Carlos E. Perez, creator of @IntuitMachine, argues model distillation narrows the open-source performance gap, undermining frontier model export controls

GLM-5.2 and Zhipu getting some CNBC coverage:

'But the real story underneath the price tag is the open source AI moment.

GLM 5.2 is free to download, fine-tune, and run on an enterprise’s own servers, putting pricing pressure on frontier labs at the same time that access looks shaky. Anthropic had to pull its Fable Mythos-class model after an order by the Trump administration, and OpenAI announced Friday that it is limiting its GPT 5.6 models because of a government request.

The federal oversight has made a model that no one can revoke increasingly look like the safer bet.'

@altcap Don’t blame the jujitsu, blame @AnthropicAI and it’s Elitist band of “researchers”

@altcap The Chinese have nothing to do with this.

The Epstein Class is seizing control of AI in America. The same Epstein Class that dragged us into one disastrous war after another.

@altcap This is a fake useless and made up Chinese UBS “study” stop spreading it!

It’s not real, there’s no data and it’s chasing panic over nothing.

Pure FUD

@altcap Brad, you shoulda taken Sam up on his offer to buy your shares

@altcap Very likely Chinese models get banned. A lot of murkiness in these waters, the “cheaper alternatives” wouldn’t even be a thing but for Anthropic innovating hard on agentic coding.

They’re stealing our tech thru distillation. We can and should end that.

@altcap 100% self-inflicted friction

@altcap From one Wabash guy to another:

If distillation is the primary method that Chinese labs are using to catch up, doesn’t a phased roll out increase the spread between US frontier and Chinese open source?

FWIW, I don’t like the regulation play

@altcap Can you really call it jiu jitsu if you shoot yourself in the foot and then someone comes and puts you in a hold?

@altcap @altcap I encourage you and Bill to step in if you can. I have little faith that the government understands these issues. Not only are they clueless about this tech, but they also can't comprehend second- and third-order effects like those being discussed here. We need your help!

China doesn't have to do anything. Tech leaders embellished and misrepresented their products while having the deepest pockets.

AI could have been sold as a bulldozer and crane, leading into a technology revolution. Instead, tech leaders jumped straight to digital nuclear weapons.

The tech industry did this to itself.

Contextual innovation is the inevitable future: https://medium.com/intuitionmachine/building-our-agi-future-a-world-of-affordable-contextual-innovation-3e36293ec750

@altcap Let's have that not be happening.

@altcap And you see the numbers from the big 2. So clearly you are validating the slowdown in revs @bgurley

@altcap Mate, China plays the long game while DC plays bureaucracy. Open source eats closed models anyway.

@altcap Can a closed model win long-term against free, open-source alternatives at scale?

@altcap @NickNemo17 Sorry you get what you paid for you fucking idiot congrats on electing a leader who is more communist than the actual Chinese at this point. You fucking dweeb

@IntuitMachine It's a real thing, it's also a non-issue. Distilled models remain behind frontier development always, provide funding tot he frontier company, and create access outside the frontier ecosystem allowing the frontier providers to more efficiently manage their compute dynamics.

@altcap My internal plan is as follows

👇Details are as follows 👇 👇

@altcap Anthropic would love the Admin limits usage of Chinese models by US entities for national security reasons, no?