zerobees

You're arguing with LLM-generated text and yes, I don't think the photos actually show that. They don't seem to be making any political point at all.

The thing to understand about the USSR is that Moscow was a flagship city of a continental-scale empire obsessed with projecting an image of power and technological progress. It had grand construction projects, cultural events, subway, good schools, paved streets. Sort of like Pyongyang, if North Korea was a global superpower. The thing that sucked about Moscow wasn't that it looked drab, it was that you could get disappeared to a gulag or outright murdered for political speech or merely pissing off a government official, and that the government managed almost every aspect of your life (including where you work and live). Forget foreign travel, you even had restrictions on domestic travel. People born in rural areas couldn't move to Moscow unless they had political connections of some sort.

Life was far more miserable in the rest of the USSR, including all the republics and satellite states that Moscow approached as sources of cheap resources and labor to prop up the capital. Around 6-8 million people starved in the USSR in the 1930s (most of them in Ukraine). Another million starved around 1947.

Replying as a person who acted a hiring manager for decades: their loss.

Hiring for technical roles, I love long-winded answers as long as they are coherent. I don't want slogans, I want to understand how you reason through problems and that I can trust your judgment. Everything else is secondary.

Note that this doesn't mean "rambling". Get to the point. But if you want to show me nuance / reasoning, I want to hear it. Also a good way to spot bullshitters.

> Not having used agents and not being able to comment on what to do and what not to do with them is immediate out since early this year.

I've not observed that anywhere. It takes a couple of hours to figure out how to use agents. It might be a slight negative if I suspect you're not demonstrating enough curiosity about what's one of the most significant developments in tech in a long time, but the assumption is easily overcome if you geek out about something else.

I'm not as negative about this, but with the benefit of hindsight, it's easy why the current initiatives didn't go anywhere.

It's not that people are not willing to make sacrifices. We repeatedly did this in times of famine or war. Europe during WWII is a perfect example. Another good example of a major cultural shift in response to a new threat was the AIDS epidemic. The entire sexual revolution went out the window and we're now in a world where young people have a lot less sex than ever before. We like to talk about gender and sexuality, but we do a lot less with it, so to speak.

Anti-consumption / degrowth arguments face an uphill battle because they basically say "you should live a harder life". There should be less stuff, the stuff should be more expensive, there should be less of you. So you need a good answer why this is the right choice. Doing it "for the planet" doesn't sound too convincing because we're also a part of the planet and most people feel entitled to it. It's hard to get others to make real lifestyle sacrifices because you showed them some photos of koalas or coral reefs.

Because koalas don't cut it, we started giving increasingly apocalyptic, doomsday-type answers, all the way to renaming "climate change" to "climate crisis" or "climate disaster / catastrophe". That was probably a mistake. It created a sense of inevitability (so might as well have fun while you can) and undermined the credibility of the proposed solutions. Is it really going to save us all if I'm sorting my recyclables into five different bins?

So in a sense, I think this is a PR disaster more than anything else.

You're being pedantic. By your standard, I also have a relationship with the DMV, and to avoid a "toxic relationship" (parent's language), I should be honest with them about all the times I rolled past a stop sign.

The parent was talking about giving a diplomatic answer about your attitude / use of AI. I'm not talking about Enron here.

> Have you considered just answering truthfully? ... That sounds not like a job but a toxic relationship.

It's a job, not a relationship. It's best not to confuse the two.

In any workplace, you will occasionally have to do things you find boring or objectionable. And if you're hoping to find a corporation that is a "perfect match", it will only hurt more when they unceremoniously fire you because the quarterly revenue growth is 1% off or because you cracked an off-color joke.

I don't think that's fair. There's a lot of talent and grit behind ffmpeg. But for better or worse, getting the code to do what it's supposed to do requires a different mindset than getting it to not do anything else (i.e., to handle malicious inputs correctly).

The developers of ffmpeg are very good at the first thing and not very good at the second. But few people on this planet, if instructed to write a complex video format parser in C or assembly, can produce something that's secure on the first try. The main failing of the ffmpeg team is that they should have spent more time on architectural hardening and mitigations. Most other large projects of this type do.

My impression is that in the past year or so, IEEE journals have been leaning pretty heavily into low-quality, AI-generated articles. And looks like this author produced not one, not two, but three career advice columns in a single day - impressive:

https://spectrum.ieee.org/u/brian-jenney

Ffmpeg has an exceptionally terrible track record when it comes to security. People have been throwing fuzzers at it for as long as I remember and coming back with a nearly inexhaustible supply of memory corruption bugs. Here's an effort by one Googler a decade ago:

https://security.googleblog.com/2014/01/ffmpeg-and-thousand-...

So, while it's a demo of the capabilities of LLMs, this should not be at all surprising. Ffmpeg is absolutely not something you should be running outside of a sandbox if you're touching any untrusted or user-supplied content. I know that people do, and these people are taking unreasonable risks.

