cloudflare confirms mythos is a harness, not just a model
https://blog.cloudflare.com/cyber-frontier-models/
https://blog.cloudflare.com/cyber-frontier-models/
The Cloudflare Blog
Project Glasswing: what Mythos showed us
In recent weeks, we pointed Mythos and other security-focused LLMs at live code across critical parts of our infrastructure. We share what we observed, the models’ strengths and weaknesses, and what the work around them needs to look like before any of it…
🤔3
I am quite optimistic about AI empowerment, and one observation of George Hotz I definitely agree with -- /everything/ now depends on your ability to see if the agent has produced slop, and the habit to ruthlessly reject it.
The slop percentage is very high, 30-90%.
https://geohot.github.io//blog/jekyll/update/2026/05/24/the-eternal-sloptember.html
The slop percentage is very high, 30-90%.
https://geohot.github.io//blog/jekyll/update/2026/05/24/the-eternal-sloptember.html
the singularity is nearer
The Eternal Sloptember
I’m calling it now, the adoption of AI agents into software development will be one of the most costly mistakes in the field’s history. Agents cannot program, and it’s taking longer and longer to realize that they can’t. They are a highly sophisticated statistical…
❤4🥴1
https://en.wikipedia.org/wiki/Orient_Express_Corinthian
https://www.youtube.com/watch?v=-fhiDsLw8Sk
what a sailboat. 220m, recently launched, they say 15 knots under sail
https://www.youtube.com/watch?v=-fhiDsLw8Sk
what a sailboat. 220m, recently launched, they say 15 knots under sail
Wikipedia
Orient Express Corinthian
largest sailing cruise ship built 2025-2026
🐳1
Oops. Linux package repository takeover is like the nastiest kind of supply chain attack. Do audit your Arch right now
https://archlinux.org/news/active-aur-malicious-packages-incident/
https://archlinux.org/news/active-aur-malicious-packages-incident/
Arch Linux
Active AUR malicious packages incident
We are currently experiencing a high volume of malicious package adoptions and updates in the Arch User Repository.
😱3
side effects of fear mongering.
"including foreign national Anthropic employees"
https://x.com/AnthropicAI/status/2065597531644743999?s=20
"including foreign national Anthropic employees"
https://x.com/AnthropicAI/status/2065597531644743999?s=20
X (formerly Twitter)
Anthropic (@AnthropicAI) on X
The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees.…
👏2🤣2🌚1
Forwarded from gerontion
Can sleep’s core benefits be delivered without actually sleeping? Scientists induced sleep-like neuronal “off periods” during wakefulness in mice, reducing local sleep pressure, weakening synaptic strength, and even restoring memory consolidation during sleep deprivation. The findings suggest key functions of sleep may depend on specific brain activity patterns rather than sleep itself.
https://www.nature.com/articles/s41593-026-02318-9
https://www.nature.com/articles/s41593-026-02318-9
Nature
Induction of cortical on/off periods in awake mice fulfills sleep functions
Nature Neuroscience - Driessen et al. show that core benefits of sleep—reduced local sleep pressure, renormalized synaptic strength and memory consolidation—can be reproduced in awake,...
🔥2🤩1😴1
interesting, it seems there is a place in the brain where the entire body movements are exposed, unlike the theories followed previously
https://www.nature.com/articles/s41586-026-10653-x
https://www.nature.com/articles/s41586-026-10653-x
Nature
A mosaic of whole-body representations on the human precentral gyrus
Nature - A comprehensive map of the human motor cortex at single-neuron resolution is described.
👍1
Forwarded from χаотичні нотатки
All models are wrong, some models are useful, language models are useful too.
https://open.substack.com/pub/cortex/p/models-as-tools-for-thoughts?r=1clcn&utm_campaign=post&utm_medium=telegram
https://open.substack.com/pub/cortex/p/models-as-tools-for-thoughts?r=1clcn&utm_campaign=post&utm_medium=telegram
blog.cortex.im
Models as Tools for Thoughts
Language. Models. Hypotheses.
Experiments and the Confidence. Hallucinations.
Thinking with Large Language Models. Draft.
Experiments and the Confidence. Hallucinations.
Thinking with Large Language Models. Draft.
🥱1
a great writeup on the different facets of the software/systems architecture
https://shapeofthesystem.com/
https://shapeofthesystem.com/
The Shape of the System
Structure over vigilance: twenty-five tenets for engineering around the limits of human and machine cognition, building correctness into the shape of the system.
