Old Ilya email on AGI openness resurfaces as a meme about OpenAI’s shift
tszzl · x · 2026-07-27
A meme post resurrects an old Ilya Sutskever email arguing that as AI gets closer to being built, it may make sense to be less open.
The screenshot contrasts that position with OpenAI’s original “open” framing:
- the hope that everyone would benefit from AI after it was built
- but the suggestion that sharing all the science is not necessarily the right move in the short or medium term
- especially if open sourcing makes it easier for someone with huge compute access to build an unsafe system in a hard-takeoff scenario
The post is essentially a culture-war joke about how early AGI idealism can turn into proprietary caution once a lab is leading.
More from AGI Musings
- Altman says the singularity is here; Musk says money may not matter by 2036 — davidpattersonx · 2026-07-27
- X is described as a machine that makes people feel uniquely sane — nabla_theta · 2026-07-27
- YC says physical AI now feels like GPT-2 and DALL·E did a few years ago — ycombinator · 2026-07-27
- How do you design products for neck, face, and pocket computers with superintelligence? — RileyRalmuto · 2026-07-27
- Law professor argues closed-book exams may be fairer in the AI era — hoofnagle · 2026-07-27
- AI for biology splits into “AI scientists” and harder problems that need wet-lab validation — rishabh16_ · 2026-07-27