… yesterday turned strange quickly.

5 hours into the test to confirm whether Anthropic’s new Fable 5 model could play Montezuma’s Revenge (and therefore qualify as weakly Artificial General Intelligence), this happened:

Of all the bugs that could have unplugged the experiment, sudden U.S. Government enforcement of export controls over the specific frontier frontier AI model being used, was not on my bingo card!

I won’t comment on that further, except to say that it created a cliffhanger ending. Before becoming an official peer of high-powered lasers, milling machines for submarine propellers, and Uranium-enrichment centrifuges, Fable 5 was able to propel Panama Joe into 17 of the 24 rooms, very much on pace to defeat the game and officially become AGI. So I’ll call this a partial validation.

Speaking of, in 2017 Mark Cuban made a prediction that: “…the world’s first trillionaires are going to come from somebody who masters AI and all its derivatives and applies it in ways we never thought of…”. Yesterday, SpaceX (which had ingested X.AI which had ingested X/Twitter) successfully IPO’d to a market cap of $2T, propelling Elon to 1e12-ionaire status, positively resolving both Mark C’s and the market’s prior predictions— while Mark A’s remains in suspense.

Let me sum up. No, there is too much. Let me conclude:
P(doom) = no idea ; P(weird) = 1

References:

https://www.metaculus.com/questions/3479/date-weakly-general-ai-is-publicly-known/

https://www.anthropic.com/news/fable-mythos-access

https://www.cnbc.com/2017/03/13/mark-cuban-the-worlds-first-trillionaire-will-be-an-ai-entrepreneur.html

https://kalshi.com/markets/kxtrillionaire/first-trillionaire/kxtrillionaire-30?op_market_ticker=KXTRILLIONAIRE-30-EM