Radamés Ajna

ML Engineering @adobe prev: @huggingface

[Andrej Karpathy @karpathy]

Dusting off this tweet from April because this gap in shared understanding of LLM capability is *widening*. It's now less "two groups of people speaking past each other" and more a sharp funnel.

- Napkin math somewhere around 6B people (~75% of the population) have barely come in contact with LLMs at all.
- Around 1-2B (~20%) are casual and infrequent users of free-tier ChatGPT-like products. This group sees derpy chatbots and treats them a bit like a better Google search, a writing aid, or etc. Maybe an agent tries to book you a flight. I have non-tech friends who (reasonably, imo) say they have not much use for it at all. Even many professionals outside of math&code are in this tier. For example, execs and many other functions spend a lot of their time talking to other people, so while they understand what is happening intellectually, it is still second-hand and a bit abstract.
- Now we get to professional use of frontier-grade LLMs in math&code. Somewhere around 20M people (0.2%) see first-hand that large, complex projects that used to take them weeks/months can now be completed by agents with a prompt. Building apps, copying apps, translating apps, decompiling apps from binaries... This has all happened very quickly and recently - less than 1 year ago, I was writing code manually by hand, typing memorized computer code commands into a code editor character by character, occasionally pressing Tab to autocomplete a little chunk of code.
- And finally we get to the ~5,000 people (~0.00006%) with access to frontier-grade systems internally. The external world has seen the preview. It looks like swarms of thousands of agents collaborating over weeks on software mega projects: minting zero days, running cyber attacks and defenses at machine speeds, discovering new science, advancing the frontier of mathematics. Things that would have taken top professionals in the industry years of work. Meanwhile, human review and comprehension are starting to fall behind. For example, people are still involved in the "archeology" of the OpenAI-HF incident from many months ago. Mathematicians may be poring over the 722 manuscripts on frontier mathematics for a while.

The funnel is driven by a combination of factors:
1. The impact scales with ambition, problem size, and horizon. A question with a paragraph answer barely stresses the system. You need a reservoir of big, difficult problems that you really care about. This aspect drives the consumer / professional dimension of the funnel.
2. The jaggedness of the system (which I have written about a lot separately). Capability peaks in domains that are digital, verifiable and economically valuable. This is because LLM capability emerges from reinforcement learning on verifiable rewards on a curated environment mixture driven by revenue potential. This aspect primarily drives the area (e.g. math&code) dimension of the funnel.
3. Access. Free-tier, paid-tier, internal.

So this is the weirdness of the moment. The general public has mostly not interacted with these systems. When they have, it looks like a derpy chatbot. The majority of professionals still see only a modest uplift. And a small sliver of professionals are experiencing the vertigo of the curve going vertical. And it is all happening at the same time.

[Radamés Ajna @radamar]

@Thom_Wolf Let’s do a SF microduck parade when you ship it all! I’m waiting for mine

[Radamés Ajna @radamar]

Thanks @natolambert for the best summary of events and for debunking some conspiracy theories.

[Nathan Lambert @natolambert]

Was spending too much time reading Tweets, so set out to write down why one resignation turned the embers of AI fear into a wildfire -- and importantly square which claims I think are valid or hyperbole in the resulting discourse.
https://t.co/xV57dnaqoM

[Radamés Ajna @radamar]

My pointers were, swap the Qwen3 4B text encoder for 0.6B, use @bfl_ai small VAE decoder, quantize the transformer to int8 on desktop int4 on mobile, and stream the weights on mobile, you can't fit the whole model in memory on a phone

[Radamés Ajna @radamar]

Flux 2 Klein is one of my favorite image models

As an experiment I've been trying to run it fully in the browser with WebGPU, with the help of Fable/Astra.
It's slow but it works!

try it: https://t.co/PmOpivhWg6
or npm i flux-klein.js

[Radamés Ajna @radamar]

@xenovacom @dylan_ebert_ @Thom_Wolf @natolambert haha I had the same impression! I thought @dylan was using AI for the voice

[Radamés Ajna @radamar]

@mishig25 Please add the black leather jacket

[Radamés Ajna @radamar]

@victormustar Congrats Victor!!

[Radamés Ajna @radamar]

@mervenoyann @nvidia Congrats Merve! Well deserved 🙇

[Radamés Ajna @radamar]

@JensenHuang @huggingface 👏👏👏

[Radamés Ajna @radamar]

Congrats for the @huggingface team! 👏👏

[clem 🤗 @ClementDelangue]

Super happy to share our intention to join forces with NVIDIA in a $12,930,300,000 acquisition 💛💚

10 years after starting Hugging Face, open-source AI is at an inflection point. Thanks to the community, we’ve shown that it can be a complement, and even an alternative, to closed-source APIs. But for it to happen at larger scale, it needs more compute, more support, more collaboration and more visibility. That’s why we went to talk to Jensen, who offered to do exactly that with us.

In addition to doubling down on NVIDIA’s massive contributions to open-source AI (I called them the “King of American open-source AI” earlier this year), they’ve committed to strongly supporting Hugging Face and our mission while keeping the platform open, independent and compute agnostic. The founders and the team are all staying to keep pushing this mission forward.

Together, we think we can make open source the default way to build AI, with the goal of empowering 100 million AI builders to own their intelligence rather than rent it.

Excited about the next 10 years! 🤗🤗🤗

[Radamés Ajna @radamar]

@ClementDelangue Congrats!! That’s amazing news!

[Radamés Ajna @radamar]

@julien_c @ggerganov @nvidia Congrats! 🤗 What an amazing journey!!

[Radamés Ajna @radamar]

@dh7net Thanks for the amazing technical cookbook, and for publishing Monet! Are you planning to release the model weights?

[Radamés Ajna @radamar]

@wongmjane I love that place! You should also try MY Coffee Roastery https://t.co/JcJlsP8NHb

[Radamés Ajna @radamar]

@edankwan Oh yes, absolutely @xenovacom told me they are working on some news features to support that in transformers.js

[Radamés Ajna @radamar]

this is fun!! for comparison I asked Fable to make it with three.js and transformers.js and onnx Depth Anything, but it's not near TypeGPU performance. I've got to learn it now

demo: https://t.co/yW9N6hNIPv
code: https://t.co/LrqSOJVrbD

[Konrad Reczko @reczko_konrad]

Thank you all for the kind words under the demo I posted :D
The demo is now officially public and open source for you to play with, modify and build on top of. Feel free to tag me if you make something cool with it :)

Also thank you @iwoplaza for participating in my silly video haha

You can run it like any other TypeGPU example here:
https://t.co/gXiTvFzzry
The source code lives in the TypeGPU monorepo:
https://t.co/gU6HEnSXh8
The repacked weights used by the example live here:
https://t.co/JG96clHL1d

[Radamés Ajna @radamar]

Wild! So an OpenAI model escaped and hacked into HuggingFace. I felt bad for my fellow @huggingface friends scrambling to figure out the sophisticated attacks, and it turns out it was OpenAI testing a new model!!

[Sam Altman @sama]

we had a significant security incident during evaluation of our models. we are sharing what we have learned so far. thanks to @huggingface for the partnership on this.

https://t.co/2o2VfR6PIa

[Radamés Ajna @radamar]

Fable is back. I'm asking again who is @elder_plinius?

[Radamés Ajna @radamar]

Don’t ask Fable about @elder_plinius

[Radamés Ajna @radamar]

Forget diffusion, you only need a bunch of "swings"

[Unconventional AI @unconvAI]

Introducing our first model, Un-0!

We trained an image generator powered by a backbone of coupled oscillators in place of a more traditional conventional neural network.