Jyunmi

It wasn't a lucky guess.

The model conjectured a formula and spent 12 hours verifying it against the Behrens GL recursion.

which is apparently some methodology for proofing something.

And so this is just further indication of the advanced intelligence that's being created in artificial intelligence.

And even though we haven't gotten to general intelligence, in certain silos like advanced mathematics, like Axiom AI that we talk about, which has developed a platform for doing, you know,

proofs that humans have not been able to achieve and there are some really incredible savant minds out there that have been applied to these problems but ai is now synthesizing what knowledge we have along with reasoning capabilities and pushing it beyond what humans have been able to do that's just an important point to you know putting your pipe in smoke

And OK, that's that one.

I want to just mention another one.

Google DeepMind, as you know, is really a center of excellence of advancing AI.

And they have just shown the ability, using a general purpose bioacoustic model, to best any of the prior models in classifying whale sounds.

And the training data that was used to build this bioacoustic model was on bird sounds.

Now, there's not a genetic relationship directly between birds until you go way back to this level, I guess, between birds and whales.

But there are some commonalities to the way those vocalizations occur.

And training on bird sounds made it possible for the model that they built to understand the whale songs better.

And that's pretty interesting.

And here's just a little bit of additional information about that.

Visual contact underwater with animals like dolphins and seals and marine mammals, orcas, for example, is not generally possible.

So they can't really correlate a lot of data around that.

So they just have to listen to the sounds that are being generated.

And