calmoo

Ironically the author isn't immune from publishing slop themselves. https://john.hartnup.uk/backpressure/

Gary is not a voice worth listening to. A narcissist, fraud and has terrible epistemics.

That comment doesn’t read like slop in the slightest to me, so the people DMing you have a bad eye for it. Regardless, I think your ‘human’ sample is not a good indicator of the quality of pangram. I would update your priors a bit.

I’ll be honest, i’ve never read text that is human written that reads as close to an LLM as your sample sounds.

I think discounting Pangram’s accuracy based on that sample isn’t very reasonable. Really nobody writes like that other than LLMs!

I'll take your word that it's human written, but reading that gist, it absolutely reads like Claudeslop / GPT slop, it doesn't surprise me in the slightest that Pangram flagged it as AI when it reads identically to LLM output - this feels like a very acceptable edge case to me (assuming you are telling the truth).

Are you absolutely sure you wrote this by hand? If so it's kind of remarkable how close to an LLM you write like.

Please share examples

I’ve never actually seen someone give an example of what they say pangram get wrong, so please go ahead and share.

Pangram is very accurate

Will this likely fix stability issues in the Bun Workers API? https://bun.com/docs/runtime/workers

Interesting - now that I think about it, feeding Claude my particular Xcode theme from VS code to replicate in Zed, the color tones were quite far off how it looked in cursor. I don't know enough about color rendering to know why that is, but I feel you!

I would usually agree, but I think we should be obligated to call out slop when we see it.

You could solve this fairly easily by giving an LLM the json definition of your sublime text theme and just transforming it onto a zed spec theme? No need for screenshots.

I think this article (and the author's previous articles on their blog) is quite clearly AI written. It has such a frustratingly punctuated cadence and really does not serve the reader anything valuable.

Is GC useful for a static type checker? Or did they make a new runtime?

Specifically with YAML, Go templating requires some special operators to get anything useful done and it ends up being an unreadable mess.

I've recently been dipping my toes into k8s / kustomize / helm, and I recently had a situation where having a base deployment yml template that I wanted to reuse across various deployments. I had a look at Helm and I was frankly shocked how bad the templating was with Go templates, it was close to unreadable and felt very brittle!

It’s fast, looks nice and since i really just review agent output these days, that’s good enough. They don’t move everything around and it moves at a nice pace.

On the contrary, anecdotally, myself and every engineer I know have switched fully from cursor to claude code since the start of the year. I now use zed with cc. I personally could not stand the buggy mess and constant UI changes of cursor. It’s also not good value in terms of claude tokens compared to claude code.

HN already does a decent amount of content-policing, which helps keep the discussion higher quality. I don't see a huge diversion here from the usual moderation.

I think a more generous interpretation of dang's comment is that it's fine to use LLMs / tools to fix grammatical errors / spellchecking, but a heavier pass where the prose, wording and tone is altered (even mildly) can create a 'slop ambience' over time, death by a thousand paper cuts.

If you're referring to speaking in English - in general I think there is a huge amount of flexibility for making mistakes in English. I'm a native speaker, I am so used to hearing various levels of English from different nationalities that i'm almost blind to it. I much prefer to hear someones true voice even if there are a few inaccuracies, so much of a person's personality is conveyed through their quirks and mistakes.

I literally got a taxi today in Berlin because the trains were on strike, and the other day because the trains and trams were broken due to ice. You really can’t think of valid uses for a taxi?

For one, it has dynamic memory allocation and is much faster and resource efficient. Drop in replacement. It also has a rather nice UI,

no, definitely not. It was 100% LLM written. Look at their post history.

Your original comment is completely void of any substance or originality. Please don't fill the web with robot slop and use your own voice. We both know what you're doing here.

Ironically your comment is clearly written by an LLM.

Your comment is AI generated

It doesn't matter what the law is, it's setup to allow these malicious practices. It's bad legislation.

Perfectly explained, thanks!

How much fraud we talking? I like to think along Matt Levine's ideas that at least some amount of fraud is acceptable - there will always be some amount of it and under or over-regulation creates a net-negative.