I disagree, for two reasons. First, blogging is nearly-free for people who have non-demanding, well-paying jobs. Sure, I find it tacky when an engineer working at Google is asking for donations, but I accept that for some people, a small trickle of extra cash may be the difference between being able to blog or having to work two shifts.

Second, blogging is cheap if you're writing about what you had for lunch, but there are numerous personal blogs and websites that are obviously costly and time-consuming to run. For example, I'm sure that Ken Shirriff's blog costs him money. Similarly, if you wanted to recreate Marcin Wichary's blog, it would cost you tens of thousands in photo gear and travel costs.

> I'm so tired of rampant capitalization and constant advertisements. That's why I'm here on the indie web, to get away from all of that, and now you are here, begging me for my money.

And that just comes across as entitled. The indie web doesn't owe the author anything, and it especially doesn't owe him the adherence to his views on capitalism.

I do. I'm familiar with his research from the pre-Google era, and that's that.

Right. I have respect for Rene, but going by the "HN quitting manifestos" like this, Google has lost its moral compass at least 50 times at various points in the last 20 years.

The company is constantly changing, but also hasn't changed all that much. It always talked the talk and was eager to tell others how to behave, but was almost never willing to give up any real revenue to do the right thing. The usual justification was that if Google doesn't do it, someone else will (and that someone else is obviously not as moral as Googlers are).

If you're old enough, you might remember that they vocally opposed privacy-violating, disruptive display ads. That was their whole schtick. But that was before they realized there's a lot of money to be made by acquiring Doubleclick.

No. There was no "unit". This was before the days of modular PC PSUs and switching wall warts (which started proliferating only later). So it was just a custom circuit that used commodity components. For these components, the volume of orders from Apple would have been tiny compared to overall market demand.

Outside of some niche specializations like cryptography, math isn't practiced because of "consequences". Most mathematicians take pride in their work not having any obvious practical applications. They're also overwhelmingly working in university settings where they're not expected to generate revenue or deliver practical results.

We basically subsidize the practice of mathematics as an art form, and if you try to take the artistry away, you might find that the artists don't want to play along. And I guess you can imagine future robo-math production lines without any human involvement, and then LLMs finding applications for the resulting theorems, but it's not possible today.

> Arguments about beauty don't lead anywhere constructive because they are too observer- and context-dependent.

Meh. You can successfully argue that there is no objective anything. It's all just our perception and the emotions we associate with it. We built entire civilizations on subjective notions of good, evil, beauty, and so on. So where do you draw the line between "acceptably subjective" and "too subjective"? And are you sure it's not just a subjective code name for "the thing I don't like"?

Ultimately, people practice mathematics mostly for abstract reasons. It's not a field where you routinely ship products and get rich by meeting market demand. If 99% of contemporary mathematicians don't want to become prompt engineers, there's nothing that makes the transition to AI math inevitable. If not mathematicians, the only party with vested interest in that would be the PR departments of frontier labs.

This has been available in the US for a long time. I had it on the car I bought in 2016, and on another I bought in 2023. It's just not mandatory, so it usually comes as a part of some safety / driver assistance package. And even if you have it, you need to enable it.

> Gift your house to the city with a deed that says they have to rent it back to you forever for $1 a year?

What if I write a note saying that you need to pay me $10,000? That's not a contract, that's just a fever dream. But if you shake my hand and sign that piece of paper, that's a different story.

The same applies here. If you get the city to agree (and they don't get raided by the FBI after that), then sure, they should be bound by the deal they made.

Note that this works both ways. If you own nice rural acreage, the federal or state government will often be happy to pay you some token amount and give you a tax break for a conservation easement that prevents not only you, but all future owners, from using the land in certain ways. It's still yours, but it's now a scenic corridor and you can't build there anymore. There's plenty of such easements in California and other Western states. If I'm bound by such a perpetual, deed-attached restriction, why can't the government be?

Having had a subscription to Hack-a-Day for a long time, I firmly believe that the vast majority of "weird one-off" Raspberry Pi projects don't actually need anything as capable as a Raspberry Pi SBC. It's just a matter of brand recognition and familiarity. If it gets too expensive, I suspect that more users will migrate to microcontrollers than to gutted notebooks.

You don't even need to learn anything new, I'm sure you can ask Claude to vibe code something on RP2350 nowadays and there's an 80% chance it will work.

It is a property of any large bureaucracy that a large proportion of the bureaucracy exists to serve itself. And it's not BS, it's a natural consequence of growth. Imagine you start in the mode of "moving fast and breaking stuff". Eventually, you break enough stuff that someone says "enough". So you develop some launch standards and guidelines. Then hire a team to enforce them. Then someone to build a launch tool. Then you realize you also need to manage legal risks, have standardized UIs, make sure that production services have backups and redundancy, and all of sudden, you have ten review processes. And then, it gets so difficult to navigate the process that teams hire PMs just to coordinate. And on, and on.

