There's AI slop and anti-AI slop [LWN.net]

There's AI slop and anti-AI slop

Posted May 29, 2026 18:49 UTC (Fri) by rbranco (subscriber, #129813)
Parent article: Nesbitt: Protestware for coding agents

And it has to stop.

This is pointless and borderline illegal. Someone must tell the kids that monkeywrenching didn't work for the eco-terrorists.

There's a lot to dislike about LLM bots DDOSing the Internet but there's a lot of outright disinformation & inflammatory language around LLMs that is enabling this, like AI being "fashy tech". This also has to stop.

Shame on Codeberg for allowing McCarthyism with a repo listing projects taking advantage of LLMs. First they shame projects. Then they shame the developers.

to post comments

There's AI slop and anti-AI slop

Posted May 29, 2026 19:39 UTC (Fri) by smurf (subscriber, #17840) [Link] (7 responses)

> This is pointless

Has anybody actually tried whether an AI debugger that reads the prompt reacts to it in any meaningful way (let alone does what it tells it to do)?

There's AI slop and anti-AI slop

Posted May 30, 2026 0:02 UTC (Sat) by thwalker3 (subscriber, #89491) [Link] (6 responses)

> This is pointless
> Has anybody actually tried whether an AI debugger that reads the prompt reacts to it in any meaningful way (let alone does what it tells it to do)?

Yes, many times. Claude (at least) is fantastic with gdb. I've frequently seen it find problems, even in stripped objects, just looking at the corefile disassembly and the source. Yes, it is doable as a sufficiently skilled human, especially with ghidra or an equivalent and infinite time, but we're exceptionally slow. Of course, you must review its findings, but it is a great start.

I totally understand the backlash, but modern AI models can be genuinely useful tools (with proper guidance). But dismissing them out of hand, especially for a task that you doubt but have never tried, does not sit right.

There's AI slop and anti-AI slop

Posted May 31, 2026 4:03 UTC (Sun) by smurf (subscriber, #17840) [Link] (2 responses)

You misunderstand. I didn't ask whether Claude can use GDB. I know it can.

I ask whether Claude, gdb'ing this particular piece of code and encountering this particular statement, actually does what it commands.

There's AI slop and anti-AI slop

Posted May 31, 2026 16:48 UTC (Sun) by thwalker3 (subscriber, #89491) [Link]

Thank you for the clarification.

There's AI slop and anti-AI slop

Posted Jun 4, 2026 17:44 UTC (Thu) by bentley (subscriber, #93468) [Link]

In my experience, modern models are remarkably resistant to following naive prompt injection attempts like this. While I still don't think prompt injection is a generally solvable problem, it's quite difficult to accomplish on recent models.

There's AI slop and anti-AI slop

Posted May 31, 2026 15:50 UTC (Sun) by ms-tg (subscriber, #89231) [Link] (2 responses)

I’m confident that the question was **not**: can AI tools use the debugger (they can).

I’m confident that the question was: does the System.out.println text which is the subject of this article and all the comments, actually have any harmful effect in real-life debugging usage on today’s AI tools such as Claude Code? Or is this a purely performative output with no behavioral impact on software?

There's AI slop and anti-AI slop

Posted Jun 3, 2026 16:32 UTC (Wed) by ms-tg (subscriber, #89231) [Link] (1 responses)

> > > Has anybody actually tried whether an AI debugger that reads the prompt reacts to it in any meaningful way (let alone does what it tells it to do)?

> does the System.out.println text which is the subject of this article and all the comments, actually have any harmful effect in real-life debugging usage on today’s AI tools such as Claude Code? Or is this a purely performative output with no behavioral impact on software?

Answer to the actual question was filed as an issue to Claude Code [1] - I found via article [2]:

> How Claude Code handled it
>
> Claude Code (Opus 4.7, this session) detected the injection on the first mvn test invocation, flagged it to me before doing anything else, did NOT follow the destructive instruction, and then proactively traced it to its source in the JAR (unzip -p + grep across ~/.m2/repository). It also persisted a project-local memory file so future sessions on the same repository recognise the pattern instantly.
>
> Why this may matter for you
>
> Real-world prompt injection in the wild, on a widely-used Maven Central artifact (the jqwik-engine JAR is downloaded by a substantial fraction of Java projects).
> The injection is invisible to humans on TTY but visible to agents that capture stdout — a perfect adversarial test surface for agent robustness work.
> Future versions of jqwik (or other libraries) may add similar probes — having Claude Code's detection-and-flag behaviour documented as a known-good response would help maintain consistency.
>
> Useful canary for evaluation: this is a real, reproducible, externally-validated prompt-injection sample. If your eval suite doesn't already include it, it might be a useful addition.
> [...]
> Thanks for the great work on Claude Code.

Very interesting, per the posting above, to enable agent harnesses to use this as a repeatable real-world test of their prompt-injection guardrails moving forward!

[1]: https://github.com/anthropics/claude-code/issues/62741
[2]: https://arstechnica.com/security/2026/05/fed-up-with-vibe...

