AI abundance will not become shared prosperity by itself; ownership and distribution structures decide who benefits. Anthropic's proposed 5% share commitment to a public-benefit fund, subject to stockholder approval, is a significant step on the Third Crucible's ownership question in The Three Crucibles. Its early timing may anticipate demands to share AI's gains, build goodwill before Anthropic's IPO or pressure competitors to follow. The fund remains company-controlled, so lasting distribution still needs public policy.
In my book The Three Crucibles, I argued that the people who build AI will not voluntarily share the abundance it creates. My reason was structural rather than moral: the logic of capital accumulation has no "share with everyone" function.
On 18 October, Anthropic announced a commitment of 5% of its shares to a public-benefit fund, subject to stockholder approval. I think it is a significant concrete step on the question my book ends with, though not a solution. AI abundance will not become shared prosperity by itself, because ownership and distribution structures decide who benefits.
The fork in the Third Crucible
My book describes three tests of the AI transition: The Correction, The Void and The Alignment. The third is a fork between abundance for all and techno-feudalism, where a few own the engines and everyone else rents intelligence.
My route to abundance combines distributed AI ownership, universal basic income and a tax on AI labor. The income is a Citizen's Dividend, which I frame as a royalty on humanity's accumulated knowledge rather than charity.
Shares instead of a donations budget
Most corporate generosity sits in a budget that gets cut in hard times. Anthropic's fund would hold equity that gains value as the company grows, and the board may add 2% for each 50% rise in value above a baseline. I want every AI company to copy that link to the upside.
Where my warning still holds
The fund is a wholly owned Anthropic LLC, not a charity. Direct payments to specific groups are one possible use, and governance details are still to come.
So this remains a voluntary, company-controlled commitment, the arrangement my book warns against relying on. Who chooses the beneficiaries, who checks the results, and when does share value reach people whose work AI changes?
*Metropolis* (1927) ends on the steps of a cathedral. Freder, son of the man who rules the city, joins his father's hand with the hand of the workers' foreman, because the film's motto says the mediator between head and hands must be the heart. It is a moving scene. However, nothing in it changes who owns the city or its machines, so the peace rests on the master's goodwill. I see the Third Crucible in that frame. A handshake can ease the conflict, but who shares in what the machines produce depends on ownership and the rules around it.
Why so soon?
I think Anthropic expects the demand for a share of AI's gains to arrive sooner rather than later. Its announcement already says "bolder approaches may be needed" for workers facing instability. Whoever has a structure ready when it arrives starts in pole position, with trust, goodwill and a structural head start.
The cynic in me notes the calendar. Anthropic confidentially submitted a draft IPO filing in June, then filed publicly eight days after announcing the fund. Goodwill is worth a lot before an IPO.
The optimist in me hopes Anthropic genuinely wants to be a force for good for all humanity, steering away from techno-feudalism and pressuring competitors to follow.
What would change my mind
If competitors follow with equity commitments, independent oversight and payments that reach people, I will gladly soften my warning. Even then, a fork this large cannot rest on corporate goodwill, whatever the motive. Anthropic has shown that public benefit can be written into ownership. Next, I want it in public policy, beyond any single board. I'm watching closely.
