What? No. Software is something people go looking for and choose to download, unless we’re talking about malware which I think is fair to say is obviously outside the bounds of this conversation. Spam emails are forced on people without their asking or looking for them.

Yeah, and that’s totally a criterion people use when labelling something “slop” or not, right? Right??? Oh wait, no, it isn’t.

They’re not at all interchangeable or the same thing.

That is not even remotely close to what I said. Not bothering further with a liar (or worse) who distorts what others say.

He did. You may have misunderstood.

I mean he should have simply stated “the Claude code stays”, instead of hiding which were the Claude commits.

Yeah, nah. Damage has been done already.

He should’ve stuck with his guns and said “the Claude code stays”.

When it comes to the usage of both words, that difference you listed is completely arbitrary and obviously irrelevant. People also use the word “slop” to refer to commercial software (see: “Microslop”) and “spam” to refer to any sort of undesirable email being mass sent, even if non-commercial.

Unless you’re trying to argue something else; that the slop in this specific case is more justified. Then refer to the top comment in the chain; frankly the main issue here is not adding slop to their software, it’s the eagerness to treat users as braindead trash undeserving transparency.

The way y’all overuse the word “slop” is like calling all e-mail “spam.”

It’s more like calling automatically sent e-mails “spam”. From the PoV of the [software | e-mail] user saying the word, both [slop | spam] are undesirable, even if the [coder | marketing team] in question is doing it on purpose and with purpose, to further their goals of [pumping out more software | reaching a wider audience].

If any interaction with spicy autocomplete is treated as equally bad, to the point of aggressive mockery - no kidding people will tune that out.

For me at least the worst part isn’t using it, but trying to hide it. I don’t think it’s justified, even if some users return snide comments because of it.

Even if it was the same conversation (as TrickDacy said, it isn’t), “you can’t avoid the harm completely” is no excuse to avoid causing at least some of that harm.

What decides if something is slop or not is the thing itself. It is not your “KwaLiFiKaShunz”. Bringing up “muh 30 years of XP lol lmao” means jack shit.

If he was co-authoring the code with Claude this means he submitted code made by Claude; he didn’t just ask for some examples and implement in his own way. The later would be far more reasonable than the former.

What he said about the problem being capitalism instead of the tool itself is, I believe, valid. However, it should be no excuse to unnecessarily feed that very same economic system, by paying for the bloody tool.

Finally. He could’ve fixed what people complained about, by removing the commits, so he would keep them happy. He could also stick to his guns, and say “no, I’m not changing it. The Claude code stays”. But he did neither; instead he’s hiding it from the users. That’s pretty much the same as saying “I’m going to treat users as gullible filth and easy to fool, instead of human beings deserving honesty.”

A good thing open software can be forked.

Some of the guidelines say sensible-sounding things. This is one of them.

However, this doesn’t help when the community goes out of its way to shit on the spirit of those guidelines, while still abiding to their letter. And that’s exactly what I’m predicting for this one: they’ll pick only the “HN is for conversation between humans.” to follow, argue that since “ackshyually errything is AI lmao” then it’s fine because the guideline is impossible to follow, or find some grey area to further normalise AI conversations.

Just like they shit on everything else there.

Some off-topic, on the guidelines as a whole:

• In most situations, like this one, rules are better than guidelines. With rules it’s clearer for both non-mod users and moderators what they should do.
• “Please” is fine when interacting directly with users, as there’s already some hierarchy pressure from the mod talking with you. But they don’t belong here; here they make the guidelines sound like requests. And requests can be declined, unlike policies.
• If you’re listing too many items as rules/guidelines, odds are most of them are either related to each other or fluff. That’s the case here; good cleanup is imperative.

Be kind. Don’t be snarky. Converse curiously; don’t cross-examine. Edit out swipes.

“Be kind” is, most of the time, a bad rule/guideline. Not just because everyone has a different interpretation on what is supposed to be “kindness”; but because it gives sea lions a nice cover.

See, a lot of the sealioning strategy is to annoy the shit out of you, feigning to not understand what you say, while being superficially kind. If you shut up, the sea lion wins, so often disengaging is not an option; if you snap out and go overboard, the sea lion wins too; and if you sealion back, you’re legitimating their approach.

