Rafael Pinto

Michael Levin would like to have a talk.

Oh, thank God!

After using Gemini 2.5 Pro, Claude Sonnet 4.5 and GPT-5 for highly novel theoretical neural networks research, I must say that, at least for my case, GPT-5 is the clear winner. The "Ph.D. level" claims are correct (and of course, like any human Ph.D., it also makes stupid mistakes now and then).

Pensei que fosse o Toninho do Diabo.

Define "normal".

I wasn't familiar with that concept at all, but it seems very relevant to the subject indeed, thanks!

bsky.app/profile/rcpi...

Do you realize the abysmal difference between being a user (and "looking into it") and doing decades of research and, in fact, creating the very foundation for those things?

The hatred coming from the butthurt anti-AI mob is now comparable to that of far-right scumbags.

It's one thing to dislike the technology, but the way they're responding to some people here is sick to say the least.

Wow, didn't know such a document existed.
docs.google.com/document/d/1...

If you confidently assert that LLMs (or anything else besides yourself, for that matter) are or aren't conscious, you're wrong.

I dream of a social network that's like a generalization of that. For any topic, you have to prove knowledge before posting.

bsky.app/profile/rcpi...

YouTube video by Chris Spencer Steven Spielberg's A.I.-ARTIFICIAL INTELLIGENCE (2001) - The Flesh Fair

This is gonna happen very soon. The signs are crystal clear.
www.youtube.com/watch?v=ZMbA...

The results show exactly the opposite.

Here we go, another thread full of geniuses and experts that know everything about how AI is just autocomplete. I'm glad to grow my block list, tho.

Dumb take.

That made me chuckle @tedunderwood.me

From the accelerate community on Reddit: Anthropic: "Sonnet 4.5 recognized many of our alignment evaluations as being tests, and would generally behave unusually well after." . Image: Sonnet 4.5 score... Explore this post and more from the accelerate community

www.reddit.com/r/accelerate...

What's the GM Benchmark??

Could've been zolpidem.

Gado e crente no mesmo dia não dá, Fantástico.

Galera do BlueSky dá chilique cada vez que alguém fala em IA, é hilário. #Fantástico

E o Fantástico, com uns 2 ou 3 meses de atraso, ressuscitando aquela "pesquisa chocante" que verificou que lembramos melhor de algo que nós escrevemos ao invés do que não escrevemos (ChatGPT).

Says the man who sells non-humanoid robots.