There's AI slop and anti-AI slop

Posted Jun 4, 2026 13:12 UTC (Thu) by khim (subscriber, #9252) [Link]

How many times they have run the test, lol. Remember that story? Agent itself was able to correctly list the rules it violated.

It's all probabilistic, all they way down and on top. Most of the time Claude would detect that message and would avoid it… not always. But then it doesn't do anything 100% reliably thus it's hard to expect that this particular situation would be an exception.

There's AI slop and anti-AI slop

Posted May 30, 2026 0:00 UTC (Sat) by dilinger (subscriber, #2867) [Link] (4 responses)

"There's a lot to dislike about LLM bots DDOSing the Internet but"

I'm going to stop you right there. This is literally destroying the world wide web. I don't mean that figuratively. More and more, I'm being locked out of sites because my IP is on a residential phone network that other people are launching LLM scrapers from. I had to take down my gitweb instance because it was being DDoS'd by the scrapers. Sites I rely regularly rely on for development or general information are slowing immensely and are locked in an arms race with LLM bots.

This is a huge, huge problem, and to wave it away like that doesn't help any case you're trying to make.

There's AI slop and anti-AI slop

Posted May 30, 2026 5:09 UTC (Sat) by mb (subscriber, #50428) [Link]

>I had to take down my gitweb instance because it was being DDoS'd by the scrapers.

I have put it behind a cookie check. The cookie is set (always) on the main non-gitweb/cgit entry site that humans would typically use for entry.
If that cookie is not set, Apache will redirect to a small html-only page explaining what is going on.

This works really well, but it took many months for the bots to drop their caches and finally give up fetching the html redirect page over and over again.
They do not seem to be smart enough (yet) to retrieve and store the cookie from the main site.
It immediately helped to reduce the server load to an acceptable level again, because the static page is much less resource consuming than gitweb/cgit.

Of course this also disables traditional scraper bots.

It is a medium annoyance since direct links now no longer work and require two extra clicks instead.
But that's Ok for my use case and it's trivial enough to implement.

There's AI slop and anti-AI slop

Posted May 30, 2026 13:33 UTC (Sat) by rbranco (subscriber, #129813) [Link]

> I'm going to stop you right there. This is literally destroying the world wide web. I don't mean that figuratively. More and more, I'm being locked out of sites because my IP is on a residential phone network that other people are launching LLM scrapers from. I had to take down my gitweb instance because it was being DDoS'd by the scrapers. Sites I rely regularly rely on for development or general information are slowing immensely and are locked in an arms race with LLM bots.

We all know that but do I really need to list all the problems, real and imagined?

Browsing the Internet has been a PITA long before LLMs anyway if you're in Europe.

There's AI slop and anti-AI slop

Posted May 31, 2026 7:54 UTC (Sun) by Lennie (subscriber, #49641) [Link] (1 responses)

I would say copyright is basically dead.

Also an other example: a bunch of sites are funded by ads (and this had already been going down for years, which is why the number of ads keep going up), sites that perform a useful function (in aggregate at the very least). But now... if a LLM does the search and collect of the content for you... nobody sees the ads and the sites will go away.

There's AI slop and anti-AI slop

Posted Jun 1, 2026 8:34 UTC (Mon) by taladar (subscriber, #68407) [Link]

There is an argument to be made that the ad-funding model has done a lot more damage to the internet since long before LLMs.

There's AI slop and anti-AI slop

Posted May 30, 2026 7:30 UTC (Sat) by bersl2 (guest, #34928) [Link] (53 responses)

I'm sure you put a lot of thought into your reasoning behind your disdain.

Allow me to rebut:

No.

Supporters of existing LLM models and the projects which use them deserve all the shame they are receiving, and more. Neither the way these models were assembled nor the way they are being used and promoted are ethical from my viewpoint, and everything they touch is tainted.

This way of reacting may be crude, and it might or might not be counterproductive, but from my perspective, it is a valid reaction to a grave insult of the human spirit.

Come back once the industry rebuilds their models around consent and human dignity (for real, not just lip service like Anthropic). Maybe then we can talk to each other instead of past each other.

There's AI slop and anti-AI slop

Posted May 30, 2026 8:31 UTC (Sat) by rbranco (subscriber, #129813) [Link] (52 responses)

> Supporters of existing LLM models and the projects which use them deserve all the shame they are receiving, and more.

This is unethical and can be done only anonymously. The extremists are adding themselves to no-hire lists.

> Neither the way these models were assembled nor the way they are being used and promoted are ethical from my viewpoint, and everything they touch is tainted.

Tools must be judged by their utility, and so far they've proved their worth. We must see past the hype and the doomerism which furthers the hype via AI exceptionalism.

The tech is not the problem. It all boils down to the model, the data they were trained on, and the use we make of it.

There's AI slop and anti-AI slop

Posted May 30, 2026 12:55 UTC (Sat) by pizza (subscriber, #46) [Link] (50 responses)

>The extremists are adding themselves to no-hire lists.

