Open AI labs aren't burning money — they're dismantling the 'intelligence tax'
thatgodzillaguy · reddit · 2026-09-28
- A Reddit long post argues Meta, Chinese labs, and open-assistant builders are running a margin-destruction operation, not burning money for reach or open-source idealism.
- Core logic: if intelligence is priced, OpenAI and Anthropic collect a tax from every company on earth forever; if it's free, the surplus flows to those building on top of AI.
- So nearly every company shares the incentive to push intelligence toward zero cost — each dollar left in the model layer is value extracted from them.
- Evidence: open-weight usage has blown past closed APIs, and both labs expected to IPO this year kicked the can citing "safety" — healthy balance sheets rarely need a cover story.
More from AGI Musings
- EA organizations slammed for 'unhinged' hiring: 10-20 work trials over months — AaronBergman18 · 2026-09-28
- Peter J. Denning on tacit knowledge and what AI still can't capture — ArtificialOther · 2026-09-28
- Chesterman's AJIL essay "Silicon Sovereigns": AI, international law, and the tech-industrial complex — ProfChesterman · 2026-09-28
- AI law scholar Chesterman: don't waste a crisis to rein in self-governing AI — ProfChesterman · 2026-09-28
- Singapore proposes a UN Framework Convention on AI Safeguards at UNGA — ProfChesterman · 2026-09-28
- Banning kids from AI at school won't work — a three-layer AI literacy framework — yi111 · 2026-09-28