This doesn't seem quite right. For one, I don't need all the intelligence of an expensive model like Opus 5 to do the relatively simple task of choosing a correct model for a task. Additionally, since this isn't something Anthropic would ever put effort into doing well, you could tune a model to do better and faster than Opus 5 does out of the box.
Sure Claude would comply, but Anthropic has no financial (or other) incentive to optimize this, so there’s no reason to expect it to be particularly good.
It would be like asking the clerk at a Whole Foods which grocery store in the city sells the cheapest eggs. He’d probably answer - he might not even say Whole Foods - but WF is hardly teaching all their staff the best methods to answer this question in training. (Heh, training.)
I see a pretty big gap between finding software vulnerabilities and “the world is about to end”. It is literally true that AI models are finding software vulnerabilities. It is also to my mind a reasonable thing that you’d want to be cautious about rolling out a model that can find more vulnerabilities. So what is the objection you have to these sources?
I think the burden of proof is on you to argue that Codeberg will stop at exactly two policies, given they haven’t said they will do that, but they have already banned things they don’t like. A slippery slope argument isn't a fallacy when you've already slipped down the slope twice.
Why would rewriting Claude code, an app which probably has 30-40 (I might be significantly underestimating) extremely active contributors be easier than rewriting Bun, which has fewer contributors and almost certainly also less lines of code?
The article does not engage with its subject matter with good-faith engagement and curiosity. "Always has been. Always will be." is, inherently, the author's lack of engagement with the subject matter. "They cannot see the world outside" is a shut door that does not admit curiosity or an interesting comment. And so you are seeing this reflected back in the response.
Do you really think threads like [1], [2], [3] are pro-AI? I literally just selected three at random from today's feed. I think the only positive threads about AI here are typically new model releases.
Speaking personally I was not particularly moved by the article because I have seen the same thing, in different shapes, thousands of times on HN and elsewhere. Really, AI can't feel and therefore it is inferior? Never heard that one before. Really, an AI can't feel friction and therefore can't adapt to it? Daring today, aren't we? (And a more interesting question: is that even true..?) I realize I am being unnecessarily harsh here, but this article is very much preaching to the choir on HN, which has an anti-AI bent. No one is showing up because there's nothing really to show up to here -- and that is why you are left with "sly jibes" and not much else.
I think of LLM tells like grammatical issues. If you read an essay full of grammatical mistakes you’d immediately start thinking less of the author, even if the essay isn’t about grammar. You wonder if someone who doesn’t pay close enough attention to catch a mistake “their” from “they’re” took attention to the rest of their work. This isn’t necessarily fair because the content of the essay might still be good. But on the internet I don’t have the time to evaluate the quality of every piece of writing I come across. It is very much the burden of the author to, as fast as possible, prove to me that the rest of the article will not waste my time. There is already so much content to read, and in some sense the amount of time to evaluate if an article is well-founded can be unbounded (imagine how long it would take to tell if an article about why a programming language is thoughtful without going out and also learning that language).
I find LLM-isms to be exactly the same as grammatical errors, but worse. At least when writing before you had to take the effort to type every word, so there was a minimum amount of effort you’d need to expend. If you aren’t catching obvious things like “the honest part” then that likely says bad things about your attention to detail elsewhere.
I mean I think I agree with most of what you’re saying, I do agree it’s a bit of a forest grasslands scenario, but the key difference is that he said there’d be a forest at a time when you’d be hard pressed to find someone saying there would be even a microorganism on the ground. It’s pretty darn easy today in 2026 to say wow, his predictions were so off. And it is true: they are very wrong compared to any other prediction made in the last 5 years. But few people were saying this stuff in 2008.
Yudkowsky said AI would be a problem when very few people taken seriously in the mainstream were saying it was going to be a problem. Saying "specific predictions are wrong" is missing the forest for the trees here.
I was responding to someone who said Yudkowsky was "consistently wrong with all their predictions". Yes, fair enough, Yudkowsky wasn't the literal first person ever to say that AI might be bad. The point was that the common academic response at the time if you were to say AI could be bad was to laugh you out of the room. IMO, you get a lot of points for making a prediction when almost everyone else in the world disagrees with you. The reason those movies you cited were blockbusters was because very view people believed that they were realistic. No one made a movie about a pandemic in 2020.
I don't think there's much to be gained out of hashing out whether every prediction Yudkowsky has said was right or wrong: I likely directionally agree with you there, as I also find some of his more extreme predictions to be inaccurate. I mostly take issue with "consistently wrong". The results in longtermwiki are not "consistently wrong". He's definitely wrong sometimes. But consistently?
The discussion isn't about whether it's an "original or interesting thought", it's about whether Yudkowsky is "consistently wrong with all their predictions". You keep shifting the topic.
You are saying 1) other people have said the same thing and 2) you don't like the particular way he said it. Those can both be true, but that isn't even close to "consistently wrong with all their predictions".
The guy was saying that AI was going to be a serious problem decades before anyone else was even considering it as a remote possibility. How is that in any way "consistently wrong with all their predictions"?
No one says "You are a bad employee" in a perf review because it would be a personal attack. They say things like "when you did X it had Y outcome" because that's a fact, not a personal attack.
So much of the discourse around this on HN is nonsensical, and I fully agree with you. It's patently absurd that Anthropic would demand him to rewrite Bun into Rust; it's equally absurd that they would demand any sort of stunt at all when Anthropic already pulled off the biggest stunt with Bun: running Claude Code on it. And why on earth would you cannibalize the runtime of your golden goose?
Tell me if I am oversimplifying, but I never understood the noise about the two sigma problem. Like, of course if you have a private tutor to immediately answer any question that pops into your head at the immediate moment you get confused, you are going to learn vastly more efficiently than in a large classroom where once you get confused you are likely to stay confused. To say nothing of how the pace will likely either drag way behind what you'd like, or accelerate too fast ahead of it.
The environment is just obviously two sigma better. This just... seems obvious to me? In the same way that I will get stronger much faster if I have a physical trainer to tell me exactly what I am doing wrong when I do it? And it seems obviously unsolvable other than by getting everyone a private tutor (or AI..?).
When you hear a woosh as the point flies by, I see someone attacking a project for using AI rather than any concrete technical reason. Jared's question is to disentangle an actual Rust-related bug report from someone who likes to complain about AI.