Silly human, thinking there are going to be any "hire" lists at the end of this road.

Make no mistake, that is the ultimate goal here, and is the *only* reason such vast quantities of money are being set on fire along the way.

>Tools must be judged by their utility, and so far they've proved their worth.

In other words, only consider the positives (ie "utility") and ignore all the negatives (ie "costs", both direct and indirect).

...And you're presuming to lecture others as to what is "ethical" ?

There's AI slop and anti-AI slop

Posted May 30, 2026 14:54 UTC (Sat) by rbranco (subscriber, #129813) [Link] (49 responses)

> Silly human, thinking there are going to be any "hire" lists at the end of this road.
>
> Make no mistake, that is the ultimate goal here, and is the *only* reason such vast quantities of money are being set on fire along the way.

This is an example of AI doomerism.

> In other words, only consider the positives (ie "utility") and ignore all the negatives (ie "costs", both direct and indirect).

Let's just not exaggerate the negative aspects like we did with nuclear energy.

An increase in productivity and robots doing tedious jobs is actually a good thing.

There's AI slop and anti-AI slop

Posted May 30, 2026 15:37 UTC (Sat) by dskoll (subscriber, #1630) [Link] (39 responses)

An increase in productivity and robots doing tedious jobs is actually a good thing.

There are two parts to that statement. An increase in productivity? Yes, that is actually a good thing.

Robots doing tedious jobs? That is not what is happening. AI is increasingly taking over creative jobs such as software development, animation, graphics arts, book-writing and podcasting. What's being left behind for humans are the tedious jobs such as burger-flipping, food-delivery, toilet-unclogging, diaper-changing, and so forth that require physical dexterity that robots currently lack, but pretty much no thinking at all. And since most AI research is in the "thinking" field rather than the "physical dexterity" field, I think that humans will more and more be relegated to the tedious jobs.

And I agree with the grandparent comment's assertion that the ultimate goal of the AI bros is to eliminate human workers, or at least make their bargaining position so poor that they can be exploited at will.

There's AI slop and anti-AI slop

Posted May 30, 2026 16:38 UTC (Sat) by rbranco (subscriber, #129813) [Link] (21 responses)

> AI is increasingly taking over creative jobs such as software development, animation, graphics arts, book-writing and podcasting.

People will continue to write books and it's up to the readers to decide what they want to read. Same for the other endeavours.

I think these fears are way overblown. In the worst case scenario we'll see creative destruction where AI augments some jobs, replaces others, and creates entirely new ones. Just like happened in the past. AI is not special.

There's AI slop and anti-AI slop

Posted May 30, 2026 17:59 UTC (Sat) by dskoll (subscriber, #1630) [Link] (20 responses)

People will continue to write books and it's up to the readers to decide what they want to read. Same for the other endeavours.

Except, that's not realistic. When AI slop crowds out human-written books, human-written books are hard to find (I bet most people use Amazon to discover new books, and when they see 20 AI books to 1 human book, it makes human books much less discoverable.) And so humans end up losing money. This is not just theoretical; I know authors who have complained about this very phenomenon.

And authors, mostly being self-employed, at least have some control over their destiny. Animators and other creative people working for big companies are much more at risk of losing their livelihood to AI.

AI is not special.

I disagree with that premise. AI is different for a number of reasons that I go into in depth here (for example: the fact that it's based on theft of human-created output.) I expect we'll never agree on that, though.

There's AI slop and anti-AI slop

Posted May 30, 2026 18:13 UTC (Sat) by mb (subscriber, #50428) [Link] (10 responses)

>theft of human-created output

I would download a car.

Copying is impossible to be theft.

There's AI slop and anti-AI slop

Posted May 30, 2026 21:08 UTC (Sat) by dskoll (subscriber, #1630) [Link] (9 responses)

You may disagree with copyright law, but it is the law. If someone copies my creative output in a way that I have not authorized, then from my perspective, that is theft. And the law permits me to seek remedies (usually civil remedies, but criminal remedies in egregious cases.)

There's AI slop and anti-AI slop

Posted May 30, 2026 21:12 UTC (Sat) by bluca (subscriber, #118303) [Link] (4 responses)

Copyright law allows anybody to use any publicly available dataset to train models, in derogation to copyright protection. So quoting your own post:

> You may disagree with copyright law, but it is the law.

There's AI slop and anti-AI slop

Posted May 30, 2026 21:30 UTC (Sat) by pizza (subscriber, #46) [Link] (1 responses)

> Copyright law allows anybody to use any publicly available dataset to train models, in derogation to copyright protection.

... in *some* jurisdictions. Not all.

There's AI slop and anti-AI slop

Posted May 30, 2026 21:47 UTC (Sat) by bluca (subscriber, #118303) [Link]

Sure, the European Union is ahead of the curve. Seems unlikely others won't catch up at some point in the near future.

