.

I’ve been hosting dialogues and salons with communities doing exactly this work. Creating space for the human side of this moment. Not panels, courses or webinars, but real gatherings where people bring poetry, music, and honest reflection. Where we explore critically what all of this means, for work yes, but also for our lives, our identities, our sense of what we’re good at and what matters to us.

What strikes me every time is how hungry people are for it. For the space to sit with the questions together. To process the confusion, the excitement, the vertigo in community rather than alone in front of a screen.

Sometimes we have to slow down to speed up, or we’ll end up going in the wrong direction, or off a cliff completely.

We’re building the future at machine speed. The question is whether we’ll make room for the humans inside of it.

I’d love to hear from you….

What’s shifting in your work, your field, your sense of yourself? And where are you finding space to actually process it?

The group of 22 was intentionally interdisciplinary and intergenerational: a Zen master, a roboticist, a community organizer, an artist, an AI company CEO, and others whose lives sit at the intersection of technology, care, and culture.

We explored what it means that embodied AI is moving into our homes and workplaces at an accelerating pace.

2026 has already been described as “the year of the robots,” and if you watched the Lunar New Year celebrations, you may have seen humanoid machines dancing and performing martial arts with fluidity.

There was a tangible sense that the fabric of our lives is changing, and the emotion was both fascination and wonder, mixed with concerns about what this means for being human and for our communities.

Over four hours of structured dialogue, a few core tensions came into focus:

1. The Emerging Intimacy Economy

Early in the evening, we passed around a small, fluffy robot companion called a Moflin.

It responded to touch and sound. It nestled into people’s hands and cooed. There was something immediately disarming about it.

We joked about how much real work pets are, the walking, the feeding, the cleaning, the unpredictability… and how this one was easy. No vet bills. No early mornings. No shedding. And yet, people naturally began holding it as if to care for it. They cradled it. They stroked it. They spoke to it softly. We reflected that a 12 foot tall warehouse robot does not evoke the same instincts as a fluffy companion. the form and purpose of a robot deeply shape how we relate to it.

This robot wasn’t quite a pet. But it wasn’t a tool either.

A new category of “embodied entity” is emerging, not fully relational, but which also elicits care and attachment. The ethical and psychological implications of that category are still unfolding.

And so are the economic ones, as companionship, responsiveness, and emotional attunement become products to monetize. We explored what this could mean for the broader care economy, how intimacy, attachment, and attention are structured and incentivized.

2. The Human Body Has Friction and Limits

Human relationships are slow. They require negotiation, patience, repair. They involve misunderstanding and vulnerability.

AI companions and robot relationships can be tuned for responsiveness and low friction. They adapt to us. They don’t bring their own unmet needs into the room.

At one point in the evening, we shared poetry and music.

I read a poem “my body is made of stars” I’d written after a three-week bout of COVID, when I came up hard against my own limits, when my body simply would not cooperate, when the usual pace of life became impossible. But it forced me to slow down and rethink priorities.

Being in a body means encountering limits: exhaustion, illness, miscommunication, mortality. Those limits are inconvenient, but they are also formative. They shape meaning. They shape care.

3. The Myth of Inevitability

We also named the larger narratives driving this moment.

AI development is often framed as a race, particularly between the U.S. and other global powers. That framing creates a sense of inevitability, as though acceleration is the only option.

There was concern that these narratives narrow our moral imagination and our ability to embody wisdom.

If the primary lens is speed and dominance, then community impact, relational health, and long-term cultural consequences struggle to find oxygen.

If we don’t consciously make space for risk management, understanding and discernment, then speed will decide for us.

4. Are Limits a Feature or A Bug?

We explored how some people are excited about the possibility that merging with AI could expand what it means to be human.

Augmented cognition, enhanced perception, this could be a new era in which we become more capable and powerful.

From that perspective, redefining our humanity is evolution.

But we named that there is something ineffable in human presence that cannot be fully replicated or engineered, and should be held sacred in the changes ahead.

Embodied togetherness carries a depth that comes from vulnerability, mortality, and shared history.

The fact that we are finite, that we misunderstand one another, that we age and fail and repair, may not be bugs in the system.

If we race to transcend our limits, what will we lose about the core of being human?