So a good strategy against sea lions is to be snarky and slightly impolite, while highlighting what they are doing. That works because it forces the sea lion to either stop sealioning or out themself as such. But… well, that contradicts a blanket “be kind lol”. And predictably, HN is a full raft. [inb4 yes I had to websearch the collective for “sea lions”. :-) ]

Please don’t post comments saying that HN is turning into Reddit. It’s a semi-noob illusion, as old as the hills. [insert bunch of links]

Because yeah, linking people making the same claim in the past totally rebukes the claim. Ppffffft.

May I be frank? I consider Reddit a shithole infested with braindead trash. But in 2026, HN is already worse than a lot of subreddits. Some submissions are still fun (that’s why I’m here!), but the comments section is a cesspool.

I mentioned this in another thread, but I never ditched RSS. It’s simply too convenient to keep track of multiple sites, that you want to check regularly but would be a bother to do manually. Youtube channels, anime/manga/comics sites, even the communities you moderate.

I’m glad to see people picking it up again, though. Perhaps then Liferea will see a bit more love.

Grammarly was already trash before the AI craze.

Liferea is still going strong for me. Sure, it doesn’t start minimised in the system tray for years; but it works. I use it for stuff like

• seeing if someone posted new comments in the comms I moderate here in the Fediverse ( !linguistics@mander.xyz and !linguistics_humor@sh.itjust.works ). So if shit happens I can act quickly.
• youtube videos. Yup, YT still has RSS. I don’t know for how long, but once it removes that ability I’ll probably circumvent it.
• some webcomics I follow, like The Order of the Stick and Sarah’s Scribbles.

I also used it to keep track of manga and anime in the past, but nowadays Suwayomi and Hayase covered that role for me.

I guess you could also use it for news? …meh, frankly I retrieve most of my international news from Lemmy, and local news from a few sites I visit ~once a day.

[Off-topic] I got curious about the comment chain, checked it in a private window, and… well, I don’t remember when I blocked that poster, but by their profile I’m glad I did it — it’s a waste of time to chat with assumptive fools, you spend more time brushing off their assumptions (only so they vomit yet another assumption, and another, and another…) than actually saying what you want, or reading something meaningful. You probably won’t miss them.

[On-topic] I got the same experience as in your second link, but with translation instead of programming — using machine translation to give me ideas on how to translate specially problematic excerpts; idiomatic expressions, tricky grammatical distinctions lacking in the target language, stuff like this. Just ideas, mind you; I wouldn’t copy the machine translation, I’d pick one or two words from it and come up with my own, so it was still human-made.

Then I noticed the “problematic excerpts” were becoming more and more common.

Some might argue “than mite as well not uze calculatorz lol lmao u’ll get rusty math”… you know what, it’s actually a fair comparison, and one of the reasons I do think people should do maths by hand sometimes. Tools are supposed to allow you to do more, not to cripple you until you’re doing less.

We are tools assisting them. I don’t want to spend my life as an “LLM output checker”.

It’s possible you read this text already, but if you didn’t, Cory Doctorow wrote a great piece about this. Some good excerpts of it that fit really well what you said:

Start with what a reverse centaur is. In automation theory, a “centaur” is a person who is assisted by a machine. You’re a human head being carried around on a tireless robot body. Driving a car makes you a centaur, and so does using autocomplete.

And obviously, a reverse centaur is machine head on a human body, a person who is serving as a squishy meat appendage for an uncaring machine.

Obviously, it’s nice to be a centaur, and it’s horrible to be a reverse centaur. There are lots of AI tools that are potentially very centaur-like, but my thesis is that these tools are created and funded for the express purpose of creating reverse-centaurs, which is something none of us want to be.

The AI can’t do your job, but an AI salesman can convince your boss to fire you and replace you with an AI that can’t do your job.

The results revealed a troubling paradox. Workers who were more susceptible to corporate BS rated their supervisors as more charismatic and “visionary,” but also displayed lower scores on a portion of the study that tested analytic thinking, cognitive reflection and fluid intelligence. Those more receptive to corporate BS also scored significantly worse on a test of effective workplace decision-making.

This is only a paradox under the assumption that gullible people are smarter. Because, yes, you need to be at least a bit gullible to see “charisma” in the others, or to not acknowledge everyone and their dog has a “vision”.

The study found that being more receptive to corporate bullshit was also positively linked to job satisfaction and feeling inspired by company mission statements.

“Chrust me, you’re happy!” “Yay, I’m happy!”

This applies also outside working environments, I think. It’s more of a general thing, on how bullshit spreads and gets enforced over sanity. I think the vicious cycle the text points out should appear elsewhere too.

