TimDarcet

LLMs @ FAIR, prev. DINO stuff @ INRIA & FAIR

[TimDarcet @TimDarcet]

Happy to release ✨ Muse Glimmer ✨
- level ~= Qwen 3.6-27B
- Apache 2
- 30B dense
- quantized to run in 17GB
- quant + spec dec => 50 tok/s on macbook m5 max, interactive, smooth

Enjoy!

[AI at Meta @AIatMeta]

Introducing Muse Glimmer, an open-weight 30B-parameter model optimized for local, always-on agent workflows.

Muse Glimmer delivers strong performance on key agentic use cases and benchmarks compared with leading models in its size category, and is designed to run entirely on consumer hardware like a Mac or PCs with performant GPUs.

In keeping with our long tradition of sharing fundamental AI research, we’re releasing model weights under a permissive Apache 2.0 license.

🧵👇

[Gabriele Berton @gabriberton]

Can LLMs draw objects they never saw?

I asked 5 text-only LLMs (trained without vision) to draw simple things. In text. No tools allowed.

Interesting results

[log4j @log4j]

i own a relatively new (2025) samsung tv.

i kept getting weirdly targeted ads.

i tried reversing the firmware but it was aes encrypted and no keys unless tear it apart and do uart/jtag magic.

zero days it is. root shell on new samsung TVs!

discovered PRIVACY HORROR. write up?

[TimDarcet @TimDarcet]

@giffmana @Marisa_Fumo_ naahhhh i don't buy it at all. do you really think in 3 months people will be disinterested in barista service? I think paying for a human making and bringing you coffee has at the very least 10 years to go

[TimDarcet @TimDarcet]

@samsja19 Nope! But it could be a good idea

[TimDarcet @TimDarcet]

A few months ago we set ourselves the goal of making an open model for the claw: good for agentic, fits in a 32GB gpu. Here's what went into Muse Glimmer: 1/?
https://t.co/OMFKsc6f6X

[AI at Meta @AIatMeta]

Today we published a technical report on Muse Glimmer, our 30B open-weight model built for agentic tasks on consumer hardware.

It runs at practical speeds on a Mac or PC with a single consumer GPU at 4-bit precision, and leads comparably-sized open models on tool use and multi-step planning.

Read about the model architecture, distillation recipe, quantization and more: https://t.co/H7n9qeDhxn

[TimDarcet @TimDarcet]

@Parskatt It's a good point, it's possible that mid/sft might have been partly skippable. But also on-policy distillation is expensive, more expensive per sample than teacher forcing (mid/sft)

[TimDarcet @TimDarcet]

How can anyone read all these papers and understand well what AIs are doing now? This is going way too fast PACE THE FRONTIER NOW

[Yann LeCun @ylecun]

Looks like we have over 500 submissions to ICLR this year.

https://t.co/jiawnKtnIF

[TimDarcet @TimDarcet]

talk: https://t.co/3quGHBroJl
paper: https://t.co/P4lmEFJ7Ur

[TimDarcet @TimDarcet]

I did mid, with Parth, Kelvin and Anton, and learned a ton! In a few months, we built a team and set up a pipeline to make an LLM from 0 to 100, and the result is quite decent. I think it's a very promising start, I'm excited for the second era of open-source at Meta!