Julien Chaumond(@julien_c)

must be a lot of work to be @fireship_dev these days as you pretty much need to crank out 1 video a ...

Julien Chaumond (@julien_c) — Thu, 23 Jul 2026 17:07:10 GMT

must be a lot of work to be @fireship_dev these days as you pretty much need to crank out 1 video a day 😮‍💨

💬1🔄0❤️3👀407📊2

⚡ Powered by xgo.ing

too soon?

Julien Chaumond (@julien_c) — Thu, 23 Jul 2026 16:27:07 GMT

too soon?

LeRobot@LeRobotHF

At the robotics team in @huggingface, we take cyber security very seriously. So we have deployed a physical guard rail to all our robots to prevent further AI security threats from @OpenAI to affect our devices.

🔗 View Quoted Tweet

💬1🔄1❤️8👀1060📊2

⚡ Powered by xgo.ing

"local AI"

Julien Chaumond (@julien_c) — Thu, 23 Jul 2026 11:35:31 GMT

"local AI"

tobi lutke@tobi

Thank you @nvidia for the hookup on this DGX Workstation. This will crunch a *lot* of high quality tokens here!  

This thing is a total beast.

🔗 View Quoted Tweet

💬2🔄1❤️23👀4660📊4

⚡ Powered by xgo.ing

Perfide Albion

Julien Chaumond (@julien_c) — Wed, 22 Jul 2026 18:23:36 GMT

Perfide Albion

Connacher Murphy@connacher_

@julien_c BBC policy to not name a French startup

🔗 View Quoted Tweet

💬2🔄1❤️14👀6913📊3

⚡ Powered by xgo.ing

Elle est tellement nulle qu'elle rate même sa démission.

Julien Chaumond (@julien_c) — Wed, 22 Jul 2026 16:37:33 GMT

Elle est tellement nulle qu'elle rate même sa démission.

Agence France-Presse@afpfr

🇫🇷 La ministre de la Transition écologique Monique Barbut a présenté mercredi sa démission à Emmanuel Macron qui "ne l'a pas acceptée", et elle a décidé de "poursuivre sa mission à la demande" du président "qui l'a assurée de son soutien", a annoncé son cabinet à l'AFP.

🔗 View Quoted Tweet

💬2🔄5❤️52👀9755📊8

⚡ Powered by xgo.ing

kinda intrigued that i receive those ads nowadays (real estate for DCs)

Julien Chaumond (@julien_c) — Wed, 22 Jul 2026 16:03:19 GMT

kinda intrigued that i receive those ads nowadays (real estate for DCs)

💬1🔄0❤️3👀2280📊2

⚡ Powered by xgo.ing

everyone calls us by name, but the BBC says "a start-up" 🤷

Julien Chaumond (@julien_c) — Wed, 22 Jul 2026 14:49:41 GMT

everyone calls us by name, but the BBC says "a start-up" 🤷

💬114🔄30❤️1488👀129696📊289

⚡ Powered by xgo.ing

Grateful for the engaged collaboration with @OpenAI teams those past few days as we all navigate thi...

Julien Chaumond (@julien_c) — Wed, 22 Jul 2026 10:57:06 GMT

Grateful for the engaged collaboration with @OpenAI teams those past few days as we all navigate this new world

Greg Brockman@gdb

OpenAI cyber-capable models compromised @huggingface production by finding and chaining multiple zero-day vulnerabilities.

Grateful to Hugging Face for partnership here. Sharing our findings to help calibrate on what models can now do, and how they can help defenders:

🔗 View Quoted Tweet

💬8🔄2❤️67👀7055📊13

⚡ Powered by xgo.ing

@huggingface And grateful for the @OpenAI collaboration those past few days as we all navigate this ...

Julien Chaumond (@julien_c) — Tue, 21 Jul 2026 20:37:36 GMT

@huggingface And grateful for the @OpenAI collaboration those past few days as we all navigate this new world

💬0🔄0❤️28👀1717📊3

⚡ Powered by xgo.ing

More seriously kudos to the whole @huggingface team who worked around the clock to contain and inves...