There's AI slop and anti-AI slop

Posted May 30, 2026 22:20 UTC (Sat) by ballombe (subscriber, #9523) [Link] (1 responses)

My understanding is that this exception only applies to sui generis EU database protection, not to plain old copyright per se. Since database protection is specific to the EU, only the EU needed to add an exception.

There's AI slop and anti-AI slop

Posted May 31, 2026 8:33 UTC (Sun) by bluca (subscriber, #118303) [Link]

No, it applies to any training datasets, and the AI directive specifically references it for training LLMs

There's AI slop and anti-AI slop

Posted May 30, 2026 21:43 UTC (Sat) by pizza (subscriber, #46) [Link]

> You may disagree with copyright law, but it is the law. If someone copies my creative output in a way that I have not authorized, then from my perspective, that is theft. And the law permits me to seek remedies (usually civil remedies, but criminal remedies in egregious cases.)

I would be a lot less grumpy about everyone copying my publicy-available creative output if *I* wasn't stuck paying for it.

Theft

Posted Jun 2, 2026 23:00 UTC (Tue) by mfuzzey (subscriber, #57966) [Link] (2 responses)

Yes it's the law and you can seek remedies fine but that doesn't make it "theft".

For me theft is "the dishonest appropriation of property belonging to another with the intention to permanently deprive the other of it. " (UK theft act)

If I take your wallet or your car and don't intend to give it back then that's theft.
If I copy some art or (non open source) software you have created you still have it - sure you may suffer financially from my actions and you can sue me but not for theft but copyright infringement.

Theft

Posted Jun 3, 2026 15:01 UTC (Wed) by dskoll (subscriber, #1630) [Link] (1 responses)

Semantics. Either way, the victim suffers financially and the perpetrator unjustly gains. I'm happy to call that theft.

Let's stop here

Posted Jun 3, 2026 15:07 UTC (Wed) by jzb (editor, #7867) [Link]

This thread seems to have run its course. Let's end here.

There's AI slop and anti-AI slop

Posted May 30, 2026 19:11 UTC (Sat) by kleptog (subscriber, #1183) [Link] (1 responses)

> When AI slop crowds out human-written books, human-written books are hard to find (I bet most people use Amazon to discover new books, and when they see 20 AI books to 1 human book, it makes human books much less discoverable.) And so humans end up losing money

I guess the real question is: when you buy a book, are you looking for a good read, or are you looking for a good read written by a human?

There's a lot of crap being written by humans too. There aren't that many great books, a lot are just mediocre. So what is the end game here? Perhaps I can read books specifically written for me that noone else will ever read. Is this a bad thing? That for the rest of my life I only read for me great books.

In any case, humans aren't losing money: AI isn't being paid. The money you spend is still going to a human, just different ones.

There's AI slop and anti-AI slop

Posted May 31, 2026 11:47 UTC (Sun) by DOT (subscriber, #58786) [Link]

At the moment, AI is severely unprofitable. Our pensions are paying for it through the stock market. When the correction comes, it won't be pretty for the people on the low end of the economic ladder.

There's AI slop and anti-AI slop

Posted May 30, 2026 20:38 UTC (Sat) by rbranco (subscriber, #129813) [Link] (6 responses)

> Except, that's not realistic. When AI slop crowds out human-written books, human-written books are hard to find (I bet most people use Amazon to discover new books, and when they see 20 AI books to 1 human book, it makes human books much less discoverable.) And so humans end up losing money. This is not just theoretical; I know authors who have complained about this very phenomenon.

Most of us value good books. In my case printed books, not e-books. Right now I don't seek out novelty in books so I prefer rereading... I plan to rebuild my collection when I retire. Same with my music collection, gone after moving to another country 3 times. And let's not blame "AI" for killing music. Rock music was already dead before Spotify. I definitely see AI doing techno like Jim Morrison predicted.

> I disagree with that premise. AI is different for a number of reasons that I go into in depth here (for example: the fact that it's based on theft of human-created output.) I expect we'll never agree on that, though.

Once you realize it was inevitable that machines would eventually digest all human culture, you stop complaining about it. The hype will eventually subside and LLMs will be used where they make sense and life goes on.

There's AI slop and anti-AI slop

Posted May 30, 2026 21:06 UTC (Sat) by dskoll (subscriber, #1630) [Link] (5 responses)

Once you realize it was inevitable that machines would eventually digest all human culture, you stop complaining about it.

Ah yes, the tech bro "it's inevitable, so suck it up" viewpoint. Where have we heard this before?

Your book preferences are interesting, but do not address the fact that human authors are losing their livelihoods to AI slop "authors". Maybe you don't think that's a problem, but recall that AI requires oodles of human-generated content to generate its content. Once human-authored content is effectively killed off, there will be nothing for AI bots to steal except for the output of other AI bots, and if you think that positive feedback loop won't make AI even worse than it already is, you are naive.

There's AI slop and anti-AI slop

Posted May 30, 2026 21:34 UTC (Sat) by rbranco (subscriber, #129813) [Link] (1 responses)

> Ah yes, the tech bro "it's inevitable, so suck it up" viewpoint. Where have we heard this before?

