AI is absorbing the work; the question is what happens to income, purpose, and power. Twenty-four monthly essays — June 2024 to today — tracking the transition as it actually unfolded, by someone who called it early and is now living it.
One essay a month on where the post-labor economy is actually heading — capability, evidence, policy, and a live experiment in post-labor income. No hype, receipts included.
Follow Kenn on LinkedInIn 2024 I said AGI-class capability was ~18 months out, that displacement would hit before institutions moved, and that planning beats pain. The scorecard — and the playbook for the next two years.
Two years of essays arguing the post-labor economy is coming. This month I bet my own income on it: a portfolio of AI-run ventures, built from a beach town in Thailand. The newsletter is the experiment.
Two years into the transition, the hardest question isn't income — it's identity. What Star Trek got right about a society that works without jobs, and what it skipped.
Two years after I warned 'plan before the pain,' the pain is here — but it has a zip code, an age bracket, and an industry code. Averages are hiding the emergency.
Agents now have logins, budgets, and performance reviews. Five predictions for the year the org chart officially gets a non-human column.
Twelve months ago 'agent' meant a demo. Now it means a coworker with a login. A year-end accounting of what actually changed — capability, adoption, evidence, and the political vacuum where the response should be.
Gemini 3 and a new Claude Opus landed while half of finance debated an 'AI bubble.' Railway mania bankrupted investors — and still left the railways. Plan for the shift, whatever the market does.
Amazon cut 14,000 corporate roles with leadership pointing to AI-era restructuring. The displacement phase now has a corporate letterhead — and the euphemism era is ending.
Nvidia pledged up to $100B to OpenAI's buildout. Oracle booked hundreds of billions in AI compute contracts. The capex supercycle is, at bottom, a wager on replacing the world's largest expense line: wages.
Stanford economists found a 13% relative employment decline for young workers in AI-exposed jobs. MIT found 95% of enterprise AI pilots failing. GPT-5 shipped to everyone. All three are the same story.
Frontier labs are paying athlete money for a few hundred researchers while entry-level hiring stalls everywhere else. The labor market didn't tighten or loosen — it split.
Altman says we're past the event horizon and it feels surprisingly normal. New-grad hiring data says it feels less normal if you're 22. Both are true, and the gap between them is the whole story.
Anthropic's CEO went on record: AI could eliminate half of entry-level white-collar jobs within five years, with unemployment spiking to 10–20%. The most credible insider just said the number out loud.
'Before asking for more headcount, demonstrate why AI can't do the job.' A leaked CEO memo just made the post-labor hiring bar official — and a forecasting group published a scenario where 2027 is the pivotal year.
Manus put autonomous agents in the group chat, MCP became a de facto industry standard, and one image model turned the internet into Studio Ghibli. Creative work just joined the conversation.
OpenAI's Deep Research produces in half an hour what a research associate produces in a week. At the Paris summit, world leaders pivoted from 'safety' to 'opportunity' — with labor still the footnote.
A Chinese lab matched frontier reasoning for a fraction of the training cost, Nvidia shed half a trillion in market cap in a day, and OpenAI's first agent started clicking through websites. The economics of cognition just repriced.
OpenAI's o3 just scored 87.5% on a reasoning test designed to resist AI for a decade. Google shipped its 'agentic era.' The goalposts didn't move this time — they fell over.
America just held an election in the first year of the reasoning machine — and neither party had a labor-transition platform. Meanwhile, a quiet protocol release started wiring agents into everything.
Anthropic's CEO published the best optimistic case yet written for AI. The same month, his company taught Claude to use a computer like we do.
OpenAI's o1 thinks before it answers — PhD-level performance on science problems. And Sam Altman just told us, in writing, where he believes this goes.
Waymo is doing 100,000 paid rides a week. Figure just showed its second humanoid. The post-labor economy isn't only coming for keyboards.
The largest basic-income study in U.S. history just reported out. It neither proves the utopians right nor the skeptics. It proves we can learn — if we start early.
A 165-page memo from a 22-year-old just said the quiet part out loud: the people building this believe AGI is close. Our institutions are planning for a world that no longer exists.
The thinkers I argue with, borrow from, and recommend. Start with my essays above for the through-line; go here for the primary sources.
The frontier-lab case for what powerful AI is for — and the essay that takes the labor question seriously.
Real usage data on which occupations AI is augmenting vs automating.
'It is possible that we will have superintelligence in a few thousand days.'
The takeoff-feels-normal thesis, from inside the takeoff.
The 2024 memo that reset expectations about the decade ahead.
A concrete, contested scenario for the next capability leaps.
The largest U.S. basic-income experiment: what $1,000/month for three years actually did.
Brynjolfsson et al. — the 'canaries in the coal mine' entry-level employment evidence.
A parallel framework for the economics of a world after wages.
The institutional consensus view — useful as a baseline to argue with.
One essay a month on where the post-labor economy is actually heading — capability, evidence, policy, and a live experiment in post-labor income. No hype, receipts included.
Follow Kenn on LinkedIn