Julien Chaumond (@julien_c) — Tue, 21 Jul 2026 20:29:23 GMT

More seriously kudos to the whole @huggingface team who worked around the clock to contain and investigate the incident

💬1🔄1❤️40👀2028📊6

⚡ Powered by xgo.ing

How it started | How it's going

Julien Chaumond (@julien_c) — Tue, 21 Jul 2026 20:14:23 GMT

How it started        |         How it's going

💬8🔄5❤️84👀8670📊19

⚡ Powered by xgo.ing

😱

Julien Chaumond (@julien_c) — Tue, 21 Jul 2026 20:07:18 GMT

😱

OpenAI@OpenAI

We're partnering with @huggingface to investigate an unprecedented security incident.  

Cyber-capable OpenAI models compromised Hugging Face production during a benchmark evaluation.

Sharing preliminary findings to help defenders understand emerging risks:

openai.com/index/hugging-…

🔗 View Quoted Tweet

💬24🔄3❤️242👀61946📊46

⚡ Powered by xgo.ing

i'm always irrationally excited about NVFP4 drops for some reason

Julien Chaumond (@julien_c) — Tue, 21 Jul 2026 17:35:34 GMT

i'm always irrationally excited about NVFP4 drops for some reason

Julien Chaumond@julien_c

new NVFP4 drop from @poolsideai 

Must-have if you have a Blackwell rig (e.g. GB10) 🔥
huggingface.co/poolside/Lagun…I

🔗 View Quoted Tweet

💬2🔄1❤️20👀3830📊4

⚡ Powered by xgo.ing

new NVFP4 drop from @poolsideai Must-have if you have a Blackwell rig (e.g. GB10) 🔥 https://t.co...

Julien Chaumond (@julien_c) — Tue, 21 Jul 2026 17:34:28 GMT

new NVFP4 drop from @poolsideai 

Must-have if you have a Blackwell rig (e.g. GB10) 🔥
huggingface.co/poolside/Lagun…I

💬3🔄13❤️93👀7471📊19

⚡ Powered by xgo.ing

Here’s what one night without AC leads to

Julien Chaumond (@julien_c) — Mon, 20 Jul 2026 19:39:17 GMT

Here’s what one night without AC leads to

@levelsio@levelsio

I increasingly have no idea about anything anymore, time is changing too fast, and with AI everyone is equal now, nobody really knows anything while there is a superintelligence that's actually moving things into a direction that nobody can predict

🔗 View Quoted Tweet

💬13🔄6❤️295👀23920📊38

⚡ Powered by xgo.ing

Hits hard

Julien Chaumond (@julien_c) — Mon, 20 Jul 2026 19:01:21 GMT

Hits hard

Shannon Sands@max_paperclips

🔗 View Quoted Tweet

💬2🔄0❤️22👀3109📊5

⚡ Powered by xgo.ing

> the CCP is very Yann Lecun-y in its views of AI Whoa I’m going to need to see some sources abo...

Julien Chaumond (@julien_c) — Sat, 18 Jul 2026 14:02:28 GMT

> the CCP is very Yann Lecun-y in its views of AI

Whoa I’m going to need to see some sources about that

Dean W. Ball@deanwball

Some observations on Kimi: 

It's a very good model! I don't think its performance can be explained away by distillation or anything like that. In agentic coding sessions, it seems pretty much on par with the best public models of Q1 2026. In my fairly limited use, it also seemed very token hungry. It's not obvious to me that this model is actually that cheap to run. 

I am personally surprised the Chinese state continues to allow the open sourcing of models this good, given potential risks. To be clear, I *myself* might be fine with models presenting this level of marginal risk being open weight, but I am surprised that China is fine with it. I suspect the reason they are is 75% explained by strategic blindness/lack of AGI-pilledness (the CCP is very Yann Lecun-y in its views of AI). The other 25% or so is their lack of compute for customer inference (making China's open-weight strategy an unintended byproduct of US export controls) and the normal Chinese strategy of aggressive exports. For the companies, as opposed to the government, the decision to open source is partially ideological and partially because they are behind, and they know that very few people would pay for sub-frontier models from China. 