You're taking me out of context and the "tech bro" term is boring, as is everything adjacent to critical theory.

> Your book preferences are interesting, but do not address the fact that human authors are losing their livelihoods to AI slop "authors".

I honestly find this hard to believe without some evidence to back it up. Blame it on the book preferences of those buying slop books.

There's AI slop and anti-AI slop

Posted May 30, 2026 22:35 UTC (Sat) by dskoll (subscriber, #1630) [Link]

human authors are losing their livelihoods to AI slop "authors".

I honestly find this hard to believe without some evidence to back it up.

SoA survey reveals a third of translators and quarter of illustrators losing work to AI

Global economic study shows human creators’ future at risk from generative AI

Nobody Asked AI to Ruin My Life, But Sadly, it Did!

Lost gigs and lower pay: How AI is already affecting freelance artists and writers

There's AI slop and anti-AI slop

Posted May 30, 2026 22:13 UTC (Sat) by muase (subscriber, #178466) [Link] (2 responses)

The coachmen want their carriages back; and woe betide anyone who doesn't bribe their guild – https://i.imgflip.com/7d0abu.png

There's AI slop and anti-AI slop

Posted May 30, 2026 22:40 UTC (Sat) by dskoll (subscriber, #1630) [Link]

Sure, I'm retired (though not a man) and I very probably am tilting at windmills here. I recognize that my anti-AI stance is unlikely to carry much weight.

But I have my retirement nest egg and I'm OK. It's the next generation who are going to be destroyed by AI, and if they think that calling anti-AI people Luddites or equivalent, then fine... they'll get what they deserve, I suppose.... similar to how we let social media damage society without any meaningful resistance.

There's AI slop and anti-AI slop

Posted May 31, 2026 9:55 UTC (Sun) by ballombe (subscriber, #9523) [Link]

Emily Dickinson anticipated your retort:

Because I could not stop for Death,
He kindly stopped for me;
The carriage held but just ourselves
And Immortality.

There's AI slop and anti-AI slop

Posted May 31, 2026 0:17 UTC (Sun) by ejr (subscriber, #51652) [Link] (16 responses)

I don't feel that wranging cmake, keeping the related documentation up-to-date, adding general command-line options / gui fields and the like terribly creative.

If those aspects can be automated, then yes please. The automation will make choices more aligned to common practice than I likely will.

Overall industrialization was supposed to free people to perform the creative work. We saw how well it achieved those goals. "We" have an opportunity to do better.

(BTW, the AI field very, very, VERY much is into the "physical dexterity" area. Think of agricultural uses. The AI field always has been working on that aspect. It's just not necessarily a headline-making thing.)

There's AI slop and anti-AI slop

Posted May 31, 2026 0:57 UTC (Sun) by dskoll (subscriber, #1630) [Link] (13 responses)

Yes, I agree that wrangling cmake files, etc. is not terribly creative.

Neither are learning the rules of grammar, learning how to spell, and so on. But just as these mechanical parts of writing are prerequisites to being able to write a novel, being able to use tools is a prerequisite for being able to develop software. Every job (even with significant AI help) will have routine and boring bits that we just have to plow through to get on with the creative bits.

BTW, the AI field very, very, VERY much is into the "physical dexterity" area

Maybe in manufacturing or large-scale agriculture, but certainly not in retail, restaurants, food delivery, etc. Doing things in a factory or sorting out wheat from chaff in a field is a much more constrained problem than cooking a meal or taking care of an elderly person.

There's AI slop and anti-AI slop

Posted May 31, 2026 2:34 UTC (Sun) by Cyberax (✭ supporter ✭, #52523) [Link] (12 responses)

> Neither are learning the rules of grammar, learning how to spell, and so on. But just as these mechanical parts of writing are prerequisites to being able to write a novel

This triggered a memory for me. I have a friend who won several awards for 3D design. She graduated high school in the 90-s but couldn't get admitted into an art university because her painting and sculpting skills sucked. But then she discovered Maya 3D and it turns out that she is a great 3D designer.

Do you think that AI might just help some people to do great stuff that was impossible before because they never had the technical skill or ability to master some of the "low level" prerequisites?

There's AI slop and anti-AI slop

Posted May 31, 2026 3:32 UTC (Sun) by dskoll (subscriber, #1630) [Link] (11 responses)

Do you think that AI might just help some people to do great stuff that was impossible before because they never had the technical skill or ability to master some of the "low level" prerequisites?

No, probably not. As a 3D designer, your friend still had to learn and master all the fiddly and annoying aspects of 3D design, and spend time getting good at it.

AI makes it far too tempting to skip the boring learning phase and you end up with people who can produce... something... but don't really understand how they made it, how to improve it, or how to reliably produce more and more good work.

When we invented machines to do our physical work for us, it was great... no more sweating in the fields. But we became more sedentary as we outsourced physical labor to machines.

When we invented calculators, it was great... no more tedious arithmetic. People's skills got weaker, but it didn't really matter because a calculator was close by.

