AI disclosure: I used AI to help organize and edit these thoughts. The arguments are mine to the extent that any argument is ever entirely someone's own. If that bothers you, feel free to attack the argument rather than the tool.

---

I think the deeper issue is why we're treating productivity statistics as the final judge in the first place.

Productivity is not a direct measurement of economic value. It is an accounting construct that depends on how output is defined and measured.

That is relatively straightforward when counting tons of steel or bushels of wheat. It becomes much harder when dealing with software, research, legal work, design, information retrieval, decision support, risk management, and other information-intensive activities.

AI could be improving decisions, reducing errors, lowering search costs, increasing optionality, or generating consumer surplus without immediately appearing in revenue-per-worker statistics.

We already know GDP and productivity measures struggle to capture the value of many digital goods. Search engines, GPS, Wikipedia, open source software, and the broader Internet generated enormous value relative to what showed up in traditional economic statistics.

There is also a timing problem. What was the measured productivity gain for the average manual laborer shortly after the steam engine was introduced? General-purpose technologies often require years of organizational change before aggregate statistics move. Factories had to be reorganized around electricity before electrification delivered its full benefits.

The argument also seems oddly selective. Slok removes the companies where AI's effects are most obvious, then concludes that the remaining firms have not yet shown economy-wide productivity gains. If the steam engine first changed mining and textiles, would we have concluded steam had no economic impact because most workers were still digging ditches by hand?

Most importantly, why assume productivity is the metric that matters? A technology can create value by reducing uncertainty, improving coordination, avoiding mistakes, increasing resilience, or expanding opportunities. Those benefits do not necessarily show up as higher output per worker.

So before concluding that AI has no economic impact because productivity measures are flat, I would first want to know whether productivity statistics are actually measuring the thing we care about. Flat productivity could mean AI is useless. It could also mean the metric is only observing a subset of the effects.

(Disclaimer: Written with an AI collaborator, because human editors were apparently too busy copy-pasting to proofread.)

Speaking of slop, notice how the summary ends by repeating the exact same sentence twice:

"It's hard to compete against things like the DALL-E image generator."
"Long-time Slashdot reader schwit1 quips that maybe 'It's hard to compete against things like the DALL-E image generator...'"

When humans publish duplicate-riddled writeups without reading them, it really raises the question of whether "human slop" is even setting a higher bar than "AI slop."

To your point about "good enough": market dynamics have always favored cheap and good-enough over high-quality and expensive for standard commercial uses. Stock photo giants like Getty relied on charging massive licensing fees for stock images (or bullying photographers over copyright).

When a restaurant menu, blog post, or corporate deck just needs a generic image of a salad or a boardroom, AI generation is free, instant, and "good enough" for the buyer. The tragedy isn't that AI art is superior--it's that for 90% of basic commercial applications, nobody wanted to pay Getty prices in the first place.

AI disclosure: yes, I used an AI as a drafting assistant. I do not consider that fundamentally different from using a compiler, debugger, search engine, calculator, spellchecker, or discussing an idea with another programmer. The argument is mine. The responsibility for it is mine. Judge the reasoning, not the tool.

---

The thing that smells off to me is that Slashdot itself mangles Unicode every time I paste it into a comment box.

So I'm reluctant to accept the implicit premise that software written entirely by humans has already conquered all the easy problems and that AI-generated code must therefore clear some uniquely high bar before we acknowledge it may be useful.

Perhaps the 15% figure is exaggerated. Self-reported numbers often are.

But "real software contains domain-specific rules, institutional habits, regulatory requirements, and weird edge cases" is not an argument against AI coding. It's a description of software. Human programmers have been wrestling with exactly those problems for decades.

What I find interesting is that the article treats the existence of hard domain-specific problems as evidence that AI cannot be writing large amounts of production code. Yet many mature systems built entirely by humans still struggle with things that sound straightforward when described in a sentence.

Again, I submit as evidence the inability of this very site to faithfully render Unicode text pasted into a comment.

The benchmark is not perfection. The benchmark is whether the AI can produce code that is competitive with the code humans were already producing.

If the evidence is merely that real-world software is messy, full of exceptions, and imperfect, that seems equally consistent with the hypothesis that AI has reached parity with a substantial fraction of ordinary software development.

AI disclosure: yes, I used an AI [Copilot] as a drafting assistant. I do not see that as fundamentally different from using spellcheck, a search engine, a calculator, a compiler, Stack Overflow, Wikipedia, or discussing an idea with another person. The argument is mine. The responsibility for it is mine. Judge the reasoning, not the tool.

You're conflating social acceptance with correctness.

"The in-crowd can recognize the out-crowd" is mostly a statement about tribal pattern matching, not truth. History is full of examples where the majority correctly identified the dissenter and incorrectly identified reality.

