We work at the limit of our tools. When new tools come along, new things are possible. So maybe we haven't "solved physics" because we aren't smart enough to do it with current tools, but new AI tools will help us solve it quickly.https://twitter.com/sama/status/1274762559694897152 …
-
-
Replying to @sama
I've sometimes wondered if a quantum AI might be needed to solve the problem of quantum gravity. Not a terribly serious though, but an amusing possibility that I can't rule out, and can even make a half-serious case for.
5 replies 2 retweets 29 likes -
Replying to @michael_nielsen @sama
Related piece of wild speculation: the notion that a quantum driven singularity might be qualitatively different to a classical one. So, you'd see two phase shifts in civilization: a first singularity, then a gap to quantum AI, and then a second.
2 replies 1 retweet 11 likes -
Replying to @michael_nielsen @sama
Ofc, it's possible recursive self-improving quantum AI might occur during the first spike, in which case no double singularity. Depends on how tough a problem RSIQ AI is.
1 reply 1 retweet 7 likes -
Replying to @michael_nielsen @sama
Is there a handwavy argument that RSIQ AI could be theoretically easier than RSIC AI, if improvement includes some kind of Grover-like search step? (Assuming no hardware limitations.)
1 reply 1 retweet 2 likes -
An amusing possibility. To pile wild speculation on top of speculation, if BQP contains NP (& P != NP) then you could imagine RSIQ might well happen first (informally: QCs could find valid scientific arguments roughly as fast as they could check them, which presumably helps...)
1 reply 1 retweet 3 likes
Michael, will you indulge me? can you write the best description of a first principles cut at *what* the problem is of quantum gravity --- in 280 characters?


Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.