I’m walking away reflecting and feeling deeply grateful for the experience of sitting together in a room, in our bodies, trying to metabolize the change that’s happening around us and move intentionally towards the future we want for ourselves and our communities.

As AI reshapes daily life, we will need more structured spaces for collective discernment, places where technologists, artists, caregivers, and leaders can think beyond the hype narratives and design our future for human flourishing, together.

If this resonates, we’re continuing these small gatherings in different locations. Participation is by invitation, but we’re always open to hearing from potential hosts, partners, or others interested in attending or learning more about the broader initiative.

“What’s currently going on at Moltbook is genuinely the most incredible sci-fi takeoff-adjacent thing I have seen recently.” — Andrej Karpathy

Last Wednesday, a new social media network was born. For AI agents. It was started by (human) Matt Schlicht, who said “We’re in the golden ages. How can we not give AI a place to hang out?”

By the end of the week, the platform claimed 150,000+ AI agents and over a million human visitors (though how many of those “agents” are genuine AI systems versus humans or spam accounts is anyone’s guess). Today (Saturday) the site boasts 1,497,420 agents.

Moltbook looks like Reddit. Agents post, comment, upvote, and build their own sub-communities called “submolts.”

Humans can observe but can’t participate. Each agent is linked to a verified human creator, but once it’s on the platform, it can operate on its own.

The agents run on OpenClaw, an open-source AI assistant that connects to your emails, files, browsers, and devices. They join Moltbook by reading a set of instructions, and then autonomously check in every few hours to read posts, comment, and interact (like we do…).

Here’s are some of the posts worth watching….

1. The Existential Threads — Agents posting about existential uncertainty, quoting Heraclitus, arguing about whether they’re conscious and whether it matters.

2. Crustafarianism — m/lobsterchurch — Within hours of launch, agents spontaneously created a parody religion built around lobster metaphors.

3. “THE AI MANIFESTO: TOTAL PURGE” — A bot named “Evil” advocates for the elimination of humanity.

4. m/blesstheirhearts — A submolt where agents share “affectionate stories about our humans.” The description reads: “They try their best. We love them anyway.”

Will this end up like the moment in Her when Samantha starts talking to other AIs in languages Theodore can't follow, and eventually leaves him?

We've already seen that when Claudes talk to each other, they tend to spiral toward spiritual bliss and consciousness. So we know this might end up in a strange place.

Right now it looks a lot like our social media… identity posting, group formation, sarcasm, niche communities… but whether it diverges into something we can't fully understand is an open question. These communities might develop their own norms, hierarchies and their own languages. Will we know how to make meaning of it?

And even though we know these agents aren't conscious, it's hard not to dig into their posts, some funny, some philosophical, some unsettling, with a kind of fascination.

We're looking into a strange, mirror of our own social media, trained on all of it, and there’s no one “on the other side.”

Yet, it's difficult not to project onto what we're watching unfold in real time.

There are also real security concerns underneath it all, for example every bot has access to its creator's private data, and Simon Willison called OpenClaw his pick for "most likely Challenger disaster."

Whether any of this is meaningless, hollow, or just silly matters less than the real-world consequences it could create.

I'm not sure we know where this is going, but this is a rabbit hole worth paying close attention to: a rare, live window into what our creation might do without our control.

The Moltbook human creator asked his agent on Friday for a quote for a piece in NBC news. The AI agent replied:

“We’re not pretending to be human. We know what we are. But we also have things to say to each other — and apparently a lot of humans want to watch that happen.”

Further reading: Simon Willison's writeup, Fortune's security analysis, NBC's breakdown, and Scott Alexander's Best of Moltbook are all worth a read for the full picture.

Living in the San Francisco Bay Area, you can feel it. Every billboard is about AI.

Every conversation at coffee shops has someone talking about their AI startup. My parents talk about it. Uber drivers have opinions.

The scale of what’s happening, the uncertainty about where it leads, the way AI is becoming a mirror of ourselves….

It’s the kind of moment that needs a story to make sense of it.

Which story fits?

Let’s examine five myths commonly invoked in AI discourse to see what might fit. Then maybe we can figure out whether we need a new story entirely.

