It doesn't know that because it wasn't trained on any tasks that required it to develop that understanding. There's no fundamental reason an LLM couldn't learn "what it knows" in parallel with the things it knows, given a suitable reward function during training.
The part about the crucibles is somewhat interesting, if true, as it reminded me of a passage in the Manhattan Project engineering story that was posted the other day. They thought the melting point of plutonium would be a lot higher than it turned out to be, so they commissioned a lot of exotic crucibles [1] that turned out to be unneeded. They were apparently developed at MIT with cerium sulfide, a compound I'd never heard of before.
I see a few other posts by electric_muse that are LLM-suspicious, but it bugs me that the style is so much like my own writing. Only a matter of time until I'm the witch on trial, I suppose.
I don't really look at the writing style, as it's a very good way to fall for false positive. And I actually won't blame a non native speaker for editing their thoughts using an AI much more proficient in English than them.
Instead I look how much actual information (be it anecdotes, personal opinion, different perspective, etc.) a comment contains.
Here it's just a very straightforward TL;DR; of the original article, it contains no substance beyond that so I don't think anyone would bother write it (or at least they would likely advertise it as a TL;DR;).
A correct summary without anything added is a very valuable thing. I personally don't have the time to read every article, and highly value the scientific paper format, with an abstract given upfront.
A correct summary like this is what any LLM will give you for free so it's not that valuable anymore, but that would be OK if it was advertised as a TL;DR;, and not presented as if it was OP's personal take like it is right now.
It’s a mixture. But yeah, today I pushed the boundary further.
A few days ago I flagged a piece someone else had written with ai. It has a specific cadence and some typical patterns. But many people seemed to buy it before I commented.
So I’ve been running a bit of an experiment lately where I write posts and LLM-ify them in varying amounts to see if an LLM actually produce something more upvotable and when people start to notice.
I started out just saying “rephrase this so it sounds tighter” and moved recently towards just jotting rough notes and saying “make an HN comment out of this” and then editing.
Today was the most AI-influenced set of posts, and people clearly noticed, so that’s definitely the threshold. Fascinating.
Check my post history to see the evolution. Using gpt-5.
Please don't do this. HN is a community, not a laboratory. As a sibling commenter asked: [do you] consider whether or not the people you interact with here want to be experimented on?
People have been doing “experiments” with trying to sneak LLM-generated content onto HN since at least 2020. It's not new or clever. Yes, sometimes it will slip through our defences. After a while the community figures out. We give a collective groan and eyeroll, kill all the comments, ban the account if it’s egregious, and move on.
We have been asking the community not to publicly accuse commenters of posting generated comments, because sometimes the accusations are false and we think the negative consequences of false accusations outweigh the positive consequences of valid accusations.
But if we're going to ask that of the community, we also have to be very insistent that people do not exploit the community’s trust with experiments or stunts like this.
did you ever stop to consider whether or not the people you interact with here want to be experimented on, or whether or not it was right to betray the confidence of folks here and degrade the experience even further for people that come here to seek real human discourse?
not a criticism, I just want to know whether or not it ever crossed your mind.
I was super hesitant at first but thought that this might be the most accepting place of any. For the most part, my comments were getting tons of upvotes and replies and so I thought “wow this is furthering the conversation. I should keep going!”
I wasn’t outsourcing the entire process after all. I moved from asking it to rephrase things I wrote (I’m sure plenty of people use grammarly and the like) to asking it to give me some drafts with a specific opinion and viewpoint, and then I would edit to my liking.
Anyway, this totally blew up and now I regret it. But it was an interesting ride because it really opened some interesting questions about the fact that these were some of the comments I made that the community upvoted the most, which to me is a sign of contributing value.
The unreadability of that example has approximately nothing to do with code formatting, which is generally understood to refer to modifying the textual representation of the code while leaving the actual logic more or less unchanged. Can you propose some alternative whitespace or indentation scheme which would make that example significantly more readable?
Now do the comparison if one endpoint is realtime user facing traffic and the other is batch processing which can easily eat up all available capacity and drive up latency.
If visa shoppers are overwhelming the normal processing of applicants who actually live in a particular country, it seems entirely appropriate to say "no, sorry, this location isn't for you" to the people who don't live there.
How about hiring more people? For that matter, why is the decision made in the local consulate or embassy anyway?
The whole worldwide visa system feels like a relic of the time before networks. Many tasks rely on physically moving pieces of paper around, leaving your passport with agents of a foreign government (!), and having documents that are stamped and supposedly authenticated. Shouldn’t everyone be able to separate the tasks of authenticating a person and authorizing that person to be somewhere?
People still make games for old consoles occasionally as hobby projects, and those are usually released freely as ROM files. I'm not familiar with Japanese law, but in most countries that would constitute a fairly solid proof that there are legal uses to which an emulator can be applied and thus that emulation itself isn't inherently illegal.
"The basic principles of the Python ecosystem" are a dumpster fire to anyone who isn't already desensitized to the situation. Just like 'uv' as a whole, this seems like a meaningful step towards making Python a little less terrible to actually use, and should be applauded.
In my opinion an AGENTS.md file isn't an artifact at all (in the sense of being a checked-in part of the codebase), it's part of the prompt. You should gitignore them and use them to give the LLM a brief overview of the things that matter to your work and the requests you are making.
every example in the wild has these checked in and treated like a lock file where everyone is scared to touch it and introduce weird behavior.
personally i think this pattern is a dead end and trying to build deterministic agentic behavior on top of inherently non-deterministic systems is a fools errand.
I mean, by construction you're only ever going to see the examples where people checked them in and published that. It doesn't mean that other people aren't getting more use out of local instructions customized to their particular work.
You're in luck! If you search for "canadian visa application" the first two results are "Visit Canada" and "How to apply for a visitor visa" on https://www.canada.ca/
On my phone I get several screenheights of sponsored links, with oh so helpful sitelinks and summaries of each scam site. The fourth result is the legit one.
Interesting, I'm surprised the results vary so much for a query with such an objectively correct answer. I tested it on my desktop in private browsing mode because I wanted to see whether Kagi was much better, but figured I should report the null result when Google actually did fine.
I don't know if GPT-5 is an exception and is overcooked on XML specifically, but in general Markdown and XML seem to work about equally well for LLM inputs, the important part is just that they like hierarchical structured formats. The example on that page could probably be replaced with:
## Code Editing Rules
### Guiding Principles
- Every component should be modular and reusable
...
### Frontend Stack Defaults
- Styling: TailwindCSS
