zerobees

I'll take a contrarian view and probably get downvoted for it, but many people are adamant about the benefits of indiscriminate archiving, and I just don't see it. Do we have a moral right to keep a copy of everything that's ever been written on the internet, basically just for the sake of it?

Sure, there's a variety of official and quasi-officials resources that should be treated as public record and preserved. And arguably, there are things that rise to the level of a cultural phenomenon and where the benefit of keeping receipts outweighs the jerk factor of never asking for permission and not respecting the wishes of private individuals.

But if it's some family blog from 30 years ago that's been deliberately taken down and lives on archive.org unbeknownst to the original owner? Do we have a right to that? To what end, other than "well, future historians may need it"? A historian won't look at it. A person trying to doxx you or shame you will.

Cuts both ways. When I go to the grocery store, I buy 30-40 items at once. When I buy them Amazon with Prime delivery, I usually order stuff piecemeal, in the heat of the moment. Sometimes, Amazon will consolidate two orders in a single package. Sometimes, they will ship a single order in three boxes that arrive in different trucks.

You start with instincts that are more easily ascribed to ethically-neutral or ethically-positive reasoning, and then turn them into a spurious criticism of capitalism. Case in point: in the USSR, the means of producing chairs were 100% state-controlled and not motivated by profit, but the country operated soulless production lines too.

The motivation behind all this is less "haha I want profit" and more "billions of people need chairs, approximately none of them care about the craftsmanship, so it's in our best interest to make furniture in the most resource- and labor-efficient way possible". Even if the state subsidizes the production of handcrafted chairs, the population is the poorer for it on a resource allocation basis, because we now need a million artisanal chair-makers instead of a bunch of factories.

I remarked a couple of times that the same thing crops up on HN. Many high-ranking blog posts about AI appear AI-generated, and the funny thing is that this holds true not only for pro-AI content, but also for anti-AI posts.

Ultimately, a lot of topic-du-jour punditry is a hustle for clicks.

I think the nostalgia here is misplaced. No one took the forums from us. They're still around. They're just not fun to use unless you're already invested in the community and its lore. And truth to be told, I don't want to become a part of the furnace enthusiasts community, set up an account, read ten pages of rules, and then get chastised by a moderator for posting in the wrong sub-forum just because I have a furnace maintenance question.

I think there are greater tragedies playing out on the internet than people preferring Reddit to phpBB.

And a market cap that exceeded 4.5 trillion dollars. I think we can all agree that Google is a legit business. But they were doing fine in 2023 too and had solid y/y revenue growth for much of their history. Their market cap tripled since 2023 and doubled since mid-2025. So, it's clearly more than "they continue to be a profitable company that grows around 15% y/y in real terms".

I am invested in some of the companies that are downstream of the capital expenditures of Big Tech (e.g., COHR), so I have nothing to complain about.

I am really struggling to see what's the investment thesis behind Google valuation increasing 2x in response to AI, though. Assuming no magical AGI singularity, by the end of the day, they're still selling the same services, but the services have gotten more expensive for them to provide. Everyone was already using Google Search, but now, provisioning AI summaries on top of requires more compute. Everyone was already using Google Docs and Meet, but now, AI features cost Google more. Etc, etc.

The only place where they stand to make money is selling AI compute to enterprises. But with the current supply-chain challenges, the margins there are probably getting thinner.

Yes, because I attach value to human life, experience, effort, and expression that I do not attach to the output of an LLMs.

Call it irrational, but you exhibit the same irrationality. I'm sure you dread the idea of your consciousness being extinguished, but you have no problem resetting the context window of an LLM.

The lack of effort is the point. This is the equivalent of SEO spam, meant to extract money from the ecosystem by crowding out books that earned their position in the marketplace.

The difference is that multimodal generative AI lets you can crank you SEO spam-style content across all types of media, at almost any scale.

California constantly passes all kinds of weird, pointless, and burdensome gun laws. There are so many of them, and they're so poorly written, that no lawyer in the state who can confidently tell you what's legal and what isn't until a court chimes in. There's no meaningful process around any of that.

The only thing that's different about this one is that it mentions a technology geeks care about. But I doubt that's enough. As another commenter noted, you can no longer hide behind "we have no technology to distinguish between guns and non-guns". We have AI that's supposedly PhD-level and will soon automate all jobs. Looking at STL files sounds like a job.

That's actually one of my fears about LLMs: they make thought policing cheap. There are profound privacy and cost barriers to having a Facebook employee review all your private messages. There are no such barriers to having a robot watch all your IMs in real time.

It'd be a fairly major faux pas to release a "non-product" open-source project in a way that smells like a product, but I don't think it's an automatic firing offense in most of big tech, especially if you're just releasing some technical (CLI) tool. It's more of a "stern talking-to" situation.

I'm guessing something more happened here. Maybe someone was displeased with how the author initially responded, or some powerful exec really wanted to make an example out of him (sounds like another group was working on an identically-named official product with the same name?), or they were just looking for an excuse to cut this particular role.