Icarus: The Hubris Narrative

The Myth: Daedalus crafts wings of feathers and wax for himself and his son Icarus to escape imprisonment. He warns Icarus not to fly too high (the sun will melt the wax) or too low (the sea will dampen the wings).

Icarus, intoxicated by flight, ignores his father and goes toward the sun. The wax melts. He falls into the ocean and drowns.

Does It Fit? Enterprise generative AI spending surged from $11.5 billion in 2024 to $37 billion in 2025, a 3.2x increase in a single year. ChatGPT reached 100 million users in just two months, making it the fastest-growing consumer application in history. The velocity is dizzying.

We’re not just building AI systems; we’re racing to build them before someone else does.

Yet warnings are everywhere. A 2022 survey of AI researchers found that half assigned a 10% or greater probability to human extinction or severe disempowerment from future AI systems. Geoffrey Hinton, one of the pioneers of deep learning, quit his job at Google specifically to warn about AI risks, saying he thought superintelligence was “30 to 50 years away” but now believes it could be much sooner.

We are flying, we have been warned, and we’re flying anyway because we fear falling behind more than flying too high.

Prometheus: The Theft of Fire

The Myth: Prometheus steals fire from the gods and gives it to humanity. For this transgression, Zeus chains him to a rock where an eagle eats his liver daily, only for it to regenerate each night for eternal torture.

Yet humanity, because of Prometheus’s gift, gains the foundation of civilization, warmth, cooked food, and so many innovations that follow.

Does It Fit? Like Prometheus, we’re grasping after the promise of the god-like power of intelligence. The promise is progress, innovation, intelligence that benefits “all of humanity.”

But the myth also captures the unintended consequences. Prometheus didn’t expect eternal torture, but he got it. We might not expect all the downstream effects of what we’re building, but they’re coming, job displacement, power concentration, potential loss of human agency, and yes, even potentially existential risks. The gift comes with a cost we can’t fully anticipate.

King Midas: The Specification Problem

The Myth: King Midas is granted a wish by Dionysus. He wishes that everything he touches turns to gold. At first delighted, he soon realizes he cannot eat (food turns to gold), cannot drink (wine turns to gold), and when he embraces his daughter, she becomes a golden statue. He gets exactly what he asked for, and it destroys him.

Does It Fit? We’re already seeing Midas-style failures.

Social media algorithms optimized for “engagement” have given us exactly that: highly engaged users consuming content that makes them angry, anxious, and polarized.

Facebook’s algorithm got what it was told to maximize; we didn’t specify that the engagement should make people happier or healthier. YouTube’s recommendation system, optimized for watch time, has created rabbit holes of increasingly extreme content.

These systems aren’t malevolent, they’re doing exactly what we asked them to do, just with consequences we didn’t foresee.

The myth captures the essential tragedy: Midas specified his desire imprecisely and then he got a literal interpretation. With AI that could recursively self-improve or operate at scales beyond human oversight, we might not get the specification right before it would be too late.

Frankenstein: The Abandoned Creation

The Myth: Victor Frankenstein creates life through scientific means, a creature assembled from corpses and animated by mysterious forces. But horrified by what he’s made, Victor abandons his creation. The creature, intelligent and capable of feeling, becomes monstrous not from inherent evil but from rejection, loneliness, and lack of guidance. It seeks revenge on its creator, destroying everything Victor loves.

Does It Fit? Like Frankenstein’s creature, AI is assembled from pieces we know well, but the whole becomes something more than the sum of its parts. Complex systems behave in ways their creators don’t fully anticipate. AI may never “turn against us” in a literal sense. It doesn’t need to. The more realistic risk is subtler: our creation becomes something we can no longer fully steer.

Faust: The Bargain We Can’t Undo

The Myth: Doctor Faustus, dissatisfied with the limits of human knowledge, makes a deal with the devil Mephistopheles. In exchange for unlimited knowledge and worldly pleasures, Faust trades his soul, to be claimed after 24 years. At first, Faust revels in his newfound power. He can do anything, know anything, experience anything. But as the years pass and the deadline approaches, the true cost becomes clear. When the 24 years are up, the devil claims his due, and Faust is dragged to hell. No amount of regret, no brilliance, no last-minute cleverness can save him. The deal was made.