Perhaps some pressure towards critical thinking might counter it?

A lot of the text is good advice for any project, not just programming. Whatever you’re working on, if it’s meaningful, should have a simple and definite scope, and clear priorities. Even it’s something like oil painting, pepper breeding, or a cardboard war tank for your cat.

A few additional tidbits. Not contradicting the text itself, but things people often get wrong about this sort of advice.

Constraints are advantages

Only to the point they force you to prioritise. You can’t really give someone raw dough and say “we were making bread under a time constrain”.

Ignore feature requests — don’t build what users ask for; understand the underlying problem instead

This does NOT mean “be an assumptive piece of shit”. You do not know what the user “wants” or “needs”, nor you should lie you do. It means instead you should look at what your project does versus what it should be doing, see if they mismatch, and address that mismatch.

Ship early, ship often — a half-product that’s real beats a perfect product that’s vaporware

This does not mean “user time is worthless trash, might as well use those things as unpaid beta testers”. Or “it’s fine to release broken shit”. It means instead “be reasonable with your expectations of perfection, and take diminishing returns into account”.

Sensible advice (don’t use this sort of grey), bullshit reason and facepalm-worthy analogy:

I actually believe increasing contrast for everyone improves the information density of our content. It literally becomes higher fidelity. It’s like taking a WAV file, converting to a 1kbps MP3, and then re-converting to a WAV file. You just footgunned yourself my dude! You should not do that.

No.

For text the information is encoded in the characters, that are abstract units. A “t” written in black conveys the exact same information as a “t” written in grey on white. What matters is if you’re using a “t” instead of a “d” or a “τ” or whatever. As such, information density won’t be affected by your questionable colour choices. Nor fidelity, because the information itself isn’t changing.

For audio things are different. Audio doesn’t work through those abstract units, you care about the sound wave; and that sound wave will get distorted once you convert the WAV into the 1kbps MP3.

The real reason to not use this sort of grey is that it’ll always give you a low contrast, no matter what you pair it with. And both excessively high and excessively low contrasts are harder to read and will tire the readers’ eyes down, doubly so for the ones with poor eyesight.

Yes. And, more important than that: EV batteries don’t just take energy to run, they take energy to manufacture. Usually you wouldn’t count the energy taken to produce a gas tank because it’s long-lasting, but if your battery lasts ~10y this amount of energy might be quite relevant — and probably is relevant due to the price.

(This shows the “fuck cars” community is spot on, when it comes to EVs: they don’t solve the problem of the environmental impact behind cars, at most they alleviate it. An actual solution would be to design cities so people don’t need to use cars willy-nilly.)

Yeah, the site coding was awful. Kind of expected from AI junk. However it seems to me like the data itself was actually researched, not “vibed”.

Edit: I see OP tweaked their AI prompt and delivered a watt-hour rating for petrol, which they must have pulled out of their ass.

I redid the maths:

• petrol heat of combustion should be 45~50MJ/kg, based on the the typical values for alkanes.
• 1 MJ = 277 Wh
• specific density of gasoline: 0.715~0.780 kg/L
• fuel consumption seems to be 5~11 L/100km = 0.05~0.11 L / km
• 1.61 km = 1 mile

Plugging all this stuff together, you get (45~50 MJ/kg) * (0.715~0.780 kg/L) * (277 Wh/MJ) * (0.05~0.11 L/km) * (1.61 km / mile) = 717~1913 Wh/mile. The estimate in the site is 1000 Wh/mile; it sounds reasonable.

Let’s roll for a moment with the worst hypothesis, and say that: his account of the facts is all bullshit, he has been submitting slop for a long time, and he only came up with his “I was sick” because this time he got caught on it.

In that hypothesis, Ars Technica has been covering his arse for a long time. Either due to excessive leniency (they caught him other times, and did nothing about it) or sheer incompetence (lack of some internal review process, either pre- or post-publication). Basically giving him the OK sign to keep doing it. Then the timing of firing him shows AT’s issue wasn’t Edwards submitting slop, but giving in to public outrage.

In either case the primary blame goes to Ars Technica, not to the individual worker. The only situation I can say they would be handling this right is if this was his second time doing this shit, in that time he was privately scolded and warned (“this shit is not tolerable, do it again and you’re out”), and still went for it. I find it unlikely.

And from our (both of us) PoV there’s absolutely no info to know if he did this shit more times. In that situation I don’t think we should assume he did.