#AI Researcher | A jumped-up pantry boy who never knew his place
Joined April 2022
- Tweets2.7K
- Following259
- Followers538
- Likes59.9K
This is much less fun than the math mania
AAAAWWOOOOOOOGAAAAAAAAA
A high severity vulnerability in "the most widely deployed DNS software on the internet" just dropped
kb.isc.org/docs/cve-2026-943…
🚨🚨🚨 ALERTE SÉCURITÉ MAXIMALE ! Python est compromis.
🔸 Mettez immédiatement pip à la version 26.2.2 ou plus.
🔸 Dans chaque projet Python (chaque venv) : python -m pip install --upgrade pip
🔸 uv : uv self update (0.13.4 ou plus)
🔸 Mettez aussi à jour les certificats du système (apt / dnf upgrade ca-certificates) et vos images Docker
pip install est compromis.
L'autorité indienne eMudhra (emSign), qui signe les « passeports » prouvant qu'un site est bien celui qu'il prétend être, utilisait des clés secrètes ridiculement faibles : 8 clés, toutes inférieures à 256. L'une valait 4. Autrement dit, n'importe qui peut fabriquer un faux passeport valide pour n'importe quel site.
Le risque pour Python : quelqu'un sur ton réseau (wifi public, FAI, État) se fait passer pour pypi.org, et ton « pip install » télécharge un paquet piégé sans aucune alerte. Aucune attaque n'a été signalée pour l'instant, mais c'est trivial à faire.
Un paquet piégé peut installer n'importe quoi. Porte dérobée, virus, vol d'informations en tout genre.
La situation est terrifiante !
Rien que dans l'écosystème Python, chaque installation de CPython, chaque venv, uv, tout est exposé.
Et je ne vous parle même pas de npm, pnpm, yarn, Go, Rust.
Ma stratégie de sécurité 2026 : rester chez moi et ne me connecter à Internet que depuis un FAI de confiance.
C'est quand il y a de la mer qu'on reconnaît les vrais marins...
Readers added context they thought people might want to know
Ce n'est pas un bug Python mais une vulnérabilité des clés privées faibles des certificats intermédiaires eMudhra/emSign (d<256 dont une valant 4), exposant les bundles certifi/pip à des MITM TLS potentiels jusqu'aux mises à jour.
discuss.python.org/t/information-…
emudhra.we-have-a-problem.com
scotthelme.co.uk/eight-bits-of-…
github.com/advisories/GHS…
Wow! Claude really refuses to work in August
New on our Engineering blog: Creating in-house source control.
In May, we merged around 2,500 code changes per day. In September, we merged more than 30,000 per day. As a result, we became bottlenecked by developer infrastructure built for human (rather than agent) scale.
Here, we share how we built a new version control system—and fully migrated our monorepo from git—in five months: anthropic.com/engineering/re…
Am I having a stroke or is there something i misunderstand -- i understand its a combo of pr, humor, and actual cool math endeavor (based on the different ppl that engage with it) -- but at these exponents, n*(log n)^(1-k) is not interesting before k reaches the proximity of 1/2
They should have a secret value baked into the model's knowledge and only if the model sees it as a response to a specific tool call can it assume it is in a simulation. Any other case, no matter how convinced it is, it must operate under the assumption it is not.
Somewhat related: someone can create a polymarket replica where people play against a continually training llm until it learns to predict better
People seem to believe (maybe: want to believe?) that Anthropic's "dooming" is a PR shtick, while it has been from day 1 aligned with this perspective and never done anything that contradicts that (read: the fact that they might have other missions too is not a contradiction)
The question whether their dooming is a reasonable concern, now, then, or in an hypothetical world is beyond the point.
And to alleviate any suspicion, Im not affiliated with Anthropic in any way, and even slightly begrudge them for never citing me even when appropriate.
Guy Dar retweeted
What can a neuron compute?
Real biological neurons are complex, but how capable are they?
Using a new method, we found that a single cortical neuron can classify cats vs dogs, recognize spoken words, and solve 10-bit parity, all tasks thought to require entire networks. (1/15)
Guy Dar retweeted
What's in a neuron? 💫 (an atypically long, almost personal post)
Neurons in LMs have always been a fascinating object to study. I've been studying them since 2020, viewing them as key-value memory cells, analyzing what they capture in vocabulary space, and how they compose together to form features.
arxiv.org/abs/2012.14913
arxiv.org/abs/2203.14680
arxiv.org/abs/2501.08319
arxiv.org/abs/2506.10920
But many neurons still remain opaque! They do many things.
Our recent work led by @AsafAvrahamy tackles this challenge by decomposing neuron weights in vocabulary space. We do this by taking the neuron weight vector and learning different ways to rotate it (just a bit) to reveal monosemantic vocabulary channels that it captures. The nice thing about our method ROTATE is that it's data-free and super efficient, relying only on vocabulary kurtosis as a search signal.
I've been thinking about this idea since 2024, proposed it to multiple students, but only Asaf was brave enough to take this ;)
Very happy with the final outcome. Check out the paper! 👇
arxiv.org/abs/2604.06005
The Kaon thing is elegant, you dont need exact orthoginalization, you can replace SV with random values. It makes sense bc whats central to Muon is to upscale weak directions not getting to SV exactly 1. The point is to fast track low SVs to noticeable values >>
Even replacing singular values with random noise works in optimization. After considering Schatten norm updates that interpolate between SGD and Muon, we could push it further towards the quasi-norm regime. And interestingly it works.
The main point here is that the geometrical explanation maybe does not work.
Instead it is possible to think about the trade-off between alignment of gradients and potential of descent. If we sacrifice a bit of alignment but gain a large amount of descent with a proper learning rate it leads to acceleration.
I think that's kinda the same point that this comment is getting at
xitter.cf/i/status/2054508313723…
