News for hackers. Stuff that matters.

julianeon

A subtle point I haven't seen mentioned here:

Most AI use cases involve the Internet. So let's pretend that OpenAI agreed w/your logic and did this. They'd have tons of businesses on the Internet, whose lifeblood is Internet traffic... and on the other side of this, controlling the spigot of all this Internet traffic, is Google.

All Google would have to do is to wake up one day and say "we've decided to depriortize low-quality AI businesses to protect the consumer" and they'd be dead.

Re:Docker Is Weird on Friday October 23, 2026 @12:36AM (#50204565)
That's a good answer that effectively rebuts the article's claim that Go was a bad choice imo.
Re:Vibecoding isn't as fun as writing code by hand on Tuesday October 06, 2026 @11:21AM (#49979868)
I think you should split the difference and use vibe coding for the parts you don't care about.

Example, the UI for his two vibe coded example apps. Why not let AI do a single pass on that? 30s for a huge improvement. Other than the personal pride of saying you did it all, I don't see why not.

I think we've overcorrected.

For what it's worth, the "first patients were lab workers" part isn't established. What is documented is that the earliest cases cluster around the Huanan market. The lab workers attribution shouldn't be carrying the argument.

Before COVID, people were just unable to be roused out of their sleep about the dangers of cross-species virus transmission. "Wake me up when something important happens." By a certain standard nothing important had ever happened: so people slept on.

Then COVID hit. "Time to wake up." Suddenly people went from basically unconscious to hypervigilant overnight. Something which was a clear and present danger - the wet markets - looked uninteresting compared to the lab. Environmental samples taken there tested positive. In other words: don't sleep on the wet markets again - they are dangerous, then and now.

While paying attention was arguably an improvement over complete indifference, I don't think we landed in the right place.

Re:I Could've Accessed 17T Microsoft Records on Wednesday September 30, 2026 @05:09PM (#49914498)
I personally think it sets a bad precedent: you want to broadcast that valuable info will be treated as such. A hacker in the future will see this, compare the "precedent" price to the black market, and not even ask Microsoft the next time.
Re:Rails World 2026 Opening Keynote [video] on Friday September 25, 2026 @01:43PM (#49847706)
Have you tried OpenRouter and the DeepSeek-like models? If you try to optimize a little you could probably slash those costs.

I send a ton of requests and even reaching $10 for the month would take more calls than I have capacity to review and use. And as I said, that's for a ton of requests.

Re:LGBT tolerance slumps in the Netherlands as debate turns vicious on Wednesday September 23, 2026 @02:30AM (#49812424)
I think we tend to forget about the importance of reversion to the mean, socially.

We think: Germans are like this, the French are like this, the Dutch are like this... without thinking that, when one moves out of sync w/the others, its likely, statistically, to move back into line w/the others.

This has nothing to do with left or right specifically. It means that, whatever the mean is, if a country is not there, it's likely to move in that direction over time. It's more likely to move closer to that than away from that, measured over years.

The Dutch were very liberal and still are in many ways, but in such a situation, they are more likely to move in the "less" than "more" direction. And this is probably such a case.

Re:The End of an Era on Friday July 31, 2026 @04:59PM (#49128456)
I agree about the writing. As for the cover art though: I'm indifferent. It's not an important part of the book.

In the literary world, I would call using AI sentences in your novel a felony, but I'm not sure using it in a cover would even qualify as a misdemeanor.

Re:The Maxwell Conjecture Is False (GPT 5.6 Sol) on Friday July 31, 2026 @11:41AM (#49124508)
Is the bar really higher? Those math PhD's can use GPT too; they benefit equally from AI assistance.
Re:The Fabled Flatbreads of Uzbekistan on Tuesday July 28, 2026 @04:35PM (#49089579)
I could probably spend a whole month eating fresh bread there every day and enjoy it more than a Michelin starred restaurant.
So how would I use these Chinese models by API? I assume I'll pay by API call.
Re:A full body MRI earns you a year of smoking on Monday July 13, 2026 @03:39PM (#48897728)
Since I think people are getting this confused:

He's saying that a full body MRI is so valuable, that if you got one and then smoked for a year, you'd be in the same place as you were before you started, health wise. No loss.

