AI thoughts index

# AI thoughts index - 18 Jun 2023 [[A lot of AI concern is driven by availability bias]] - 11 Jun 2023 — suggested that LLMs are good at navigating labyrinthine government websites - 10 Jun 2023 — use to come up with analogies and examples for - 20 May 2023 — Try “Help me build my intuition about …” may be a good prompt - 14 Apr 2023 — most good writing is at least in part a way of thinking, LLM-powered output is just going to create more not-thinking - 9 Apr 2023 — [[The digital world will become more and more like the fey realm]] - 8 Apr 2023 — feels like Everybody focusing on ai has given up on the future - 8 Apr 2023: People who can’t imagine how AI doesn’t just kill everybody have bad imaginations - 8 Apr 2023 — The abstraction and reasoning corpus seems like an actually good test - let luke know when you see someone at the frontier actually using gpt for not software - [GitHub - fchollet/ARC: The Abstraction and Reasoning Corpus](https://github.com/fchollet/ARC) as opposed to things that people just throw out - [[26 Mar 2023]]: GPT able to do some openscad modeling https://twitter.com/RealSexyCyborg/status/1639579952453656576?s=20. More broadly if you can represent a real world process in code, I’d believe it’s LLM-able. - [[25 Mar 2023]] — I think I agree with Andy that it’s a weak general intelligence but that doesn’t actually *mean* anything  [Web URL for this note](http://notes.benjaminreinhardt.com/AI+thoughts+index) [Comment on this note](http://via.hypothes.is/http://notes.benjaminreinhardt.com/AI+thoughts+index)