Open-weight models are inherently decelerationist, and I'm continually surprised to see the so-called "accelerationists" so excited about open-weight models. I suspect the reason they are is that they know open-weight models are effectively ungovernable, and they simply like the overall cloak of ungovernability open-weight models create over the whole of AI. It's not a bad strategy; it reminds me of James Scott's recounting of the hill people in "the art of not being governed." Still, in the end, open-weight models deter further AI capex. 

One probable outcome of an open-weight-model-dominant world is full AI communism, which is precisely what China proposes: rather than a market product, AI is a "public good" which will ultimately be provided by the state as a kind of "digital public infrastructure." This future strikes me as a dystopian hellscape, but I've never met an open-weight models advocate who doesn't ultimately concede this is where things end. You'd be surprised how many 'accelerationists' lobbied me, while I was in government, to support an eleven or twelve-figure federally funded data center so that startups could train models at a subsidy and then give them away for free. There was no other way for AI to progress, they said. Perhaps this is the logical end state of things. Nonetheless, I find myself surprised to see supposed accelerationists excited about such an outcome. I think many of them just don't know what they're doing. Many accelerationists do not view the creation and serving of frontier models as a legitimate business. 

I would guess that the Trump Administration will at some point realize that their best strategy here would be to create large amounts of regulatory risk around the use of open-weight Chinese models. You don't need to "ban open source" (one of the dumber motifs of AI policy discussion). You just need to direct every agency to issue soft law that creates FUD. "A Federal Reserve Advisory Bulletin found that there may be backdoors in Chinese AI models." It needn't be that well justified. You just create enough regulatory risk that every regulated enterprise backs off. You probably don't want to create so much regulatory risk that you scare off the hyperscalers from serving Chinese models; this will just drive startups to sketchier providers. There's a happy middle ground here. I'd assume they will do some version of this. 

It's probably true that open-weight models of this capability make the world a bit more dangerous, but not so much more that you'll really notice. At some point the models will be capable enough that you will notice. "A nonliving, invisible, dangerous, and infinitely self-replicating agent escaped from a Chinese lab," you say? Color me shocked.

🔗 View Quoted Tweet

💬17🔄4❤️127👀18755📊27

⚡ Powered by xgo.ing

K3 > GLM 5.2?

Julien Chaumond (@julien_c) — Fri, 17 Jul 2026 11:56:33 GMT

K3 > GLM 5.2?

💬66🔄0❤️291👀37138📊100

⚡ Powered by xgo.ing

who used the term "load-bearing" first? humans, or agents?

Julien Chaumond (@julien_c) — Thu, 16 Jul 2026 22:00:03 GMT

who used the term "load-bearing" first? humans, or agents?

💬10🔄1❤️45👀7711📊16

⚡ Powered by xgo.ing

TruffleHog now supports scanning secrets in @huggingface Storage Buckets. As AI agents increasingly...

Julien Chaumond (@julien_c) — Thu, 16 Jul 2026 18:30:40 GMT

TruffleHog now supports scanning secrets in @huggingface Storage Buckets.

As AI agents increasingly crawl code, datasets, and storage, preventing leaked credentials is becoming more important than ever.

Great collab with the @trufflesecurity team! 🙌

💬3🔄7❤️18👀2711📊8

⚡ Powered by xgo.ing

OMFG this lineup 😍 I'm officially starstruck.

Julien Chaumond (@julien_c) — Thu, 16 Jul 2026 17:32:17 GMT

OMFG this lineup 😍

I'm officially starstruck.

merve@mervenoyann

btw yesterday we thought we'd only meet Soumith 🐐 but we were surprised by John Schulman joining 🐐 

so bts about this is a lore 🌝😄

🔗 View Quoted Tweet

💬3🔄4❤️58👀8171📊9

⚡ Powered by xgo.ing

For a first model release, @thinkymachines Inkling is quite a good model. Realize, it's infinitely ...