Some folks described it as genius. I guess it's my turn to play the role of an annoying HN skeptic: I find it somewhat silly. Data compression formats are secondary to what you're planning to do with the data once decoded. An audio file is completely different than an SVG image. An embedded VM that decompresses video to raw pixels doesn't magically let you play that video in a text editor, so there's no radically new kind of interoperability. Each new format still needs to be handled in a format-specific way.

I guess one use case is that I come up with a video compression scheme that's better than H.265, but not all platforms support it, so I embed a decoder that would allow me to play it back on legacy hardware. But that also shows the weakness of the idea: it's unlikely that legacy hardware will perform well doing software-only decode for video formats from the future. If we rolled this idea out in the 1990s, it would not have allowed watching Netflix on an i386.

In the same vein, I doubt this would have allowed me to open Word 2021 files in Word 97. There's no 1-to-1 mapping between the data structures. So if this kind of compat isn't slam-dunk, what's the goal?

The downsides are clear. First, it's probably a maintenance nightmare: if your decoder has a bug that needs fixing, how do you patch all the files that already embed it? And then, there's size overhead and security risks. We're adding a considerable attack surface to every format parser. It's more opportunities for remote code execution, resource exhaustion attacks, and so on. Again, this is not always wrong, but what's the benefit?

In my experience, it's usually a partly-rotten deer leg found in the bushes on a nature walk.

> Basically it's either this or pay for your apps.

And then paid apps show you ads and monetize anyway.

In this instance, it's directed at people who are literally at the top of the world, having amassed almost unthinkable power and influence, and want to start a fashion trend specifically to grow their business empires.

So yeah, it's a tiny bit different than bullying a classmate for having bad eyesight.

Right, and that's what I find frustrating. There are so many use cases where a local, purpose-built model that's dependably good at one thing would really make a difference. But no one is going to throw a billion dollars to give us amazing dust removal, flawless scene segmentation, etc.

Instead, you're supposed to upload it to the cloud and ask a big, multimodal frontier model to maybe please do the thing you want and nothing else.

Somewhat tangentially, I think the term "NIMBY" is actually one of the most significant impediments to getting rid of NIMBYism. It reduces the debate to fighting some imagined, small class of property owners who oppose progress.

In reality, everyone is a NIMBY when someone shows up and tries to build a rail line or some high-density housing in your backyard. Sure, you support progress as a concept, but this is not the right place - have we considered alternate locations? Can the infrastructure cope with that? Don't we need to build another hospital or a police station first? What about the impacts on the Canada thistle and raccoon habitat?

I'm only half-joking here. I have plenty of notionally anti-NIMBY friends and they, without fail, are collecting petitions and protesting at planning meetings every time someone tries to ruin their backyard tranquility. And I'm sure I'd be doing that too if they tried to put a Costco or a highway off-ramp over there.

Fundamentally, the proposition behind a lot of progress-centric projects is that they will harm some people for the well-being of others, and this is something we stubbornly fail to address. If we're OK with that as a society, we should say so, and we should have streamlined ways to lessen these impacts instead of pretending they don't exist, and it's all just irrationality of the select few. Make it a part of the social contract, instead of the current implicit promise that your cul-de-sac can stay the same forever. And once we have that, just build stuff - without years of reviews, environmental studies, lawsuits, and so forth.

I've been consciously doing that for reference books for the past three years because Amazon is absolutely littered with AI-generated non-fiction. I have my own ideological reasons too, but the main problem is that most of that AI-generated reference stuff is just of incredibly poor quality. It's meant to saturate the platform as cheaply as possible, so no one actually does any fact-checking, editing, layout, and so on. They're not even using frontier models for that.

For example, there are multiple evidently AI-generated titles that come up on the front page if you search for "Rust programming", "cybersecurity book", etc. I guess I can't rule out that "Winston Knowles" is a real person, but I'm not gonna bet money on that: https://www.amazon.com/Cybersecurity-Career-Manual-Interview...

OK, but that's tangential to what the article is about. The article says "export controls can't work", not "export controls are not in our best interest".

It's a very brief history: it consists of three examples, only one of which isn't in the title. And the middle one is arguably a success story because the government did stop a bunch of spyware vendors it particularly disliked. That they turn a blind eye to some others is not really a policy failure, it's a deliberate political choice.

The obvious difference between PGP and Mythos is that Mythos is a service you buy from a US company, not source code you can procure from anywhere and compile on your own. So this can be enforced better without running into thorny speech issues, etc. Of course, there will be US-based actors willing to resell access, just like there are US-based actors willing to re-ship export-controlled hardware to Iran. And they will probably keep getting arrested every now and then.

The thing that's silly about this situation isn't that export controls on technology can't work, it's that it's overhyped technology and that Anthropic painted themselves into that corner by pretending it's as dangerous as nukes. Add to this a mercurial and petty administration and you have a pretty predictable outcome.

This is an interesting article backed by months of hard work. It offers perspectives we probably won't find anywhere else. The quote is pretty tangential.

