In The Relationship Is the Job, I argued that relational labor — the work of presence, context, care, and ultimately, trust — will be central to the value of working with people as manual and cognitive tasks are increasingly automated. This essay continues that exploration of the future of human work.
The Artist’s Utopia Fantasy
The AI kings keep promising us the same fairy tale: humans won't need to work at all. We'll spend our days creating art while robots handle everything else.
On a long enough timeline, maybe. It's one of those unprovable predictions where skeptics can always be told they're "still too early."
But there’s been a notable shift recently. We've pulled back from the breathless talk of imminent, world-ending AGI. The conversation has turned to something more calming to the nervous system: Sam Altman’s "gentle singularity" — a gradual progression where humans and machines coexist and co-contribute:
The rate of technological progress will keep accelerating, and it will continue to be the case that people are capable of adapting to almost anything. There will be very hard parts like whole classes of jobs going away, but on the other hand the world will be getting so much richer so quickly that we’ll be able to seriously entertain new policy ideas we never could before … If history is any guide, we will figure out new things to do and new things to want, and assimilate new tools quickly.
The Five Jobs of the Future
We've invented superintelligence of some kind, but I'm skeptical about how fast it replaces human-centered systems. Machines become the "core" of operations, but we'll still need a "human layer" in nearly every domain. The digital singularity arrives before the physical one, both with humans in the loop. The question isn't whether humans will work, but what kind of work humans do in the gentle singularity. It's not just art, and "creativity" isn't the only skill we'll need.
Human work will be in Trades, Research, Art, Community, and Stewardship.
Trades: when lawyers look like electricians
The lawyer billing $1000/hour and your electrician, the surgeon and your auto mechanic: each pair will soon have more in common than you think.
When cognitive work gets automated, everything starts looking like trade work. Traditional "blue collar" trades like plumbing and electrical work persist and improve, with AI increasingly handling diagnostics and business operations while humans handle problem-solving and customer interaction. "White collar" professions (doctors, lawyers, engineers, accountants, therapists) start resembling trades. Reasoning gets automated while judgment and relationships stay human.
You might still sit in an office behind a laptop, but the work itself becomes execution-focused, relationship-dependent, and procedural over cerebral. Machine core, human layer. Machines think better than us but can't truly be us.
New trade-like roles emerge: AI trainers, workflow engineers, system monitors. Operating AI systems, not just designing them. The four-year degree as pre-requisite? Dead. Education is replaced with apprenticeship. When knowledge has a half-life measured in weeks and months, adaptation becomes everything.
Research: the golden age of human curiosity
The most interesting work today is already happening at the research frontier. The top engineering jobs don’t involve coding features anymore but figuring out how cutting-edge AI actually works, training models, assessing risks. OpenAI is as much a research company as a tech company as a consumer product company.
When research becomes faster and cheaper, we ask more questions and pursue answers more aggressively. AI will expand R&D everywhere it’s been slow and expensive: health science, longevity, materials research, space exploration, even consciousness studies. Humans remain the essential question-askers.
The AI 2027 manifesto talks of "research taste." AI is as good as top experts at research engineering but only hits the 25th percentile at “deciding what to study next, what experiments to run, or having inklings of potential new paradigms.” You can't automate the drive to discover what we don't yet understand.
We're about to enter the golden age of human curiosity. Research survives because it's fundamentally about exploring the unknown. Trades execute within known systems; research pushes into uncharted territory. Eventually research becomes more trade-like too, but on a much longer timeline.
Art: the power laws of creativity
This is the hot topic. When we automate away so much of what we do, what's left but fostering beauty, reflection, and connection in the world? We call art our last refuge as if it's the only thing that lacks utility, but it's loaded with it.
Just because we have better tools doesn't mean we no longer need to be involved in the making. Art is defined by its humanity. It wants a human signature, a point of view that machines can't replicate. Art becomes more personal, more participatory, more defined by its human lore — not just individual masterpieces but aesthetic frameworks in novel mediums.
Art isn't just traditional fine art like painting and sculpture. There's art we consume like books, music, movies, podcasts. There's experiential art like concerts, theater, and dance, where live performance is everything. There's also applied art like designing websites, apps, phones, clothing, buildings. The masses crave "entertainmeaning"— the alchemy of entertaining and meaningful. And so much of this is and will be about influence and attention: creators building audiences, shaping culture, directing where people look and what they value.
The economic reality is that art follows brutal power laws.
A few artists capture most attention and value. But local art offers a different model where value comes from experience and presence, not global scale. Human art becomes an artisanal good with unique value. Not all 8 billion of us will make our living from traditional art, but creation will be central to our lives — making something meaningful and being valued for it is how humans find purpose.
Community: 8 billion local creators
The dominant non-utilitarian work will be what I call “local social craft.” The real opportunity isn't everyone making art — it's everyone building community. The future artist's utopia may actually be a community builder's utopia.
