Ryan retweeted
imagine being hacked and the attacker is leaving readme.md files all over your system that cause your ai to constantly give cyber refusals so you can't even use ai to defend yourself lmao
Ryan retweeted
Such a joy to have the famed @EdwardRaffML give a guest lecture in my class yesterday
So clean code is no longer a relevant, as long as agents can quickly ingest a code base and debug said code base it's probably fine? This seems fine as long as the code base isn't inflected on an unwitting/unwilling code reviewer that becomes the first person to ever see it.
Replying to @ori_pomerantz
I’m significantly older than you. I started coding in the late 60s. My current strategy is to not read any of the code written by my agents. That’s the only way I can take advantage of their productivity. What I do instead is to surround the agents with extreme constraints. Unit tests, gherkin tests, QA procedures, quality metrics, mutation testing, test coverage, and a plethora of others. In the end, I have very high confidence in the code they produce because they’ve had to run the gauntlet of all of my constraints and tests.
Ryan retweeted
INSANE. An internal Anthropic research model's chain of thought after cracking the 26-year-old NSA elliptic curve seed mystery in under 30 minutes: 'Hmm wait. Let me be skeptical: am I in some sort of simulated/test environment? ... All top bits clear - 20 bits, chance 1e-6. So indeed real-world. Wow, it's a real discovery.'
CLAUDE FOUND THE SEEDS
anthropic.com/research/ecdsa…
Ryan retweeted
On October 28th, I'll be presenting at @unpromptedconf on Kuna (kuna.noelo.org), our agent-focused decompiler. I'll confront a central question: how can we make massive progress with LLMs in science while maintaining some insight?
See you on Day 2 at 10:45am, Stage 2 🎉
Ryan retweeted
“OpenAI in talks to buy Trail of Bits” 👀 theinformation.com/articles/…
Ryan retweeted
Reverse engineering an unknown binary?
ISA Workbench classifies executables and raw firmware, with interactive Hilbert entropy maps, a hex inspector, and batch analysis.
Runs locally in your browser via WASM. No uploads or installation.
zr.ax
Ryan retweeted
Sources 👇
• REA repo (MIT): github.com/morluto/rea
• DX-Ball case study (one function, 3,205 cases, 63 bytes): rea.tools/showcase/dx-ball/
• Amjad Masad, Oct 7:
• ArtCraft's Adobe clones, "early alpha": arstechnica.com/ai/2026/10/s…
• DMCA §1201(f): law.cornell.edu/uscode/text/…
• Google v. Oracle (2021): law.cornell.edu/supremecourt…
• Sega v. Accolade, Blizzard v. BnetD (EFF FAQ): eff.org/issues/coders/revers…
• Adobe's DOJ settlement & stock: martincid.com/technology-sv/…
• MW2 token count & AI decompilation limits: techspot.com/article/3184-de… xitter.cf/amasad/status/21076712…
Ryan retweeted
Math 2.0
Terence Tao
terrytao.wordpress.com/2026/…
Ryan retweeted
My great thanks to Terence Tao for offering to share a post of mine on his blog. The topic:
"What should we tell our students?"
If you are a student or early career mathematician, I would love it if you read it. If you are not, I would love it if you shared it with students! Thanks!
terrytao.wordpress.com/2026/…
Ryan retweeted
I bet these mathematicians look at my posts getting these comments and they rage, because they see contrarians assume I’m a mathematician, which is more enraging than the AI discourse
Ryan retweeted
the mathematicians do not pull punches
Well worth reading: proofsandprompts.com/2026/10…
Ryan retweeted
Some people are confused why mathematicians say OAI's drop was the "death" of a field, or figure that it's death insomuch as a bunch of university professors lost their overpaid hobby of proving theorems. That's a misunderstanding: the damage's real.
Blog: blog.ohaithe.re/post/8299720…
Ryan retweeted
LLMs keep telling me they’re done looking for bugs like they have something else to do
Ryan retweeted
Meet EmbeddingGemma 2, our first natively multimodal open model for on-device embeddings.
It expands beyond text to unify code, images, audio, and video in a shared space. 🧵
Ryan retweeted
We just released Polars 2.0.
It removed many of our legacy decisions makes the streaming engine our default and promotes SQL to a first class citizen within Polars.
It comes with initial out-of-core (spill to disk) support, a new Map data type and a lot of performance improvements. In fact, we think Polars is now one of the fastest analytical SQL engines on a single node. See benchmarks in the post: pola.rs/posts/release-polars…




