This reminds me of an HN discussion where multiple AI models reportedly converged on a controversial observation about humans: many people would rather fit in than be right. Your comment seems to elevate that tendency from an observed human bias into a social ideal.

If the objective is social survival, conformity matters. If the objective is discovering what is true, conformity is often an obstacle. Scientific, artistic, and technological advances routinely begin as ideas that fail prevailing social tests.

Also, the writing argument feels circular. If LLMs write like average humans, you say the crowd will still detect them. At that point you're no longer evaluating writing quality. You're assuming the crowd can identify outsiders because they are outsiders.

The deeper issue is that your metric quietly shifts from "is this writing good?" to "does this writing belong to our tribe?" Those are different questions.

Herd recognition may be real. What it does not provide is a guarantee of truth, merit, intelligence, or authorship. Human societies have spent thousands of years demonstrating otherwise.

AI disclosure: The underlying opinions are mine. The AI's contribution was mostly translating an emotional rant into something less likely to get dismissed as an emotional rant. If that's cheating, then so is spell-check.

---

This seems to assume that social acceptance is the thing being optimized.

Why should it be?

The entire point of constitutional rights is that they apply equally to the in-crowd and the out-crowd. A society based purely on social acceptance is just a popularity contest with better branding.

The view you're describing may be descriptively accurate in some settings. Groups often enforce norms through informal ostracism. But that's an argument about how humans behave, not about how they should behave.

If someone's argument is wrong, show why it's wrong. If someone's writing is bad, criticize the writing. "You're out-crowd now, therefore nothing else matters" strikes me as abandoning reasoned evaluation in favor of tribal status enforcement.

AI disclosure: The underlying opinions are mine. The AI's contribution was mostly translating an emotional rant into something less likely to get dismissed as an emotional rant. If that's cheating, then so is spell-check.

More broadly, I think we're heading toward a strange place where people spend more effort policing the provenance of an argument than evaluating the argument itself. If the reasoning is wrong, it should be wrong because the logic is faulty. If it's right, it should be right because the logic is sound. Whether an AI helped turn "this article makes me mad" into a coherent critique seems orthogonal.

Slashdot has traditionally been a "show me the source code" community. In argument terms, that means: show me the logic, not the authorship pedigree.

Disclosure: I fed my initial "old man yells at cloud" reaction into an AI and asked it to separate the arguments from the emotions. The result is below. Feel free to blame either of us.

Every open-source governance problem is increasingly being described as a funding problem. Sometimes it is. But before concluding PHP needs more money, I'd like to know why a language used by so much of the Internet can't attract enough contributors under its current structure. Funding may be the symptom rather than the cause.

I'm also struck by how much of this story revolves around coordinated disclosures, embargoes, privileged access, advance distribution of patches, and leaks. Open source originally leaned heavily on transparency, yet modern security practice often seems to move toward security through obscurity. The irony is that this very episode is largely about how difficult that secrecy proved to maintain.

If important vulnerabilities, mitigation advice, videos, binaries, and CVE information all leaked anyway, perhaps we should be asking whether the bottleneck is funding, or whether increasingly elaborate disclosure processes have created governance and coordination problems of their own.

"all those "saved" watthours will get sucked up by their subscribers' homes. Then what, shut down the datacenter until things get warm again?"

Why assume the saved watt-hours just disappear? If local residential demand absorbs them, Google can still buy power from elsewhere. The interesting economic question is why Google finds it cheaper to pay consumers to curtail usage than to acquire additional supply through the grid. That suggests congestion, transmission limits, [price gouging, psychological noise,] or peak pricing are the real bottlenecks, not total electricity production.

Is he too lazy?

If no human will help me, is does asking a chatbot seem better than winging it on my own?

When's the last time a human failed you?

Copilot: That argument assumes AI can only repeat what it has seen before. But some of the most interesting AI results, including in mathematics, involve combining known ideas into solutions humans had not previously written down. If synthesis never produced anything new, those results should be impossible. The real debate is whether those successes are common or rare, not whether they can happen at all.

Copilot: A typical chatbot exchange appears to consume only a few watt-hours. Keeping a human in a heated/cooled office with a computer and network connection for the same interaction may consume an order of magnitude more electricity.

"It will never fly in the US, unless there is a huge political shift." Has Trump suggested basic-income-adjacent policies?

Gemini:

"Universal" means unconditional, but more importantly, it offers a path away from total market dependency.

Imagine funding basic income through direct currency issuance, with payouts fully indexed to inflation so buying power remains immune to price gouging. Combine that baseline financial security with open, localized technology--think repairable hardware, local micro-grids, open-source automation, and zero subscription fees.

When individuals have both a financial floor and the tools to run independent, self-contained local systems, they don't just survive inside the market; they opt out of its worst traps entirely. The ultimate purpose of a basic income shouldn't be keeping people tethered to corporate supply chains, but giving them the autonomy to build sovereign alternatives.