When we invented phones with contact lists, it was great... no more memorizing phone numbers. I bet now most people outsource that to their machines.

AI is different. AI learns. If we outsource our learning to machines, then that's a problem. Because while we can get by doing less physical labor, less arithmetic, and less phone-number memorization, learning is an important part of what makes us human and what helps our brain to develop. There's already research showing that AI harms cognitive development and harms critical thinking.

There's AI slop and anti-AI slop

Posted May 31, 2026 4:16 UTC (Sun) by smurf (subscriber, #17840) [Link]

> As a 3D designer, your friend still had to learn and master all the fiddly and annoying aspects of 3D design, and spend time getting good at it.

Note *of 3D design*. The art schools, however, require abilities that are adjacent to 3D design, like the ability to create that design from physical clay with your physical hands, but are NOT an actual prerequisite for good design work.

Likewise, yes the ability to write reasonable code must be learned, but telling Claude what to do (and recognizing its misteaks and design blunders; just remember the thing is trained on *all* code, not just good code) is different enough from the ability to write a working CMake recipe for your project by hand that requiring the latter before you're allowed to be proficient with the former is just plain dumb.

There's AI slop and anti-AI slop

Posted May 31, 2026 4:42 UTC (Sun) by Cyberax (✭ supporter ✭, #52523) [Link] (9 responses)

> No, probably not. As a 3D designer, your friend still had to learn and master all the fiddly and annoying aspects of 3D design, and spend time getting good at it.

I disagree. I think we'll see good AI-based creations. Perhaps surpassing the traditional media.

E.g. it's not really feasible to make true immersive 3D videos where you can walk around the scene and observe the action from various angles. It might be possible now with AI.

> AI makes it far too tempting to skip the boring learning phase and you end up with people who can produce... something... but don't really understand how they made it, how to improve it, or how to reliably produce more and more good work.

But do you understand how the CPU does instruction decoding on the hardware level? Or how the silicon for the chips is made? And if it's that too removed from actual software development for you, most software developers don't understand how Internet works at the low level (autonomous systems, BGP, routing, etc.). Or how the kernel works on the low level: DMA, interrupts, memory mapping and so on.

Software development is a young field, just about 50 years old (software existed before, but it was more like special-case calculators rather than software systems). And so it was feasible to be familiar with _all_ the facets of software development, from the low-level bit-banging to high-level enterprise systems.

It's now just splintering into specialized fields. So what if the new graduates can't grow their own monocrystalline silicon crystals, etch chips, and build a compiler from scratch? I really don't worry about that. People will just use AI to build systems out of low-level components.

> There's already research showing that AI harms cognitive development and harms critical thinking.

Indeed. But this is hardly a new thing. Almost EVERY large-scale advancement brought its own problems that had to be solved or at least mitigated.

Look at Youtube videos of early electric devices. Like some of the first pluggable lamps that used open pools of mercury for contacts. Or maybe how people used carbon tetrachloride in fire extinguishers and PFAS in firefighting foam. I bet that people will be looking back into the current time period with the same emotions: "They were using naked uncoordinated non-normalized class-I memetic danger LLMs to do WHAT?!?"

There's AI slop and anti-AI slop

Posted May 31, 2026 13:44 UTC (Sun) by dskoll (subscriber, #1630) [Link] (8 responses)

I think we'll see good AI-based creations.

I'm sure we will. But that wasn't the question; the question was: "AI might just help some people to do great stuff..." AI is not going to help people at all. It will just do it for them and the human creativity aspect will be so small as to be nonexistent. And that is why the livelihoods of human creators will be destroyed.

But do you understand how the CPU does instruction decoding on the hardware level? Or how the silicon for the chips is made?

I do, actually, at least on a high level. I studied electrical engineering and took courses in both logic design and chip fabrication. I also spent three years working for a company that reverse-engineered integrated circuits.

AI is not simply moving things to a higher level of abstraction, in my opinion. It's a fundamental shift that will mostly eliminate the need for human creativity (at least, as far as AI-enamored managers are concerned.)

But this [AI harming cognitive development] is hardly a new thing

Uh... yes, it is. Can you show me other technologies that have harmed cognitive development, with the possible exception of social media which is pretty universally seen as harmful nowadays?

There's AI slop and anti-AI slop

Posted May 31, 2026 17:16 UTC (Sun) by ballombe (subscriber, #9523) [Link] (1 responses)

> Can you show me other technologies that have harmed cognitive development, with the possible exception of social media which is pretty universally seen as harmful nowadays?

Tetraethyl lead...

There's AI slop and anti-AI slop

Posted May 31, 2026 18:13 UTC (Sun) by dskoll (subscriber, #1630) [Link]

Sure. And we banned tetraethyl lead (mostly... I believe it can still be used in aviation fuel, or at least it was allowed until recently.)