And then, someone needs to build cafeteria menus. And the tool to manage health care enrollment. And badging. And ultimately, you have a product that could probably be operated by a lean team of 100 people, but you have 5,000 employees to take care of all the auxiliary functions, from legal compliance to providing benefits. You need slack in that org structure too, because you don't want everything to grind to a halt when one important person leaves or takes a week off.

I don't understand why you find this objectionable. Would Google or Facebook be more fun if you were on a very small team with zero slack and constant grind, and there was no one to call if the printer is broken? Yes, it's a jobs program funded by the revenue from core services, but it ultimately makes life better for everyone?

In common with most of Europe, German legal system is not based on case law. It's more firmly rooted in formal laws and regulations and judges are not required to follow precedent.

> It’s all about who’s ass you kiss and who’s ass you don’t ... I think it’s exceedingly rare that a CEO is actually competent at their job.

But... that is kinda the job? CEOs are, first and foremost, the public face of the company. They're the one who talk to VCs / banks, regulators, major customers, the press. They're very highly paid PR reps / fall guys that shield everyone else, including the board of directors and all the VPs and SVPs, if something goes wrong.

For most companies, especially large companies, it's not important for a CEO to be good with software engineering, business development, etc. That, at least in principle, can be handled by other parts of the hierarchy.

Skepticism against AI text detectors on HN is as old as time, and frequently comes from people with some vested interest in filling up the internet with slop when you look at their business ideas / projects / blogs. I've done systematic testing on human and LLM-generated text and I'm confident that the accuracy is higher than 95% (and that <5% is almost exclusively false negatives).

You shouldn't crucify people based on this alone, but if it reads like AI, quacks like AI, and is detected as AI, it's probably AI.

It reads that way and Pangram says it's AI. And my experience says that if you see an AI-related headline on HN, there's a 50%+ chance that it is AI-generated and meant mostly for clicks.

This article appears to be 100% AI. I guess there's some irony that a company ships an AI feature and someone else uses AI to come up with criticisms of that feature. But the article... doesn't actually say anything?

It's just full of weird, generic short-sentence LLMisms ("Detection is observation.", "Changing the password is authority.", "The security benefit is real.", "That is a meaningful improvement.", "This is not just text generation. It is an agent taking action with a sensitive credential.", ...). It doesn't offer any insights into the actual architecture that Apple came up with, whatever it might be. It doesn't propose a better design, other than a bunch of super-generic things that apply to every single software project ever ("The system should verify the exact website and account before filling or changing anything.", "This feature deserves focused adversarial testing during the beta period."). So... it's upvoted just because the title mentions Apple and AI?

Given that the US govt is reportedly talking to OpenAI about taking a stake, your only choice might be Zimbabwean dollars.

Forums of the early 2000s were almost always sorted by recency, not upvote count. They also typically weren't dominated by vendor press releases and news stories, whereas domains such as anthropic.com, blog.google, openai.com, along with outlets such as techcrunch.com and arstechnica.com, are probably among the most popular URLs of the past year.

But I don't think it's a meaningful distinction to begin with. Usenet was an endless time sink to get angry at things that didn't matter and argue with strangers who might not even be real people. It wasn't monetized, but it still made it easy to waste years of your life.

The people on this site are, overwhelmingly, the people who already "came together" to build businesses like Facebook, Instagram, Twitter/X, and so on.

We pretend we're the victims, but none of these platforms would have been built without an army of willing, enthusiastic, highly-paid engineers who made small "ethical compromises" every day.

And now that there's money made in something else, many of us would accept a seven-digit offer from OpenAI in a heartbeat, leaving the task of figuring out the downstream effects to other people.

The LWN article is incorrect in saying that it "must copy the entire process state (including memory) for the child process". There are some kernel structures and page tables that need to be initialized, plus you need a new stack, but it's not nearly as dramatic as implied. Most of the parent's memory is "incorporated by reference", so to speak.

In fact, if you profile it, in the fork() + execve() model, execve() is far more expensive, because not only does it replace the old process with a new one, but it also involves running the dynamic linker, which opens, parses, and mmaps library files.

It still makes sense to get rid of the fork() overhead if you're going to throw away the cloned process state soon thereafter, but if you wanted to make process execution radically faster, rethinking the exec architecture would probably offer more significant gains.

I know that people with strong feelings one way or the other will comment here, but note that this is specifically about problems with known answers that can be inferred from existing literature (e.g., training data).

This is an interesting result, but as I understand it, it's not about solving frontier challenges (which LLMs can evidently do too, but that's not what's tested here). It's closer to "can a mathematician (blindly) write exercises you can't cheat on using an LLM". "Blindly" in the sense that they can't adjust the problem ahead of the time until they get a model to fail.

The conclusion in the paper is: "The concept of writing exercise-style benchmark questions based on publicly accessible research has reached its limits when it comes to the best-performing available models."

I get what you're saying, but I think there's a contradiction between wanting to be a passive index-fund investor and having opinions like that. The core tenet of index investing is that the market knows better than you.