What I learnt co-leading an AI Safety bootcamp for legal and governance practit
A LessWrong post about an AI safety bootcamp for legal and governance practitioners sparks intense skepticism on Hacker News regarding the validity of AI safety subcultures.
skippable The article text is unavailable and the comment thread consists largely of generalized skepticism and personal attacks against AI safety culture rather than technical debate.
summary
The LessWrong post shares insights from co-leading an AI safety bootcamp tailored for legal and governance professionals. Because the full text is unavailable, the story is defined by the comment thread, where HN users overwhelmingly dismiss AI safety bootcamps as insular consultant grifts or sci-fi moral panics. A minority of commenters argue that critics should engage with object-level arguments about machine learning scaling instead of relying on personal attacks.
key points
- The original post details experiences teaching AI safety frameworks to legal and governance practitioners.
- Commenters strongly criticize AI safety subcultures, comparing them to corporate consultants and hucksters.
- Defenders of the field note that critics frequently resort to ad hominem attacks rather than addressing core machine learning scaling risks.
who should read it
Anyone tracking the cultural sociology and public reception of the AI safety and governance movement.
discussion hostile
consensus
Commenters widely agree that AI safety bootcamps and broad governance frameworks are an insular grift detached from concrete engineering realities.
pushback
A minority argues that critics avoid engaging with the actual object-level arguments about machine learning scaling by relying on personal attacks.
It seems like to me it would be more effective to focus on the object level aspects of their arguments rather than their weird lives? Their argument isn’t “look at my credentials, I’m an expert you must believe me” it’s that “ML research is largely math and science engineering, as AI gets better at math and science it can use that to speed up the cycle at which point something bad could happen”.
These Bay Area people dedicated their lives to AI risk often before AlexNet even showed deep learning could scale, by definition they are eccentric people.
At the same time, people like the AI 2027 …
It cuts through the hostility to refocus the conversation on whether core arguments about machine learning scaling and recursive self-improvement hold weight.
Generated by an LLM from the article and thread. Check the sources before relying on it.