Ask HN: Who here would agree to replace parliaments with LLMs?
A Hacker News user asks whether people would agree to replace parliaments with large language models (LLMs), arguing the technology is sufficient and that current political systems, which they describe as gerontocratic and ineffective, cannot be worse.
Background
- This is an "Ask HN" post on Hacker News, the influential tech forum run by Y Combinator. The poster proposes replacing human parliaments with LLMs (large language models like GPT-4) — a deliberately provocative idea.
- LLMs are AI systems trained on text; they generate fluent but unreliable answers, can't reason, and have no accountability or will of their own.
- The argument "no way llms can be worse than what we have now" doesn't claim LLMs would be good — only that current legislatures are so bad (old, slow, corrupt) that a flawed AI couldn't be worse. "Gerontocratic retardation" refers to elderly leaders in many governments.
- This echoes a real fringe debate in AI circles: some technologists argue AI could make fairer, data-driven policy decisions, while critics point out LLMs learn from biased human data and cannot represent actual voters or take responsibility for outcomes.