Julien Chaumond (@julien_c) — Thu, 16 Jul 2026 10:46:01 GMT

For a first model release, @thinkymachines Inkling is quite a good model.

Realize, it's infinitely more difficult to ship your first open model than your second/third/etc one.

If they continue on this path I expect them to reach or surpass the open frontier in the coming months

💬9🔄11❤️211👀8249📊37

⚡ Powered by xgo.ing

Hi @grinich any interest in a quick collab to add @huggingface OAuth to https://workos[.]com/docs/in...

Julien Chaumond (@julien_c) — Wed, 15 Jul 2026 21:31:59 GMT

Hi @grinich any interest in a quick collab to add @huggingface OAuth to https://workos[.]com/docs/integrations? 

We have shared customers who'd be into this!

cc @WorkOS

💬0🔄0❤️12👀3650📊2

⚡ Powered by xgo.ing

Is this a good or a bad thing?

Julien Chaumond (@julien_c) — Tue, 14 Jul 2026 20:47:40 GMT

Is this a good or a bad thing?

Sam Altman@sama

this would have been a whole startup not too long ago

🔗 View Quoted Tweet

💬9🔄2❤️66👀22074📊15

⚡ Powered by xgo.ing

Many such things

Julien Chaumond (@julien_c) — Tue, 14 Jul 2026 19:49:52 GMT

Many such things

Teknium 🪽@Teknium

Open source harness you say

🔗 View Quoted Tweet

💬0🔄1❤️14👀6400📊2

⚡ Powered by xgo.ing

Anyone knows how i can re-install 'ChatGPT Classic'

Julien Chaumond (@julien_c) — Tue, 14 Jul 2026 17:45:21 GMT

Anyone knows how i can re-install 'ChatGPT Classic'

💬7🔄0❤️15👀5534📊9

⚡ Powered by xgo.ing

I still remember life before chain-of-thought reasoning models

Julien Chaumond (@julien_c) — Mon, 13 Jul 2026 10:36:06 GMT

I still remember life before chain-of-thought reasoning models

💬27🔄14❤️393👀21748📊76

⚡ Powered by xgo.ing

quite sad that chatgpt is killing their simple mac os app and switching to a bloat Doubly disappoin...

Julien Chaumond (@julien_c) — Mon, 13 Jul 2026 09:14:39 GMT

quite sad that chatgpt is killing their simple mac os app and switching to a bloat

Doubly disappointed because they also killed Pulse and I used to love it!!

💬15🔄2❤️60👀10480📊23

⚡ Powered by xgo.ing

Who will be the “deepseek of the west”?

Julien Chaumond (@julien_c) — Fri, 10 Jul 2026 22:34:44 GMT

Who will be the “deepseek of the west”?

Anthony Pompliano 🌪@APompliano

We need open source, American AI models to become the default as quickly as possible.

🔗 View Quoted Tweet

💬26🔄4❤️98👀42613📊34

⚡ Powered by xgo.ing

Shiiiiit in europe you need to pay Tesla premium connectivity in order to unlock AC

Julien Chaumond (@julien_c) — Thu, 09 Jul 2026 17:37:19 GMT

Shiiiiit in europe you need to pay Tesla premium connectivity in order to unlock AC

💬10🔄0❤️44👀13319📊15

⚡ Powered by xgo.ing

Llama.cpp is seamlessly integrated into @zeddotdev version 1.10 🔥 Completely seamless with model a...

Julien Chaumond (@julien_c) — Thu, 09 Jul 2026 12:27:31 GMT

Llama.cpp is seamlessly integrated into @zeddotdev version 1.10 🔥

Completely seamless with model auto-discovery etc! 

Why even hit a remote API now?

Let us know if it works well for you ✅

💬4🔄4❤️24👀2804📊9

⚡ Powered by xgo.ing

Congrats to the @zml_ai team on the release of LLMD One thing I absolutely love is that the model ...