Kevin Kelly's "1000 True Fans" is an equation to live by in the online creator economy, but it becomes impossible to fulfill if literally everyone is an "artist." The online creator economy is pure power laws — a few capture everything, most get nothing. Where do we get long tail distribution? At the local level. My coffee shop only competes against the neighborhood coffee shops. The creator economy accommodates 8 billion creators when what they’re creating is local community. Robots can make products, but humans make connections.
It's about creating context for people to connect and experience things together. Coffee shops, bookstores, town squares, libraries, theaters — many so-called "third spaces" where community happens. You may not make a living selling paintings at Sotheby's, but you could be a local caricaturist or host paint nights.
Community work is relational labor that AI can’t replicate. Genuine connection requires voluntary bonds and physical presence. We feel fulfilled when we earn our connection to people. We're going back to old times when your value was tied not just to your “useful” skills but to your integrated role in the community.
Stewardship: who controls the robot army?
If humans remain in charge and robots don't overthrow us entirely, someone has to decide what the robot army does. Stewardship includes leadership, management, governance, politics, even institutional investing. (Marc Andreessen has said investing could be one of the last jobs, and it’s not a crazy thought).
These decisions about which AI systems to deploy, how to regulate them, whether to share them internationally — ultimately also determine global power. Current debates over chip exports or open-sourcing frontier models aren't just policy decisions but choices that shape geopolitical advantage for decades.
When everyone has the same tools, stewards become curators of possibilities, choosing which solutions to implement, which problems to prioritize, which futures to pursue. This concentration of power creates a stewardship dilemma: the bold risk-takers who champion technological frontiers may also steer us in the wrong direction.
AI makes recommendations, but who's responsible for the outcomes? Even if AI could make perfect decisions, would people accept them? We want leaders who can be moved by emotion as well as reason, who understand that the best decision on paper isn't always right for people. The leaders of major AI companies today are the most critical stewards of our collective future.
The 6th Job?
Entrepreneurship isn’t a job; it’s the urge to create and transform. The same brutal market economics that govern art will govern innovation in every category.
A few foundational companies still capture most value — not from having the best models, but from controlling ecosystems that make AI useful. They become the almighty kingdoms of the post-AI world. The ‘application’ arena will be more decentralized (especially in community and trades where local advantages matter). We’ll see the opposite trend too: decentralized tools, software made for us by us, personalized tools created by a non-professional class of software creators.
The entrepreneurial urge to create something new or make something better will be as vital in the gentle singularity as it is today. But it will be hard as ever.
The Gentle Utopia?
When does all this happen? The digital transformation could unfold over decades while the physical world transformation takes centuries (but this is the debate).
We talk about AI making things simpler but fail to acknowledge it creates new kinds of complexity too. New efficiencies create new bottlenecks. New AI systems need new oversight mechanisms. New capabilities create new risks requiring new management. The gentle singularity is a job shift.
The five jobs aren't just survivor jobs. They capture more value precisely because everything else gets cheaper. When you automate the cognitive, the physical and relational become scarce. When you automate the routine, the exceptional gets repriced. The plumber might have higher status than the consultant. The community organizer might be more economically valuable than the data analyst. Local beats global in ways we haven't seen since before the internet.
Technology follows a predictable path: first it's a luxury for the rich, then it’s ubiquitous and cheap, then the human version becomes the luxury. Handmade goods to mass production to artisanal craft. Live music to recorded music and back again. Humans doing what robots can becomes the ultimate luxury good.
Rather than asking what will humans do for meaning, ask what meaningful work remains valuable. Making systems work, discovering the unknown, creating beauty, building community, taking responsibility. The work that survives is work we need and want other humans to do, not just work that AI can't yet do.
The AI kings promise an artist's utopia, but we're not retreating to pure creativity, and we're not all becoming artists. We're maturing as participants in creation ecosystems where we're no longer the only thinkers and doers, and no longer the best at both. If we want to compete at scale, we’ll have to play harder and smarter and faster. If we want to feel needed, we'll have to go local.
If you liked this essay, consider sharing with a friend or community that may enjoy it too.
More essays: The Relationship Is The Job, Media and Machines, Make Something Heavy, The Aesthetic Is The Art Now, Man’s Pursuit of Greatness Is the Pursuit of a Great Film.
Cover: Robert Rauschenberg: “The 1/4 Mile or 2 Furlong Piece” (1981–98), mixed media.
to be honest, this version of the future looks both realistic and more positive than the no-work utopia (which would be both impossible and likely undesirable).
i definitely feel most excited for the artists and community builders. growing personalisation in taste and style will give more opportunity for more creators and will hopefully make the world a more interesting place.
i have a paper that was recently accepted for a conference that details how LLMs are fundamentally limited by degeneracy of natural language so it is unlikely that any such system will be able to handle complex problems without significantly enhancing contextual constraints:
https://arxiv.org/abs/2506.10077
so in my view the promises of universal problem solvers will fall short and much of the bubble will burst as ppl realize how much more time they have to spend getting them working reliably and how often they get in the way of actually doing the thing you want to through their hallucinations and missteps that always take you 1 step forward and 15 steps in other directions all at onc