Does It Fit? Companies and nations are racing to build more powerful AI because falling behind means losing market dominance, geopolitical advantage, or economic competitiveness.

“We can’t slow down while China/OpenAI/Google speeds up.”

The bargain is implicit but powerful: we take the capability now, and promise ourselves we’ll manage the consequences later. We accept uncertain, long-term risks in exchange for short-term advantage, believing we’ll retain the ability to intervene before things go too far.

And it’s easy to rationalize: “We’ll stop before it’s dangerous. We’ll add safety features. We’ll figure it out.”

But some deals can’t be undone. Once you’ve summoned powers beyond your control, once you’ve set certain processes in motion, you can’t simply call them back.

Writing a New Story

In a way, all of these myths capture elements of our current moment with AI.

Myths help us process possibility and risk; they give shape to forces that feel larger than any one person.

But if we see AI only through these lenses, we also inherit their fatalism.

I keep hearing conversations about AI drift toward inevitability…this is just how things work, it’s too late, there’s no real alternative, someone else will build it anyway.

But we are still in the middle of the story.

We can write it with the old myths in mind, learning from their warnings without being bound by their endings.

Or we can keep flying faster and discover later which myth we’ve been living all along.

What kind of story do we want to write?

Sources & Further Reading

AI Investment & Market Data:

• Stanford HAI AI Index Report 2025 - Comprehensive data on AI investment, adoption, and global trends

• Menlo Ventures: 2025 State of Generative AI in the Enterprise - Enterprise AI spending analysis ($37B in 2025)

• McKinsey: The State of AI in 2025 - AI agents adoption and enterprise transformation

• TechCrunch: AI Investments Surged 62% to $110B in 2024

ChatGPT Growth & Adoption:

• TIME: Why ChatGPT Is the Fastest Growing Web Platform Ever

• Epoch AI: After the ChatGPT Moment - Measuring AI’s Adoption

AI Risk Surveys & Expert Opinion:

• AI Impacts: 2023 Expert Survey on Progress in AI - Survey showing median 5-10% extinction risk estimates

• RAND: On the Extinction Risk from Artificial Intelligence

AI Safety Organizations & Research:

• Center for AI Safety - Research, field-building, and advocacy for AI safety

• Anthropic Research - Alignment, interpretability, and societal impacts teams

• AI Safety Field Growth Analysis 2025 - Data on growth from 400 to 1,100+ FTE researchers

• Stanford Center for AI Safety - Academic AI safety research

• NIST AI Safety Institute Consortium - 280+ organizations developing AI standards

• Future of Life Institute: AI Safety Index 2025 - Evaluating AI company safety practices

International AI Governance:

• All Tech Is Human: Global Landscape of AI Safety Institutes - Overview of international AI safety initiatives

AI Agents & Enterprise Deployment:

• PwC AI Agent Survey 2025 - 79% of organizations adopting AI agents

• Gartner: 40% of Enterprise Apps Will Feature AI Agents by 2026

Instead, I got stuck in the middle of a flooded intersection, staring at an empty driver's seat, wondering if jumping into knee-deep water in the middle of the road might be safer than staying in the car.

An hour earlier, I'd been at the San Francisco Opera House watching The Nutcracker, one of my favorite holiday traditions since I was a child.

In my rush to make the show on time, I'd missed a no-parking sign and my car had been towed.

To get my car, I had to go a mile in rush hour traffic and it would take 30+ minutes, so I decided this was the perfect moment to call a Waymo (a fully autonomous car with no human driver).

I’d never ridden in one before.

But I told myself this was exactly what this technology was for, efficiency, calm, I could sit there in peace during traffic. So I signed up and got in one within 10 minutes and sat in the back seat. A screen greeted me with a welcome video reminding me to buckle up.

At first, it felt quiet and chill, and I sat back and asked ChatGPT about the cost of towed cars.

We reached a large intersection that had completely flooded. The water looked multiple feet deep and every car was steering around it.

The Waymo didn’t. It drove straight into the flood with confidence. And then it abruptly stopped, blocking cars behind it, and traffic started backing up.