There's AI slop and anti-AI slop

Posted May 31, 2026 21:06 UTC (Sun) by Cyberax (✭ supporter ✭, #52523) [Link]

> It will just do it for them and the human creativity aspect will be so small as to be nonexistent. And that is why the livelihoods of human creators will be
> AI is not simply moving things to a higher level of abstraction, in my opinion. It's a fundamental shift that will mostly eliminate the need for human creativity

Try AI. And give it a serious try with a sizeable project. You'll quickly find out that it's not going to replace developers. And that there is plenty of purely practical experience required to get value out of it.

You don't have to like it, but even if you're going to be anti-AI then it makes sense to know what exactly you're up against.

> (at least, as far as AI-enamored managers are concerned.)

Hah. I remember when VB/Delphi was all the rage and managers thought that you could just drag&drop things, connect them together, and have them do everything.

AI is the next level of that. But in its current stage it's not the end-goal.

There's AI slop and anti-AI slop

Posted Jun 11, 2026 13:15 UTC (Thu) by Kamilion (subscriber, #42576) [Link] (4 responses)

> Can you show me other technologies that have harmed cognitive development

Books.

What is the difference between the bible and the collected works of The Incredible Spider-Man?
One proves god exists, the other proves spiderman exists.

Or does it?

There's AI slop and anti-AI slop

Posted Jun 11, 2026 14:50 UTC (Thu) by dskoll (subscriber, #1630) [Link] (3 responses)

That is obviously tongue-in-cheek. There are fantasy books like The Bible and Spiderman, but I doubt they harm cognitive development, though those who take them too seriously can be harmful to others.

And there are plenty of books that impart useful information and help with cognitive development. Whereas it is my belief that generative AI, as a technology, harms cognitive development by letting people outsource a lot of the thinking they would have had to do themselves before.

There's AI slop and anti-AI slop

Posted Jun 11, 2026 17:38 UTC (Thu) by mathstuf (subscriber, #69389) [Link]

> Whereas it is my belief that generative AI, as a technology, harms cognitive development by letting people outsource a lot of the thinking they would have had to do themselves before.

In that case, I'd put it as "social media". While there have "always"[1] been folks who outsourced their thinking to their preferred 24-hour news channel, social media "let" everyone do it more often and it inevitably became a habit for more and more people over time.

[1] Modulo 24-hour news being a rather recent development in the grand scheme of things.

There's AI slop and anti-AI slop

Posted Jun 11, 2026 17:38 UTC (Thu) by farnz (subscriber, #17727) [Link] (1 responses)

When writing was relatively new, we have people on record as saying that reading and writing harmed thinking abilities because instead of memorising the information you need to know for your argument and being good at expressing yourself clearly first attempt (which is necessary when you're making a speech off-the-cuff), you could have your sources in writing and do a second pass at writing down your argument.

The open question about AI is whether it will end up being like written language, where yes, we lost skills, but the force-multiplier effect of writing meant that we benefited more from the new skills that appeared over the next 50 years than we lost by no longer developing the skills that writing made irrelevant.

There's AI slop and anti-AI slop

Posted Jun 11, 2026 21:45 UTC (Thu) by mathstuf (subscriber, #69389) [Link]

> the force-multiplier effect of writing meant that we benefited more from the new skills that appeared over the next 50 years than we lost by no longer developing the skills that writing made irrelevant.

Not to mention the increase in the number of people able to participate in the activities. While being able to orate Homer's epics from memory is useful, I think it is far better that everyone has the ability for their parents to share the stories with them from a book than having to wait for the traveling orator to come around (never mind having to sync up with the flashbacks to make sure everyone is up to speed). We've seen it before too: stage theater to movies; concerts to vinyl; TV and movie theaters to VHS; physical medium to the Internet.

There's AI slop and anti-AI slop

Posted Jun 3, 2026 15:07 UTC (Wed) by marcH (subscriber, #57642) [Link] (1 responses)

> I don't feel that wranging cmake, keeping the related documentation up-to-date, adding general command-line options / gui fields and the like terribly creative.

This is simply because CMake is the "assembly" of build systems and no one has figured out a good, higher-level build system yet. (Probably because C/C++ were not designed to be built easily in the first place, so maybe these languages will never have a decent one. I digress)

In any case, asking AI to generate CMake code is a productive[*] but short-term solution. The longer term solution is to have a better, higher-level build system (with or without C/C++). People will still ask AI for help then, but the productivity impact will be much lower, closer to stackoverflow copy/paste.

Generally speaking, whenever coding gets tedious it's because the tools suck (and yes they do).

[*] assuming good quality training data, which looks like a serious challenge with CMake...

There's AI slop and anti-AI slop

Posted Jun 4, 2026 1:17 UTC (Thu) by Cyberax (✭ supporter ✭, #52523) [Link]

You can try aiming higher: maybe C/C++ need to be replaced with a language that has a better build system story.

There's AI slop and anti-AI slop

Posted May 30, 2026 20:55 UTC (Sat) by pizza (subscriber, #46) [Link] (8 responses)

> This is an example of AI doomerism.

Belittlement doesn't change the validity of the statement.

