LESSWRONG
LW

Bohemian
12130
Message
Dialogue
Subscribe

Posts

Sorted by New

Wikitag Contributions

Comments

Sorted by
Newest
1Bohemian's Shortform
25d
3
Bohemian's Shortform
Bohemian24d21

Interesting! AI 2027 was likely one of the seeds of the idea (aside from actual world happenings and the discourse here).

I haven't looked into the methodology of AI 2027 in detail before, it makes sense it is a sort of wargame -- thanks for the link!

Reply
Bohemian's Shortform
Bohemian25d51

Rough concept idea for a board game: each player represents an AI lab and attempts to develop ASI. The actions involve raising money / hiring talent / building data centers, as well as interacting with other players in competitive manner. The idea lends itself well to the typical "snowball" style of board games (starting small, growing, later stages of the game actually concerning world events).

Game mechanics twist is that the game can only be won if the developed ASI is aligned, i.e. developing a misaligned ASI would lead to all players losing. Ideas how to achieve alignment would probably involve the need for cooperation among players (multi-lateral actions) as well as spending own resources.

If designed well, the game could be both fun and "educational".

Reply
On "ChatGPT Psychosis" and LLM Sycophancy
Bohemian1mo80

On the topic of sycophancy: lately I have been basically exclusively using the "thinking" ChatGPT models, o3 and o4-mini-high, and, although the resulting outputs are of higher quality, I noticed that sometimes I feel a sting of missing the "Great idea!" / "What a great question" / ... responses of 4o. The "coldness" of the response is especially visible when reading the CoT summaries, including sentences like "the user is aiming to... " / "it sounds like the user wants...".

Reply
No wikitag contributions to display.
1Bohemian's Shortform
25d
3