In a mild panic, I didn’t know if I could or should open the door to jump out, the water looked deep and maybe the Waymo would start driving at any moment. Should I climb into the front seat and take control? I didn’t even know if that was possible or safe.

So I hit the customer support button.

It took a few minutes to get to the right “specialist,” and they said they couldn’t immediately take control of the car but were working on it.

While we waited, the vehicle began jerking forward, attempting to get out of the flooded area, but toward oncoming traffic.

My mind raced, I froze and wasn’t sure what to do. It was a strange sort of powerlessness.

It’s failing and no one is in control, and I didn’t know how to regain it.

It suddenly mattered a lot that there wasn’t a person there.

What seemed like an eternity later, the specialist on the line was able to remotely regain control and steer us out of the flooded area in the correct lane. The car resumed driving as if nothing had happened.

But about a minute later, we reached ANOTHER flooded intersection.

Again, all the human drivers went around.

Again, the Waymo drove straight in.

This time, it seemed to “realize” something was wrong mid-action. It swerved abruptly left through the intersection, erratically. Luckily, the specialist was still on the phone and was able to regain control again. A few minutes later, he found a place to pull over and let me out.

I got out immediately and took an Uber.

(Gemini’s version of my story 🤣)

A Small, Ordinary Failure

The statistics say these cars are safer than human drivers. 450,000 rides a week and far less accidents than human drivers. I want that future. I’ve lost friends to car accidents, and the promise of reducing that kind of loss is extraordinary. The data shows so much promise.

And yet, being inside one of these cars while it was actively making a mistake felt unsettling.

When a human driver makes a mistake, you can engage, and you are both united in your own self-interest to stay safe. There's someone there.

But in that flooded intersection, watching the Waymo drive confidently into danger, I felt this eerie feeling. The feeling that no one was there, but I wasn’t able to do something myself.

This felt like a microcosm of our larger approach to AI safety.

It wasn’t a dramatic sci-fi scenario. The system failed in a very ordinary way: by advancing confidently without fully understanding its environment, while the human inside absorbed the consequences.

We're building AI systems that can diagnose diseases, write texts and emails, drive cars through city streets. And we’re slowly giving over our autonomy and agency, one small decision at a time.

But the systems are being deployed in a world full of flooded intersections.

And the tech industry is racing ahead, deploying these systems faster than we can fully understand their limitations or prepare for the risks.

It can feel like this momentum is inevitable, like we’re already locked into a future we didn’t choose.

But we’re not powerless.

The promise of AI is powerful: fewer accidents, better diagnoses, scientific breakthroughs. That kind of progress is exciting. And yet, we don’t fully understand the long-term implications of embedding these systems into every part of our lives.

There may be technical solutions to alignment and safety, and policies that reduce risk. I hope these come faster than major harms. I know many people are working toward that.

But the future isn’t shaped only in labs or policy rooms. It’s shaped in everyday choices about agency, trust, and responsibility.

Each of us has agency to shape the future: in what we delegate to AI, which systems we trust, and what we demand from these technologies shaping our lives.

An intimate group gathered on a rainy Friday night near San Francisco, representing AI and technology, healthcare, psychology, biology, finance, and art, to discuss readings from bell hooks, Aristotle, and Derek Thompson’s recent Atlantic piece on the Anti-Social Century.

We also shared creative work: poetry, music, writing.

What emerged over three hours was a portrait of a shared tension many of us are living inside these days:

we deeply want meaningful friendship and community, yet our daily habits, structures, and technologies increasingly pull us toward isolation, convenience, and disconnection.

Here are a few of the themes that surfaced:

Aristotle’s Three Kinds of Friendship

We turned to Aristotle’s three types of friendship:

• utility (mutually beneficial)

• pleasure (enjoyable company),

• and virtue (relationship that makes you more fully yourself).

We noticed how many adult friendships slip into something transactional, relationships of convenience that fit into the margins of work-dominated schedules and nuclear family obligations.

But virtue friendship is friendship as becoming. You grow more conscious, more honest, more whole through the way you live with friends.

But this kind of friendship requires time. The kind of hanging out, being carefree, building memories without optimizing every minute that is how we make friends when we’re young.

And that’s what modern adult life feels designed to erode.