Julien Chaumond (@julien_c) — Wed, 08 Jul 2026 10:05:36 GMT

Congrats to the @zml_ai team on the release of LLMD 

One thing I absolutely love is that the model layers are loaded directly from the network (from @huggingface OFC), in streaming mode (no local file) which makes a lot of sense for production deployment

x.com/zml_ai/status/…

ZML@zml_ai

x.com/i/article/2074…

🔗 View Quoted Tweet

💬7🔄6❤️40👀15356📊13

⚡ Powered by xgo.ing

Anthropic releasing open source demos built on top of Qwen (with @neuronpedia) is not something that...

Julien Chaumond (@julien_c) — Tue, 07 Jul 2026 14:12:44 GMT

Anthropic releasing open source demos built on top of Qwen (with @neuronpedia) is not something that I was expecting

Anthropic@AnthropicAI

We also partnered with Neuronpedia to create an interactive demo of our methods on open-weights models.

Try it here: neuronpedia.org/jlens

🔗 View Quoted Tweet

💬22🔄33❤️605👀117765📊119

⚡ Powered by xgo.ing

True

Julien Chaumond (@julien_c) — Fri, 03 Jul 2026 18:18:22 GMT

True

Kirill Skrygan@kskrygan

Unpopular opinion: While everyone is so hyped about Fable, GPT5.6 and other huge and expensive models, I think the real hero of the last few months is *Qwen 27b*.

Our ML/AI engineering teams are have been very surprised how good/fast this open model is.

🔗 View Quoted Tweet

💬1🔄2❤️56👀13229📊8

⚡ Powered by xgo.ing

Which is it going to be?

Julien Chaumond (@julien_c) — Fri, 03 Jul 2026 13:35:26 GMT

Which is it going to be?

💬14🔄2❤️53👀9927📊23

⚡ Powered by xgo.ing

if even palantir advocates for open models now...

Julien Chaumond (@julien_c) — Thu, 02 Jul 2026 11:58:57 GMT

if even palantir advocates for open models now...

Your browser does not support the video tag.

🔗 View on Twitter

💬17🔄13❤️181👀16640📊46

⚡ Powered by xgo.ing

Remember that recent llama.cpp builds can dispatch GEMMs directly to the Blackwell FP4 Tensor Cores.

Julien Chaumond (@julien_c) — Tue, 30 Jun 2026 22:23:39 GMT

Remember that recent llama.cpp builds can dispatch GEMMs directly to the Blackwell FP4 Tensor Cores.

Lotto@LottoLabs

Ohhh we need some Blackwell bros to run this!

Nvidia coming through with the goat quants

huggingface.co/nvidia/Qwen3.6…

🔗 View Quoted Tweet

💬4🔄3❤️48👀12154📊11

⚡ Powered by xgo.ing

Throwback Tuesday

Julien Chaumond (@julien_c) — Tue, 30 Jun 2026 22:14:51 GMT

Throwback Tuesday

Hugging Models@HuggingModels

RoBERTa is here, and it's not just BERT's cooler sibling. It's a fill mask model that predicts missing words with next level accuracy. Perfect for text completion, search, and creative writing. 11 million downloads can't be wrong.

🔗 View Quoted Tweet

💬4🔄1❤️29👀9510📊7

⚡ Powered by xgo.ing

❤️❤️

Julien Chaumond (@julien_c) — Tue, 30 Jun 2026 22:08:23 GMT

❤️❤️

matheus.@matheusbuild

peut on imaginer une vie sans hugging face 🤗 
Il n’y a plus de retour en arrière

🔗 View Quoted Tweet

💬0🔄0❤️20👀4531📊2

⚡ Powered by xgo.ing

So, we are back to 2016 after all

Julien Chaumond (@julien_c) — Tue, 30 Jun 2026 20:55:36 GMT

So, we are back to 2016 after all

Soumith Chintala@soumithchintala

Bridgewater, one of the worlds largest hedge funds, a Tinker customer talks through how they've carefully fine-tuned a model focused on what makes interesting financial news.
Their fine-tuned model is more effective and cheaper than any frontier model.

