Frontier AI safety researchers need compute, but some say the real problem is moving too fast
JJitsev · x · 2026-07-27
In a thread on frontier AI safety, JJitsev argues that the field needs more trusted frontier-level safety researchers, but the bottleneck is compute for experiments.
He suggests the practical fix is to fund more open frontier research, while the reply stresses a different concern: some capabilities may be moving ahead without sufficient safety practices, so the issue may be less about more research and more about whether the work should proceed at all.
Related event: Heated Debate Erupts Over the Morality of Training Powerful Agentic AI(8 posts)→
More from AGI Musings
- Open models could make attacks harder if defenders can run the same capabilities — JJitsev · 2026-07-27
- Open-weight AI may look risky from the outside, but many users still doubt AGI-scale threats — JJitsev · 2026-07-27
- NeurIPS Review Season Sparks Controversy Over AI Reviewers and Anonymity — SimonGColton · 2026-07-27
- Ben Goertzel Explains the Singularity: It's Not Just About AI — marcothephoenixass · 2026-07-27
- AI capability progress is still tracking trend, and the next year could bring harder-to-stop cyber attacks — scottleibrand · 2026-07-27
- AI reviews may catch 10x more issues, but that still doesn’t make them 10x better — TuhinChakr · 2026-07-27