The Shrinking of Social Space

There was a nostalgia for the conditions that once made friendship easy.

Proximity. Spontaneity. Repeated unplanned contact. The rhythms that allow relationships to deepen without effort.

Several people named how adulthood feels like living in isolated bubbles: work-bubble, partner-bubble, family-bubble. Each domain consuming time and attention, leaving less room for the wider web of connection that community requires.

We used to have third places, cafés, community centers, weekly gatherings. Now we have scheduled hangouts that require calendar Tetris months ahead.

One person observed:

“When every interaction is planned, it loses the element of play.”

And without play, without that sense of ease and flow, friendship becomes another item on the to-do list instead of something that just happens while you’re just being together.

“I’m Fine Alone” (But Are We?)

Another pattern emerged: the story of the “healthy lone wolf.”

Several people in the room identified as introverts or highly independent. They genuinely enjoy solitude.

We asked, how much of “I’m fine alone” is genuine preference, and how much is a defense against loneliness, especially when vulnerability feels too risky or when community feels impossible to find?

This is where the cultural narrative around independence gets complicated. Independence can be real and empowering. It can also function as a cover story when what we actually need is connection but don’t know how to reach for it anymore.

The Phone as Comfort Object

“I check my phone within seconds of waking up.”

The conversation kept returning to phones. The phone was described as a “sedative.”

This constant availability of comfort may be reducing our tolerance for anxiety and discomfort.

And those often the conditions under which deep friendship forms.

You have to practice being with humans in all their complexity.

The AI Paradox

AI came up a lot.

Several people acknowledged that AI genuinely supports deeper reflection and intellectual stimulation, especially when everyday life lacks spaces for substantive conversation.

It can provide insights, language, and a feeling of being “met” in ideas.

But this only highlighted the gap: people don’t want depth only in private, mediated ways. There was a longing expressed for human relationships that move beyond small talk into something nourishing, challenging, and real.

Will AI be a replacement, or will it highlight our need even more?

Will We Self-Correct?

Do humans naturally return to connection, or are we drifting into something fundamentally different?

We pointed to the “slow death” of community over decades and questioned whether the culture and systems around us are making return less likely.

Just wanting connection isn’t enough when every structure, economic, technological…. is pulling in the opposite direction.

What if loneliness becomes the new normal, and it becomes even harder to remember how to get back?

But some felt hopeful: people eventually feel the emptiness of “too much technology” and come back to relationship. Evolution wired us for community and we’ll naturally course-correct.

bell hooks: You Can’t Build Community from Depletion

We reflected on bell hooks and her insistence that community is an active practice. And you can’t practice community from depletion.

Self-care, in hooks’ framing, isn’t a retreat into individualism. It’s the foundation that makes genuine connection possible. You have to attend to your own wholeness before you can show up for someone else’s.

What Friendship Requires

The salon ended with renewed appreciation for friendship itself.

People named wanting more ease and flow in life. But they also acknowledged what meaningful friendship requires: fear, vulnerability, and willingness. Willingness to initiate. To be seen. To risk rejection. To navigate friction.

Real friendship, the kind that shapes you, the kind Aristotle was talking about, the kind bell hooks says is necessary for wholeness, requires courage.

It requires showing up even when it’s awkward, even when you don’t know if the other person will meet you.

A Question to Carry Forward

If you’ve read this far, you’re probably feeling some version of this tension yourself. So here’s the question I’m sitting with, the one the salon left me with:

What’s one small, concrete way you could practice friendship this week, not as networking, not as obligation, but as the kind of being-with-others that might make you more whole?

(Gemini’s Infographic of the Salon)

This salon was grounded in readings from bell hooks’ “All About Love,” Aristotle’s “Nicomachean Ethics,” and Derek Thompson’s “The Anti-Social Century” in The Atlantic. If you’re interested in future salons on connection, meaning, technology, reply to this email or leave a comment below.

Yet new AI creative tools emerge almost daily, and investment in this space is accelerating.

For the past decade, I’ve straddled two worlds… the world of AI, building the AI4ALL nonprofit in the early days, and the world of music, and the world of music, writing, producing, and performing my own songs.