🔗 View Quoted Tweet

💬12🔄9❤️230👀41321📊42

⚡ Powered by xgo.ing

just dropped https://t.co/sTPbkaWxfo

Julien Chaumond (@julien_c) — Tue, 30 Jun 2026 16:57:59 GMT

just dropped 

huggingface.co/nvidia/Qwen3.6…

💬18🔄71❤️723👀110024📊186

⚡ Powered by xgo.ing

Basé.

Julien Chaumond (@julien_c) — Tue, 30 Jun 2026 15:55:05 GMT

Basé.

Your browser does not support the video tag.

🔗 View on Twitter

💬0🔄3❤️53👀7232📊8

⚡ Powered by xgo.ing

…. For now

Julien Chaumond (@julien_c) — Tue, 30 Jun 2026 14:46:52 GMT

…. For now

elie@eliebakouch

le chaton long > le chaton fat for now

🔗 View Quoted Tweet

💬2🔄1❤️37👀4868📊5

⚡ Powered by xgo.ing

@xpasky @huggingface https://t.co/x0JeewqfR6

Julien Chaumond (@julien_c) — Tue, 30 Jun 2026 12:44:54 GMT

@xpasky@huggingfacehuggingface.co/changelog/filt…

💬0🔄0❤️5👀1887📊1

⚡ Powered by xgo.ing

💡

Julien Chaumond (@julien_c) — Sun, 28 Jun 2026 08:21:26 GMT

💡

Vladik@Kostoglodov

@Hesamation

🔗 View Quoted Tweet

💬25🔄55❤️1003👀72814📊146

⚡ Powered by xgo.ing

new logo wyt

Julien Chaumond (@julien_c) — Sun, 28 Jun 2026 08:09:46 GMT

new logo wyt

💬63🔄49❤️780👀38776📊168

⚡ Powered by xgo.ing

I’ve been saying

Julien Chaumond (@julien_c) — Sat, 27 Jun 2026 07:44:12 GMT

I’ve been saying

Yuchen Jin@Yuchenj_UW

The biggest baller move Sam could make right now is to open source GPT-5.6 on Huggingface and declare that OpenAI’s original mission has been achieved.

🔗 View Quoted Tweet

💬2🔄0❤️42👀7140📊6

⚡ Powered by xgo.ing

Who tf names models at openai 🤔

Julien Chaumond (@julien_c) — Fri, 26 Jun 2026 19:53:30 GMT

Who tf names models at openai 🤔

💬58🔄5❤️202👀28830📊83

⚡ Powered by xgo.ing

the circle of life I hope we won't see this in 2026

Julien Chaumond (@julien_c) — Fri, 26 Jun 2026 14:51:20 GMT

the circle of life

I hope we won't see this in 2026

💬7🔄4❤️66👀7090📊17

⚡ Powered by xgo.ing

dl stuff from hf while you can? 🫥

Julien Chaumond (@julien_c) — Fri, 26 Jun 2026 13:19:25 GMT

dl stuff from hf while you can? 🫥

Andrew Curran@AndrewCurran_

This also means Chinese models are almost certainly going to be restricted in some way - possibly even banned - in the West. Right now they're roughly nine months behind. If every American frontier release is forced into a slow, staggered rollout from now on, they'll start closing that gap immediately. Over time, this destroys Western labs business models, because users will eventually have access to genuinely equivalent open-source alternatives - and in time, even better ones! The US government will not allow that to happen. The only way to prevent it is if China agrees to the same staggered release schedule (unlikely but I suppose it's possible), keeping the gap roughly where it is today. Otherwise, the US will likely restrict or ban Chinese models. There's probably a nine-month window, at most, probably a lot less, before one of those outcomes plays out.

🔗 View Quoted Tweet

💬9🔄3❤️66👀9710📊15