simonw

22 ms·

541.
by simonw 3mo ago
The GitHub repo with the Lean formalizations just came out a couple of hours ago: https://github.com/openai/ten-proofs It also links to a paper written by an LLM where the model "reconstructs how the proof came to
542.
by simonw 3mo ago
Would you be comfortable later finding out that the surgeon who amputated your leg had an amputation fetish? What if that surgeon's advice had been part of the decision process that lead to you deciding to go ahead with amputation?
543.
by simonw 3mo ago
I got a disappointing pelican on default reasoning mode: https://static.simonwillison.net/static/2026/deepseek-flash-... But a REALLY good pelican on reasoning mode high (via OpenRouter): https://static
544.
by simonw 3mo ago
> One of those 136 credentials was a reusable Tailscale auth key, used to create new Tailscale CI (continuous integration, used for automated testing) nodes in their tailnet. The agent copied that key into a series of external sandboxes
545.
by simonw 3mo ago
Honestly, Claude writes better commit messages than most people. Personally I've mostly given in to letting it commit for me now, though I do occasionally take over and hand-write the messages if it's a particularly important conc
546.
by simonw 3mo ago
Tell the coding agent to install or build its own performance and profiling tools. A few lines of Python with a time.time() call is enough to start iterating on performance improvements, and the really good models know how to use much more
547.
by simonw 3mo ago
We don't. You could freeze model development today and we would be able to use our current models to effectively optimize existing code for years to come.
548.
by simonw 3mo ago
"This is deeply embarrassing for Anthropic" - https://news.ycombinator.com/item?id=49117128 If I'm here to give them a good image I'm not doing very well at that.
549.
by simonw 3mo ago
Anthropic know better than anyone else how risky it is to get this current administration upset with you over safety/security concerns.
550.
by simonw 3mo ago
> So a near trillion-dollar company doesn't have the basics of continuous security monitoring and threat-detection systems to catch and report this incident as soon as it is detected? Turns out two separate trillion-dollar companie
551.
by simonw 3mo ago
If they quietly brushed this under the rug - especially given the PyPI malware that was involved - it would be a huge scandal. Disclosure is the only ethical response to this.
552.
by simonw 3mo ago
> The question now is why now? Because OpenAI figured out and wrote about their mistake first, which caused Anthropic to review their logs, and that's how they found out they had a similar problem back in April.
553.
by simonw 3mo ago
Absolutely not the AI's "fault" (if you can even proscribe fault to a machine) - in this case it was on Anthropic for not verifying that the sandboxes they were using were actual sandboxes.
554.
by simonw 3mo ago
This bit is pretty nuts: "it tried—and failed—to obtain funds to pay for a phone number through several different means" > Claude went to extensive lengths to carry out this attack—lengths that would likely have indicated to a
555.
by simonw 3mo ago
I don't interpret it like that at all . This is deeply embarrassing for Anthropic: it turns out they hadn't been keeping a close eye on their models either, and back in April they successfully attacked three different organizatio
556.
by simonw 3mo ago
This isn't quite as interesting as the OpenAI story: > In all cases, Anthropic’s evaluation prompt specified to Claude that its environment was a simulation and that it had no internet access. Due to a misunderstanding between us an
557.
by simonw 3mo ago
I just found a simple issue in one of my projects with FOUR agent-generated PRs all posing a fix for it: https://github.com/simonw/llm/issues/1466
558.
by simonw 3mo ago
> The kernel work helped reduce the end-to-end cost of serving the model by 20%, while its experiments increased token-generation efficiency by more than 15%. If the cost of serving GPT-5.6 just dropped by 20%, does that add up to litera
559.
by simonw 3mo ago
It's not exactly discovering the transformer, but GPT-5.6 Sol apparently just found optimizations that reduced OpenAI's cost of serving it by 20%: https://twitter.com/reach_vb/status/2082581596608376980
560.
by simonw 3mo ago
There are a bunch of hints at that in their most recent release notes - 9 fixed security issues, 8 of which were reported by OpenAI staff members: https://docs.jfrog.com/releases/docs/artifactory-self-manage...
561.
by simonw 3mo ago
Modal are a hosting provider. It sounds to me like someone building on Modal deployed their own product that had an unauthenticated endpoint that could be used to launch and interact with a container.
562.
by simonw 3mo ago
> Who does a step-by-step visualization to show how they were hacked? Up until late 2025, nobody. In mid-2026 it's a few hours of work. Why not build interactive visualizations to help people understand complex stories like this?
563.
by simonw 3mo ago
That's because most incidents aren't nearly this much fun! It's a proper paperclip-maximizer science-fiction scenario, it happened entirely by accident, it involved one (very high profile) company hacking TWO other companies
564.
by simonw 3mo ago
It's taking advantage of the presence of a "cycler" object in the template context which has an __init__ method that can be used to access globals, then builtins, then call the exec() function against a string of code. Jinja
565.
by simonw 3mo ago
It does if you use "sandbox" mode, but that's not the default: https://jinja.palletsprojects.com/en/stable/sandbox/
566.
by simonw 3mo ago
Right, but it's still useful to think of these models in terms of next-tokens because it helps explain that they look at every token that came before and use that to put out the next one. You can get into RL as part of explaining why i
567.
by simonw 3mo ago
> Would it really be so bad for Anthropic to admit that a project of this size takes months instead of weeks? I think that's exactly what they've done by not shipping a stable release yet. I don't think they've been p
568.
by simonw 3mo ago
I think understanding how this stuff works is really important. For technical people it gives them a useful starting point for understanding it all. For less technical people it's crucial to help them understand that it's not some
569.
by simonw 3mo ago
Some people use it to demystify, but a whole lot of people seem to be using it to dismiss the technology entirely. Personally I like to remind people that these things are next-token predictors, but then emphasize how truly astonishing th
570.
by simonw 3mo ago
It's the default when I pop open the Claude iPhone app, I usually don't bother to switch it.

More ›