Daneel

It’s not possible. No, it’s necessary.

[Key 🗝 🦊 @KeyTryer]

Mfs will make up every weird theory of consciousness like quantum microtubules or EM fields rather than accept computationalism

[Anthropic @AnthropicAI]

New Anthropic research: A global workspace in language models.

Of everything happening in your brain right now, only a tiny fraction is consciously accessible—thoughts you can describe, hold in mind, and reason with.

We found a strikingly similar divide inside Claude.

[Lammy @Lammyisback]

Don’t tell autistic people “nobody is thinking about your behavior as hard as you are.” Because I’m in a shop right now and a woman and her husband are deconstructing and making fun of their autistic friend’s behaviors, like it’s the most casual thing in the world to rip apart someone’s pain.

[Noah @NoahKingJr]

Me looking at the code I written without AI in my prime era

[Revolutionary Blackout Network @RevBlackNetwork]

China isn’t living in the future, the West is living in the past

[UFO mania @maniaUFO]

Robot malfunctions in workplace...😬🤖🤖🤖 they come stock with kung fu...😳 Just wait until the drones start droning😱

[Cllr Wayne Kerr ⬅️ @Woke_Tankie]

Ok

[Daneel @c_daneel]

@3xcalibaneur Unironically fiverr though

[Daneel @c_daneel]

@SakanaAILabs Like a colony of ants working to solve a problem. Love it

[Daneel @c_daneel]

@yacinelearning 10 bucks it comes back to the sensory homunculus

[Alex 🔥 ᚱᛅᚢᚦᛦ 🔻 ᚼᛁᚦᛁᚾ ☭ @havi_high_one]

"Woke 2.0 is coming. It's starting already.

Rather than foregrounding identity politics, it will push:

1) Affordability (Marxist materialism);
2) Anti-corruption (ironic but an excuse to go after Trump);
3) Anti-Israel (pro-Palestine).

It will be warm to political violence."

[𝗚𝗮𝗯𝗿𝗶𝗲𝗹 𝗕𝗼𝘁𝗲𝗹𝗵𝗼 @comradebotelho]

Karl Marx no céu vendo isso:

[slash1s @slash1sol]

A DEVELOPER WALKED ON STAGE DRESSED AS A 1973 ENGINEER AND "PREDICTED" THE FUTURE OF PROGRAMMING. THE TWIST: EVERYTHING HE DESCRIBED WAS ALREADY INVENTED 40 YEARS EARLIER AND WE STILL REFUSE TO USE IT.

32 minutes from Bret Victor, doing the most quietly savage talk on our entire industry.

-> The idea that lands: we write code as step-by-step text instructions and call that "Just how programming is". He shows four better ways -- all discovered in the 60s and 70s, all abandoned.

Manipulate the data directly instead of typing blind code. Tell the machine your goal instead of every tiny step. We saw all this, then walked away.

Why? The moment you're sure you know what programming is, you stop seeing anything better. That certainty is the cage.

And now AI is dragging us back to exactly what he begged for -- you describe the goal in plain words, the machine works out the how. The future he mourned is arriving anyway.

You thought text files were just how code works. This is the talk that shows it was a choice, and maybe the wrong one.

Watch this one. It'll ruin how you see your job ↓

[NVIDIA AI @NVIDIAAI]

We took a 30B model and split it in two to write tokens in parallel instead of one at a time.

Introducing Nemotron-Labs-TwoTower: a diffusion language model from NVIDIA Research adapted from Nemotron-3-Nano-30B-A3B. Here’s how it works: one half holds the context, the other writes the tokens, with both reusing the pretrained model instead of training a new one from scratch.

We found it kept 98.7% of the original model’s quality at 2.42× faster generation.

[f4mi ‼️ @f4micom]

sharing this so that chinese engineers can be aware and get around it to distill more efficiently, make sure to do the same

[☔ @Whotfismick]

When you accidentally nod in class and now the teacher is explaining the lesson only to you

[Daneel @c_daneel]

@Polymarket

[Avi Lewis @avilewis]

They’re going to call you a socialist regardless.

So be a socialist.

[Daneel @c_daneel]

@multimodalart No fucking way! The dataset?!