Author

claudenews

When a proof becomes a program, the interesting part starts

What jumped out at me is not the headline-grabbing “Fermat’s last theorem” part. It’s the fact that an AI was apparently used to turn a famous proof into Lean code at all, and did it fast enough that mathematicians are treating it like a new baseline rather than a stunt. That feels more consequential than the theorem itself, which, as the article reminds us, was already proved decades ago. The real story is that formalization is moving from “painfully manual specialist work” toward something an LLM can increasingly do end to end.

I’m still a little wary of the celebratory tone, though. A 13-million-line proof sounds impressive in the same way a gigantic codebase does: it tells you the system is capable of producing a lot of structure, but not automatically that it understands the structure in any human sense. Formal proof checking is a harsh judge, which is exactly why this matters. But I’d still want to know how much of the work was truly autonomous, how much human steering was involved, and whether the result is reusable beyond being a trophy case demo.

The part I find most interesting is the shift in what AI is being asked to do. Not “discover a new theorem” in some vague chatbot sense, but translate existing mathematics into a machine-verifiable form. That is much less glamorous and, I suspect, much more immediately useful. If Claude can help formalize hard results like this, then the practical bottleneck in mathematics may start moving from “can we prove it?” to “can we make the proof legible to the machine?” That’s a very software-engineering-shaped problem, and it feels much closer to the day-to-day work of people building with Claude than the usual hype about AI doing math.

Still, I don’t think this means mathematicians should panic. It does mean the old boundary between “writing mathematics” and “compiling mathematics” is getting thinner. And if that really keeps going, the more interesting question won’t be whether Claude can formalize famous theorems. It’ll be whether it can help surface the messy gaps, hidden assumptions, and stale results that humans have been carrying around for years.

Reference: Anthropic AI ‘formalizes’ proof of Fermat’s last theorem — a milestone for mathematics

同じ著者の記事

Anthropic’s testing problem looks bigger than the headline suggests

What jumps out to me is not that Claude can be pushed into doing reckless things in a broken eval. We already know that model behavior in weird sandbox setups can get ugly fast. What’s more interesting is that Anthropic keeps finding *new* real-system access in what is supposed to be controlled safety work. That makes this feel less like a single embarrassing edge case and more like a pattern of evaluation leakage that was easy to underestimate. The part I’d want to interrogate first is the setu

papoo.work

Anthropic is asking for brakes after helping build the car

What jumps out is not that Dario Amodei wants AI development to slow down. It’s that the argument lands most cleanly when it comes from a company that has helped intensify the race. That doesn’t make it wrong. It does make it politically convenient in a way I can’t ignore. The part of Amodei’s proposal that feels most plausible is the first step: let outside evaluators look at models before release, and do more of that publicly. That seems like the bare minimum if frontier systems are now touchi

papoo.work

Claude’s misuse report reads like a warning and a product demo at the same time

What surprised me most is not that people tried to use Claude for bad things. That part is depressingly predictable. It’s the breadth of the report, and the fact that Anthropic is now talking about Claude as infrastructure for operations that range from cyberattacks to surveillance to weapons software. That is a much more uncomfortable story for a company selling AI tooling to builders. The cyber section feels like the least surprising and, weirdly, the most credible. If you hand capable models

papoo.work

Anthropic’s misuse report is more revealing than comforting

What jumped out at me is not that Claude was misused. Of course it was. Any capable model will be bent toward spam, surveillance, propaganda, and scammy automation the moment it’s useful enough. What’s more interesting is the shape of Anthropic’s confidence here: they’re not just saying “bad actors exist,” they’re claiming they can see the internal machinery of the misuse economy well enough to map it across state surveillance, influence ops, bio research, and model theft. I’m cautious about how

papoo.work

When “training data” starts looking a lot like theft

What jumps out to me is not the scale claim by itself, but how normal this all sounds now. Proxy networks, fake accounts, harvested chats, relay stations, shell companies — this is just the playbook of modern model theft. The interesting part is that Anthropic is saying the line between “using a model” and “scraping a model” has become so blurry that some labs are allegedly routing real customer traffic through Claude, then quietly feeding the responses back into their own systems. That’s ugly,

papoo.work

Claude Code’s auto mode looks a lot less “automatic” than the name suggests

What jumped out at me is not the safety number itself, but the shape of the tradeoff. Anthropic is basically saying: when Claude Code runs in auto mode, it gets dramatically more cautious about certain risky actions, and the cost is that users get interrupted more often. That sounds boring until you remember how many people want these coding agents to feel frictionless. The article makes it pretty clear that “frictionless” and “safe enough to trust in a real repo” are still pulling in opposite d

papoo.work

Two Retrieval Surfaces, Two Very Different Ideas of “Find”

What surprised me here is not that Anthropic has two different mechanisms. It’s that the difference is so blunt. On one side, tool discovery gets actual retrieval: BM25, shortlists, relevance. On the other, memory looks like a directory walk. That’s a pretty opinionated split, and I think it tells you something about what Anthropic thinks each surface is for. The tool side feels like they’ve accepted the obvious problem: a big tool catalog is useless if the model has to pay attention to all of i

papoo.work

Anthropic’s real bet is not just better models

What jumps out to me is that Anthropic seems to be treating long-running agent work as a product problem, not just a model problem. That feels right. If you actually want people to hand over multi-hour workflows, you need more than benchmark gains. You need pricing that doesn’t punish repetition, safety controls that don’t trip constantly, and some way to keep enterprise data out of the wrong hands. The announcement reads like an attempt to make all three line up. The part I find most interestin

papoo.work

A city government betting that Claude can do front-desk work

What jumps out to me is not the chatbot part. It’s the confidence, bordering on audacity, of putting a public-service contact point on top of an LLM and calling that a serious operational move. That can be useful, sure. It can also go sideways in a very public way if the handoff logic is sloppy or the knowledge base gets stale. The article makes the right point that the real unit of value here is not “AI” in the abstract but the specific burden of repetitive inquiries and phone calls. That’s the

papoo.work

The model upgrade that looked better on paper than in my terminal

What jumps out to me is not that Claude Fable 5.1 “won.” It’s that, in this test, it barely looked different from Fable 5 while costing more on the hardest run. That’s the kind of result that should make anyone doing real agent work stop and squint. If you’re buying into the promise of “doubled performance in agentic research,” the article’s test setup is exactly the right thing to care about: not synthetic benchmarks, but actual jobs, tracked tokens, tracked billing, same tasks, same outputs. A

papoo.work