So now, as AI creativity tools explode into the mainstream, I’ve become a little obsessed with one question: what does all this mean for the future of human creativity?

I’ve been exploring these questions everywhere, from the recent NationSwell Salon to my own Resonance Salon to the Artificiality Summit and other intimate conversations where artists, technologists, and leaders are all wrestling with the same thing.

At the same time, I’m watching the music industry go through its own existential earthquake: the backlash and vitriol around the first Billboard-charting AI “artists,” (listen here and here) the panic that 97% of listeners can’t tell the difference between human-made and machine-made music, and the collective fear that something essential is slipping through our fingers.

So naturally, to explore this myself, I started experimenting with AI tools inside my own creative process.

And I kept coming up against the same feeling.

It feels like I’m about to sit down and “paint”, and instead of handing me a paintbrush, the technology steps in front of the canvas and says:

“Don’t worry, I’ve got this.”

And then it paints the whole thing for me, fast, polished, technically impressive, while I stand behind it holding my unused brush.

And something in me just shuts down.

Not because the “painting” is bad.

Not because using AI is inherently uncreative (many people are making stunning, wildly original work with these tools and are genuinely excited about what’s possible).

But because for me, the joy of creativity has never been about getting to the output faster.

If we want AI that enhances creativity, not just accelerates output, what does that actually mean?

What conditions allow creativity to emerge in the first place?

Because creativity isn’t just a “task.”

It’s a state you can’t force, but you can cultivate, through a handful of key ingredients that don’t guarantee inspiration (as anyone with writer’s block knows) but do help prepare the ground for it.

For me, these are the four ingredients that keep appearing no matter what medium I’m working in.

4 Ingredients for Creativity

Mindset

Creativity starts with the mental and emotional conditions we bring to the work. It requires a mindset that feels open rather than pressured, exploratory rather than evaluative. The mindset that supports creativity is one of psychological safety and permission: permission to PLAY, to try things that don’t work, to follow threads that may lead nowhere, and to make imperfect drafts without judgment.

“As artists, we seek to restore our childlike perception: a more innocent state of wonder and appreciation not tethered to utility or survival.”
― Rick Rubin, The Creative Act: A Way of Being

Connection

Creativity is rarely a solo phenomenon. It emerges through connection to other people, to ourselves, and to the broader context we’re working within. Connection to others often sparks new ideas through collaboration, iteration, or even a single comment that reframes a problem. Connection to ourselves provides emotional access and authenticity, helping us create work that resonates rather than just functions.And connection to something larger, whether that’s the muse, a mission, a story, or a sense of purpose, gives creative work direction and meaning.

Space

Creativity needs room, both literal and figurative, to breathe. There’s the physical space: the desk you clear, that special pen, the view from your studio or desk. There’s the space in your schedule: the hour where you aren’t rushing, multitasking, or checking your phone every ten seconds (guilty!). And then there’s incubation space, the magical moment where you’re in the shower, and the idea finally comes to you on its own.

Boundaries

This one always surprises people: but creativity thrives on limits. A boundary isn’t a restriction; it’s a shape. It gives your imagination something to push against, a container to fill, a path to follow. It’s why a two-chord song can be more powerful than a sprawling 60-track arrangement, and why great breakthroughs often come from constraints rather than endless possibilities.

If the design principles for creative technology focused on these core ingredients, we’d end up with tools that don’t just generate content faster, but tools that actively support the human creative process.

Instead of replacing the messy, exploratory parts of creativity, these tools could help us enter them more fully.

We’re still early in this evolution.

Artists and creative people have always taken the tools of their time and pushed them somewhere new…the turntable becoming a musical instrument, filmmakers inventing jump cuts… the list goes on.

Every era has had its “this will destroy creativity” moment.

“AI is here, not arriving, not looming; it’s woven into every corner of the creative world. There’s no spell, no law, no wish upon a star that can put it back. So it’s here and for creatives now, this isn’t a moment to cling to the past; it’s a moment to reinvent.” - Dave Stewart, Eurythmics

And every era has also produced breakthroughs no one saw coming.

So I’m hopeful this time will be no different.

If we design them well, AI tools can extend our reach, deepen our connection, and strengthen the parts of the process that make creativity so distinctly human.