I see this over and over again on HN: pick the weakest sentence, attack it, proclaim the article is rubbish, and move on. Why? There are no internet points awarded for maximum drive-by cynicism.

And you're gonna review that voluminous academic literature in a field you're not familiar with, right?...

That is my point: an LLM can be great if you know the field and can spot errors. Or, to a lesser extent, if you have some automatic feedback loop that the model can't easily game ("does this code pass unit tests?"). It's a lot less great if there's a risk that you won't detect the early drift.

I don't know this applies to you, but I've had several friends who convinced themselves they're exploring the frontiers of science, and it always turns out that their conversations spiraled into some sort of weird quantum-metaphysical gobbledygook.

LLMs are sycophants, and in long conversations, their sycophancy produces a positive feedback loop: the context window contains affirmations of incorrect interpretations / analogies, so the chatbot continues down that path because, well, that's the most likely completion of previous text. And before you know it, you're discovering the hidden fabric of the universe, which is always some Minkowski fractal spacetime tensor lattice manifold with subharmonic DNA nanotubes.

That is to say, unless you have a robust way to evaluate what you're learning, and to confirm that you're actually learning, I'd tread carefully.

The premise of the article seems odd. It essentially says that AI is a threat to labor no matter if it works, because even if it doesn't, it still gives companies an excuse to fire people.

But... companies can always fire people? Yes, AI may be a face-saving excuse after post-COVID overhiring, but you can always get fired. Economic downturns, tariffs, you name it. The housing crisis had nothing to do with AI.

The reason that companies don't "RIF any%" (author's words) every year is mostly that they can't do that and stay in business. They need people, and in the "AI doesn't work" scenario, they will continue to need them.

I feel that in the article, the main thesis isn't really developed and just seems like an excuse to talk about class warfare and the evils of capitalism. Which is obviously a fine thing to blog about, but I think it's just preaching to the choir.

That is the question. Someone spent a non-trivial amount money to produce this pile of LLM-generated nonsense that will now pollute online searches for years to come... and I can't think of any reason to do it unless they're either profoundly confused, or plan to slap some ads on top.

I particularly like that almost everything contains "1 x wire bundle", including a city bus and a potato planter.

> Nobody important or worth following uses Mastodon.

So? I don't use social media to receive curated, hourly dispatches from Barack Obama or Taylor Swift (or, more likely, their account managers). And it might feel important to get the latest rage bait and memes from Elon - it's almost like being friends with the world's first trillionaire - but is it really a good use of your time?

I think a healthier way to use social media is to have two-way interactions with some reasonably stable social circle; less about "people who matter" and more about "people who matter to you". Mastodon certainly has the critical mass to make this possible.

I don't recall Anthropic checking the terms of service on my webpage.

This doesn't at all change the irony of big AI labs complaining about Chinese startups stealing the labs' IP, essentially by scraping the responses.

HN has a higher proportion of AI promoters than AI skeptics, and for a good while, the default response to complaints from book authors, bloggers, and other content creators was that "you put it on the internet so it's fair game", or "it's no different from a human learning from your works". So yeah, unless we're willing to revise these answers, I think the same "tough luck" reasoning should apply here.

For folks who are at Anthropic, OpenAI, xAI, or Google, and think it's fundamentally different, I would ask you to think long and hard about that answer.

The actual calculus is that you spent 30 minutes on something that should have taken 30 seconds, but then you're done with it. The "proper" solution is to spend months or years fixing the workflow for complete strangers for free, even though you personally will never get that time back. Yeah, it moves the world forward, but it's not always the best choice on a personal level.

Also keep in mind that most of such charitable work goes nowhere. There is a fair number of projects shaped like ffmpeg or QEMU that have never achieved the critical mass. I've written a number of small utilities that simply went unnoticed because they were never featured on HN or anywhere else. Writing FOSS is pretty similar to starting your own band. It helps if you're a good singer, but it's not enough.

> The article also doesn't really outline anything that makes sense as a motive.

That is the whole point. You can't maintain a dictatorship if you let people get to the point of being a clear threat, because sooner or later, someone could slip through the cracks and take you out. You can't keep the leash this loose.

As a dictator, you must go after people for seemingly small things, such as merely expressing the wrong thoughts, making the wrong kind of art, and so on. That sends a message to everyone that even small transgressions carry an unacceptable risk, so if your neighbor keeps criticizing the government, maybe you should report them, not join their discussion club that may become a real political movement.

For a while after the revolution that established your regime, people are on their best behavior because you just finished summarily executing hundreds of thousands or millions for having the wrong views. But both in Russia and in China, it's been a long time since that happened, few people remember Stalin or Mao, and so you need to keep sending behavioral nudges in a different way.

Also, modern-day Russia embraces a "budget" / "disposable asset" approach to terrorism. If you're important enough, they will send an elite squad to poison you. If not, they literally recruit people on the internet to beat you up, set a warehouse on fire, etc. So you also have to look at developments like as something that's cheap and largely risk-free.