AI Safety Debate: Unvetted Models Fail to Cause Catastrophe
sebkrier · x · 2026-07-18
Prominent figures in AI safety and alignment are debating the necessity of model vetting. Some previously argued that models like Mythos couldn't be released without patching every software vulnerability and applying strict restrictions.
However, critics point out that truly unvetted models have already leaked, and the predicted catastrophic harms have not materialized. Thus, using this as an excuse to suppress open-source releases via classifiers is untenable, prompting calls for the industry to reassess the real-world value of excessive safety audits based on reality.
Related event: Industry Reflects on AI Release Risks and Doomsday Predictions(4 posts)→
More from AGI Musings
- Claude Code skill uses 10 Markdown rules to make outputs ADHD-friendly — alex_verem · 2026-07-22
- AI Power Demand Exposes US Energy Gap, Urging Shift from Scarcity to Abundance — bradneuberg · 2026-07-22
- ControlAI CEO says an international ban on superintelligence is needed to avert extinction risk — zetalyrae · 2026-07-22
- Gary Marcus says LLMs still cannot really do math on their own — GaryMarcus · 2026-07-22
- Gary Marcus says LLM math skills are like knowing only a car’s engine size — GaryMarcus · 2026-07-22
- AI may make digital work infinitely leveraged while offline life gets more human — illscience · 2026-07-22