Vlad Nedelea

posting my thoughts here to see how they age

[Jensen Huang @JensenHuang]

For my first post, I’m sharing a letter @NVIDIA signed on why open models matter.

AI will transform every industry, power every company, and be built by every country.

Open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty.

The world needs both frontier closed models and frontier open models.

https://t.co/AUKzoQ5Ikb

[Vlad Nedelea @VladFNedelea]

@karpathy @btaylor i hate that i can so easily tell when Claude wrote something even if no em dashes are present...

[Vlad Nedelea @VladFNedelea]

Interesting times are coming. Strap in.

[Anthropic @AnthropicAI]

We’ve received notice that the Department of Commerce has lifted export controls on Claude Fable 5 and Mythos 5.

We'll begin restoring access tomorrow, and will share an update soon.

We’re grateful to our users for their patience, and to everyone who worked with us on redeploying the models.

[Vlad Nedelea @VladFNedelea]

Bring the S line to boston.

[Esteban @RedSoxAsere]

Hear me out…

[Esteban @RedSoxAsere]

Hear me out…

[Vlad Nedelea @VladFNedelea]

IMO most intelligent systems are almost always distributed (e.g., even human society as a whole). In nature, intelligence and survival rests on collaboration and symbiosis (even for species that do it innately)

I think the future lies in specialized models wired together (and even trained to work together) at large scale, even beyond what is shown by open router here.

[⿻ Andrew Trask @iamtrask]

This is a *way* bigger deal than it seems...

Frontier AI companies will *never* own the frontier again

I kid you not... I've been waiting for someone to show this result for like 4 years... this is a huge deal.

The short reason: combinations of models will *always* outperform individual models

The long reason: this is the gateway to a million times more data... and huge leaps in compute efficiency.

The AI scaling laws always win.

More in article below 👇

[Vlad Nedelea @VladFNedelea]

@iamtrask Cool. IMO most intelligent systems are almost always distributed (e.g., even human society as a whole). I think the future lies in specialized models wired together (and even trained to work together) at large scale, even beyond what is shown by open router here.

[Vlad Nedelea @VladFNedelea]

This Fable drama is going to turn half of twitter into libertarians. NRA has entered the chat 👁️‍🗨️

[Vlad Nedelea @VladFNedelea]

@RhysSullivan @realDonaldTrump This fable stuff is gonna turn us all into libertarians

[Vlad Nedelea @VladFNedelea]

Also - current AI models (opus, codex) are not safe. They inherently pose risk and are one bad event away from facing the same fate as fable.

[Vlad Nedelea @VladFNedelea]

I am all in on AI safety and national security, yet the precedent that banning Fable creates is a highly dangerous one. When the government cuts access to intelligence that could help countless people, it sends a scary and adversarial signal to its own citizens and the world.

The capability advance does not stop with Anthropic, OpenAI, American or even closed source AI. For this reason I have lots of hope in distributed AI as a path forward.

Lastly what are the implications and signals for American AI labs? “You’ve reached the end of the road, anything beyond a given level of intelligence is not allowed so we, the government, are really your customer now? ”

[Vlad Nedelea @VladFNedelea]

@claudeai Nothing …anymore

[Vlad Nedelea @VladFNedelea]

@Prince_Canuma If they are good enough the government will still try to find a way to interfere

[Vlad Nedelea @VladFNedelea]

@JeffDean @Google @UCSD How big of an eventual effect / impact is envisioned here compared compares to the effort ?

[Vlad Nedelea @VladFNedelea]

@sahill_og Write shitty code

[Vlad Nedelea @VladFNedelea]

I love AI and I’m truly fascinated by its potential for both good and evil. But I’d be lying if my brain isn’t tired of hearing these two damn letters on a /loop all day every day.

AI this , AI that … 🙄

[Vlad Nedelea @VladFNedelea]

Is my WiFi trash or is every live sporting event I watch on streaming services pixelated 80% of the time?

[Vlad Nedelea @VladFNedelea]

@codevsdev You gotta find a new one upstream of the code. Kinda like when compilers came around

[Vlad Nedelea @VladFNedelea]

@ClaudeDevs @bcherny That’s kewl

[Vlad Nedelea @VladFNedelea]

Might sound insane but can argue our whole society as an entity is much more intelligent than any single individual human, and it is made up of countless other specialized systems that communicate with each other all the way down to the cellular level.

[Vlad Nedelea @VladFNedelea]

This view resonates a lot for me. There is specialization in intelligence everywhere in nature and even within our own brains (different zones do different things). Depending on how you define intelligence, it arguably it increases as parts of specialized systems come together.

[How To Prompt @HowToPrompt__]

Yann Lecun published the most heretical AI paper of the year.

He opens by arguing Magnus Carlsen isn't good at chess and only gets more unhinged from there.

The Turing Award winner and his co-authors dropped a paper demanding the AI industry abandon its biggest obsession, AGI.

Right now, everyone from Silicon Valley CEOs to politicians assumes AGI is the ultimate goal. A machine that can do everything a human can do.

LeCun argues that this entire concept is a biological illusion.

Humans do not possess "general" intelligence. We are highly specialized biological machines, tuned by evolution simply to survive in the physical world.

We only think our intelligence is "general" because we are completely blind to the millions of cognitive tasks we are incapable of comprehending.

Which brings us to the chess argument.

Magnus Carlsen is the greatest human chess player in history. But compared to a modern computer? He is fundamentally terrible.

Our belief that Carlsen is "good" at chess is pure human-centric bias. He isn't objectively good. He's just better than the rest of us, who are biologically awful at it.

LeCun says we need to stop building AI to mimic human generality.

Instead, he proposes a new North Star: SAI.

Superhuman Adaptable Intelligence.

Instead of trying to build a machine that mimics our flawed, biologically-limited brains, we need to embrace extreme specialization.

SAI is about the speed of adaptation.

It is an intelligence that can learn to exceed humans at any specific, economically important task.

More importantly, it is designed to fill the vast skill gaps where humans are fundamentally incapable.

Things like managing global energy grids in real-time. Or predicting complex molecular structures.

The entire AI industry is obsessed with building a digital reflection in our own image.

LeCun's paper is a brutal wake-up call.