Recently I've been reading some various next-level arguments that in order to carry out their anti-regulatory agenda, Elon and Vance will remove many government bureaucrats and replace them with TPOT.
Various context:
Will 80,000 hours release an article (or substantially rework their career guide) by 2028 recommending to work for the United States government? Note that currently 80,000 hours does recommend government work, but specifically in the area of policy (https://archive.is/fGKHW). This does not count, as a YES resolution requires the argument to be based around efficiency, automation, replacement of redundant or noncooperating employees, or for the purpose of enabling an agenda of removing regulations and enabling progress.
1,000
3.00@aCivic Can you provide a link? Anyways, I don't think this counts, because that is in the field of policy, and not government efficacy.
@TonyGao I am very confused about what you mean by "the area of policy" vs "efficiency"; don't they go hand in hand? And working on improving decision-making in government is arguably also about efficiency.
Anyway, for current recommendations, see:
https://80000hours.org/problem-profiles/artificial-intelligence/#ai-governance-and-policy-work
https://80000hours.org/problem-profiles/improving-institutional-decision-making/
@aCivic The distinction is internal versus external: "efficiency" refers to government efficiency, not the efficiency of others outside of the government.
I suppose one way you can think about it is: are you as the employee deciding what should be done (policy), or merely making it easier for policies already set to be completed (efficiency).
For example, if I were to resolve this market, it wouldn't be around working on AI policy or regulation (or prediction markets, or anything involving the words like "decision making", "research", "influence", or "governance". If it involved AI, it would be something like using AI internally to replace other government employees.