HN user
So having tons of AIs quoting various literary works and reproducing knock-offs of them has a positive effect on those books' sales?
I think you're wrong: there is absolutely damage to the authors and publishers from what the AI companies have done.
ROFLMAO ... you think this is a liberal problem?
Pretty short/dumb take.
It should be required that every pay-walled article posted here includes a paywall bypass link.
NO ONE IS PAYING THE WSJ FOR A SUBSCRIPTION TO READ THIS NOISE!
The "real" cost is what people pay, not what the AI companies claim.
Who is paying $200/day? My company has me on a $30/month plan (I assume they pay annually?). I use Claude constantly and only hit usage limits when I try to do 4+ projects at once.
I can't imagine why anyone would need almost 7x that.
Piracy created AI.
Its not a "study" at all, it's a marketing piece.
The founders left, and it went downhill as a result.
Wikipedia is supposed to be a reference. SO was supposed to help programmers program.
The former continues to succeed, with hiccups, because they remember their goal. The latter is dead, and it failed because it tried to be the former, forgetting their goal.
I disagree. Stack Overflow was made to succeed, not fail ... originally.
Then the founders left, and meta basically chronicles the echo chamber of mistakes made. It's not the intent: that was to help coders code. Meta was dominated by people who lost the intent, and wanted to build a library.
I feel like "do X untill tests are green" is sort of the prototypical /goal case.
The key thing is that normally Claude grades Claude's work, so you do get a"pinky promise, as all of Claude's training bias tells it to plunge ahead (even if that means "fixing" a test by commenting it out).
With /goal, a different Claude instance evaluates the goal. This is the robust version you described, and such an "adversarial Claude's" approach, with or without/goal, is exactly how you avoid pinky promises (from one Claude).
?
I'm pretending to part of a group I'm not? Is "people who read web articles on monitors" a group (to pretend to be a part of)?
Why not address the objective evidence the OP provided? To an impartial observer, it seemed quite overwhelming.
I think "If you’re using OpenRouter, you’re already looking to bypass frontier models." is false. Our company uses both Claude subscriptions and OpenRouter ... and a lot of what we use OpenRouter for is more Claude.
We do a little exploration with other models through it, but it's not at all accurate to say we use it because we are "already looking to bypass frontier models".
... or at least, no more than any other company that doesn't want to overpay for their tooling, but is basically happy (ATM) with the current state of Claude.
From the article it doesn't even sound like he cares about pelicans at all, and doesn't think they are a good way to compare models anymore ... but people are used to seeing the test now, and it does serve as a common "hello world" unit of work.
Forget aggressive, I just find the main text font harder to read (vs. ... you know ... a normal font).
I think the frontier AI companies think you're missing key details:
- they can still discover entire new untapped markets for AI (that, potentially, only their models can unlock)
- they can find (novel, unique to them) ways to drive down the cost of running their models
- they can provide other ancillary value (e.g. write better harnesses) because of their expertise, and then charge for that value
I'm probably missing a few bullet points also. However, none of those are moats (or at least not yet) ... I'd consider them more like bets. The frontier AI companies are betting "the house" on them, and if they pay off they could, hypothetically, make them financially competitive.
THIS ONE IS HORRIBLE!
Copy/paste is one of the most basic, low-level features of a modern operating system. NO APPLICATION SHOULD EVER SCREW WITH IT, IN ANY WAY!
And I say this not just as a seriously annoyed user, but also as a professional UI developer: it is a well-known anti-pattern to override the user's expectations, at any level ... and that applies tenfold to the most basic patterns that every other app follows.
I don't care if you added a magic way to write all my code for me: if the only way to invoke it is to break copy/paste, you've failed at development!
AI is clearly the correct answer to this specific problem: Claude can not identify another writer if they use Claude (or another LLM) enough to massage the text.
The problem, as the article correctly identifies, is that this is just the tip of the iceberg. There are tons of other cases (e.g. identifying a car without a license plate by the scratches) that AI is going to enable ... and in those cases we can't just get anonymity back by using an LLM.
Forget about screen readers: I'm looking at it on a monitor and I just see the robot version!
Feels like an ad for their debugging tool.
So, you know not to blindly trust "experts" ... right? Case in point: this is not an unbiased group of economists!
"Anton Korinek, a University of Virginia professor currently embedded with Anthropic, framed the urgency in historical terms: "Steam, electricity, and computers each gave societies decades to adapt; AI may give us only a few years." ..."
"... Industry representation on the signatory list is notable, with Reuters reporting that it includes Sarah Friar, who serves as OpenAI's finance chief, Jeff Dean of Google $GOOGL -1.31% DeepMind, and Jack Clark, one of Anthropic's founding figures."
In other words, these are people from an industry that has a strong financial interest in selling the idea that their products are world-changing. But, if you blindly accept the word of anyone trying to make a buck off you, without thinking critically ... well, OP you really need to contact me offline about this great bridge I have for sale ...
In practice output styles are just more context (equivalent to CLAUDE.md or memory) ... but with a slightly increased weight.
"even then it was clear that types were a good idea and the implementations were not"
I still feel this way about the state of the art. I want solid type support in my language ... but I do not feel TypeScript is the best possible implementation at all, and for many situations I feel implicitly typed (but still using all of VS Code's type features) Javascript is actually better than using TypeScript.
I eagerly await the day when someone actually gets a JS typing tool right and I prefer using it over raw Javascript.
Serious developers can make a serious argument that you can get quality production-level code with other ways besides explicit enforced type systems (eg. that if you have good enough test coverage such explicit type systems are a redundant waste of time).
Obviously "production quality" varies greatly from shop to shop, but I think there's more legitimacy to the idea than you're giving it.
It's July 2026 ... how does this qualify as news?
IF your audience is 5-year olds, maybe Hacker News isn't the best place to get beta testers?
No paywall bypass link?