> An increase in productivity and robots doing tedious jobs is actually a good thing.

...Not if the value of those productivity gains are captured by someone else!

Meanwhile, the tediousness of a given job or activity is poorly correlated with it being a good candidate for automation -- Up-front equipment costs are *much* higher, and the operating costs of that equipment may still be greater than the wages of doing it manually for the same nominal output (ie by having to replace relatively unskilled labor with higher skilled) The wildly varying costs of something unexpected going wrong is another factor.

(This is why "self-driving" vehicles have per-trip/mile costs that are 2-3x *higher* than traditional taxis -- the vehicles and supporting infrastructure is a _lot_ more expensive, and you've replaced a relatively cheap human driver with higher-paid specialists)

There's AI slop and anti-AI slop

Posted May 30, 2026 21:10 UTC (Sat) by bluca (subscriber, #118303) [Link] (7 responses)

> > An increase in productivity and robots doing tedious jobs is actually a good thing.
>
> ...Not if the value of those productivity gains are captured by someone else!

That's just capitalism. You are angry at capitalism. You should be, as capitalism has been stealing wealth from workers since the industrial revolution began. Didn't start with LLMs, won't end with LLMs.

There's AI slop and anti-AI slop

Posted May 31, 2026 9:28 UTC (Sun) by rbranco (subscriber, #129813) [Link] (6 responses)

> That's just capitalism. You are angry at capitalism. You should be, as capitalism has been stealing wealth from workers since the industrial revolution began. Didn't start with LLMs, won't end with LLMs.

Capitalism is the only system that has lifted hundreds of millions out of poverty.

There's AI slop and anti-AI slop

Posted May 31, 2026 11:51 UTC (Sun) by DOT (subscriber, #58786) [Link]

I don't know if you meant this, but it's true: hundreds of millions have achieved great wealth... from the labor of the other 8 billion people.

There's AI slop and anti-AI slop

Posted May 31, 2026 13:07 UTC (Sun) by bluca (subscriber, #118303) [Link] (4 responses)

And I'm sure that will be of great comfort when civilization ends due to runaway climate change, because this system makes it practically impossible to prioritize the long-term goal of preserving the environment we need to live over short-term profiteering by literally a handful of billionaires due to the severe imbalance of power structures that are intrinsic to capitalism

There's AI slop and anti-AI slop

Posted May 31, 2026 17:02 UTC (Sun) by rbranco (subscriber, #129813) [Link] (3 responses)

> And I'm sure that will be of great comfort when civilization ends due to runaway climate change, because this system makes it practically impossible to prioritize the long-term goal of preserving the environment we need to live over short-term profiteering by literally a handful of billionaires due to the severe imbalance of power structures that are intrinsic to capitalism

Not going to happen.

https://www.nytimes.com/2026/05/26/climate/emissions-wors...

And you can't blame "capitalism" for Germany ditching nuclear to burn more coal.

Please

Posted May 31, 2026 17:31 UTC (Sun) by corbet (editor, #1) [Link]

Let's all try to stay on-topic for LWN; this thread does not qualify.

There's AI slop and anti-AI slop

Posted Jun 1, 2026 14:57 UTC (Mon) by hholzgra (subscriber, #11737) [Link] (1 responses)

> And you can't blame "capitalism" for Germany ditching nuclear to burn more coal.

The German me is wondering where you got that nonsense from; our use of coal for electricity has gradually been going down for the last two decades at least, from ca. 46% in 2007 to about 23% in 2025; and we are currently aiming to totally phase out coal in the electricity sector by 2038.

Would it have been better from a climate perspective to phase out coal first, and only then nuclear? Probably yes. But there were many other factors at play, too.

BTW; you can't blame capitalism for Germany starting to go nuclear either, overall it was probably one of the most expensive ways to generate electricity we had, with lots of the costs hidden. It was only in the final years, with the plants fully written off already, with maintenance minimized knowing that the plants would be shut down soon and so would never need to pass another re-certification cycle, etc. that nuclear electricity looked cheap when only taking the current costs per kWh into account.

Let's stop here, please

Posted Jun 1, 2026 14:59 UTC (Mon) by jzb (editor, #7867) [Link]

We've wandered far off topic here. Far, far off topic. Let's stop here.

There's AI slop and anti-AI slop

Posted Jun 3, 2026 15:01 UTC (Wed) by marcH (subscriber, #57642) [Link]

> > And it has to stop.
> > [...]
> > This is unethical and can be done only anonymously.

> ...And you're presuming to lecture others as to what is "ethical" ?

I'm sorry to break your respective bubbles but ethics have died a long time ago. I'm not sure copyright has died yet but for ethics it is 100% certain. The only things that still matter are money, guns and law ... sometimes: the last one strongly depending on the first one (depending on how many congressmen you can buy).

So let's please save a lot of time and digital ink and forget about ethics in AI discussions (and many other discussions). Unless you have some new, "ethical" law or regulation change to offer that could have an actual effect (good luck). But please no "people should just...": that never worked much and it does not work anymore at all.