Palantir, of all places, vouching for sovereign AI!
https://x.com/palantirtech/status/2072114267776491695?s=46
https://x.com/palantirtech/status/2072114267776491695?s=46
X (formerly Twitter)
Palantir (@PalantirTech) on X
Our thoughts on the importance of AI sovereignty.
1. Your AI sovereignty dictates your institution’s future. Sovereignty is the precondition for choice. Relinquishing sovereignty transfers the future choices of your institution to others, who are likely…
1. Your AI sovereignty dictates your institution’s future. Sovereignty is the precondition for choice. Relinquishing sovereignty transfers the future choices of your institution to others, who are likely…
🤔1
if/when we'll get to the verified code generation, whether with LLMs or with formally proven generators, we'll still have vulnerabilities.
except they'll be in the task descriptions or the tests themselves
https://x.com/zellic_io/status/2073101446883823622?s=46
except they'll be in the task descriptions or the tests themselves
https://x.com/zellic_io/status/2073101446883823622?s=46
X (formerly Twitter)
Zellic (@zellic_io) on X
You "win" https://t.co/y2fdv7oJiN – ZKSecurity’s FV contest – by writing the most optimized formally verified circuit for cryptographic primitives.
But how does the best SHA256 circuit have a cost of 0?
Here's how we exploited the FV spec for the top score…
But how does the best SHA256 circuit have a cost of 0?
Here's how we exploited the FV spec for the top score…
incredible that i estimated opus at 3T params last year and now K3 is 3T and opus-level
- https://x.com/nathancgy4/status/2077817266725609544?s=46
- https://www.kimi.com/blog/kimi-k3
- https://x.com/nathancgy4/status/2077817266725609544?s=46
- https://www.kimi.com/blog/kimi-k3
X (formerly Twitter)
nathan chen (@nathancgy4) on X
man ignore the benchmarks and the race i just want there to always remain a frontier level intelligence model that is open and safe, available to everyone regardless of region
it is truly incredible to have built such an open (and strong) model with a group…
it is truly incredible to have built such an open (and strong) model with a group…
Schema, a harness that has frontier models play like physicists: write each game’s mechanism as an executable program, test it against reality, and plan inside it. The name is inspired by Immanuel Kant: a schema is a rule of construction that bridges from abstract concept to concrete perception.
https://schema-harness.github.io
https://schema-harness.github.io
the intensity of the push to /replace/ ECC with PQC instead of hardening puts doubt on the security of the whole PQ setup
https://x.com/hashbreaker/status/2079577558895726668?s=46
https://x.com/hashbreaker/status/2079577558895726668?s=46
X (formerly Twitter)
Daniel J. Bernstein (@hashbreaker) on X
Wow. After disenfranchising an unspecified list of people and making claims about the resulting numbers, the IETF TLS WG chairs refused to answer an evidence request from Fabiana Da Pieve, the European Commission's Team Leader Post-Quantum Cryptography: …
💯1
Bernard Arnault is no saint but this is a work of art
https://x.com/lulumeservey/status/2081583334434468262 (translation), https://x.com/LVMH_Presse/status/2081404246751109150 (original in french)
https://x.com/lulumeservey/status/2081583334434468262 (translation), https://x.com/LVMH_Presse/status/2081404246751109150 (original in french)
X (formerly Twitter)
Lulu Cheng Meservey (@lulumeservey) on X
Bernard Arnault going direct is remarkable and worth reading in full
The single most impressive thing is that Arnault (personally, not via company spokesman) responds to a major hit piece without at all accepting the role of defendant
He never complains…
The single most impressive thing is that Arnault (personally, not via company spokesman) responds to a major hit piece without at all accepting the role of defendant
He never complains…
🤨1
PSA: do tell your sysadmin:
nginx
nginx
CVE-2026-42533 — every release since 2011 affected, pre-auth RCE exploit with ASLR bypass available 🤯👍1😱1
daily PSA: DISABLE GIT AUTOSYNC IN YOUR EDITOR NOW.
There is an ongoing spray of attacks on (public but also private once somebody from your org gets infected) repositories that try to push code with exploits that fire once landed on your machine (malicious VSCode configs, etc)
There is an ongoing spray of attacks on (public but also private once somebody from your org gets infected) repositories that try to push code with exploits that fire once landed on your machine (malicious VSCode configs, etc)
😱3