Obviously you'll come out way ahead if you don't do the smoking part: it's very valuable, in other words.

Re:Focus on Thursday July 09, 2026 @11:47PM (#48855537)
You know what's going to happen is some budding CEO will come along and read this and conclude, "I know the solution: my company won't give to charity." It will become enshrined as this principle in their company lore, that nothing can detract from The Work. Managers will parrot it, HR may even add it to onboarding. But it will be an empty gesture, because as the company grows it will become more and more multifacted, or 'distracted' if you will.

I'm not sure the process can even be stopped, if the company is successful and the new changes appear to be, and probably are, profitable.

To a physio therapist (the article example) that is a business meeting. That's what's strange here. That person isn't there because their you're friend: they're getting paid to be there, which makes it a business meeting, which makes a secretary, human or AI, appropriate.
Re:The Revenge of the Philosophy Majors on Tuesday July 07, 2026 @12:40PM (#48820243)
I've noticed that many famous billionaires want to be viewed as philosophers: Thiel obviously, Musk arguably.

For this they do need ideological coherency and the ability to order their arguments logically, ideally as part of a larger program. Since it is such a popular destination late in life, you'd think it would be a good choice for a major too.

Re:Did my old job only exist because of fraud? on Sunday June 21, 2026 @08:50PM (#48624228)
That's the job: experiment until you find product-market fit... or die trying.
Re:Pre-2022 Books on Saturday June 20, 2026 @08:21PM (#48614429)
This will work for a little while but this can't go on indefinitely: no one is going to want to stick to 30+ year old books.

We should probably work on developing standards for what we want in a book instead of clinging to a losing position.

Re:How many of the 170k English words do you know? on Friday June 19, 2026 @08:24PM (#48604950)
I'll contest a few of these, which I thought were good.

Breviary: this was, to me, known and not uncommon. It's widely known to Catholics, but also, if you have an interest in medieval art or books, you'd likely know it too. It was one of the main types of books before the invention of the printing press. Think of an image from an illuminated manuscript, 50% chance it's from one.

Hippopotomonstrosesquippedaliophobia: it's not that you're expected to know the whole word, but they're looking for you to recognize components of it and infer the meaning from that. I knew sesquippedalian (sometimes jokingly used in "long word" contexts) so that was easy: but phobia is also easily identifiable, and hippo, from the latin root, I knew was not as obvious as the animal, but probably something like "large" (clue: the Hippodrome). So you could, even knowing only "phobia" and being able to guess "hippo", have a good basis for your choice.

Complacent and gauche: have heard both these uses, I think that's straightforwardly correct. If this was a dictionary that would, at worst, be the 2nd or 3rd definition. No complaints.

Source: I used to place in spelling bees and could've been a contender but I didn't have the discipline to study the dictionary for hours on the weekends, which is the next level.

Re:Spotify Killed the Thrill of the Hunt on Friday June 19, 2026 @02:20PM (#48601497)
I don't think Spotify should be used for music discovery. I know it has that feature and I know the app wants you to embrace it, but I would say: don't use it.

Use YouTube for this. You can find experts, fans, all kinds there who will guide you through genres and act as non-algorithmic, human-sense-making curators. Use Spotify only to find the specific tracks and albums you already know the name of, to listen.

He has identified a real problem, and the good news is there is a solution.

Re:The New Bibliomaniacs on Saturday June 06, 2026 @02:05PM (#48427388)
There are some fields where there are a real measurable advantages of physical books, essentially as an archive.

I can name two:

1) Chemical Engineering 2) Classics

In both cases the physical book may be the only place to find certain kinds of valuable info.

In the case of Chem E, I was told this by my father, also a Chem E, who said that some of these old books contained values and tests that were found nowhere else. And while a lot of that is available in modern form, not all of it is.

In the case of classics, I'm cribbing from David Butterfield here, who has a great book tour on YouTube where expands on this at length (4+ hours).

In the 18th and 19th century the level of education was higher and there were simply more people around who were working at the highest level in the field. Their speculations were written down in physical books and nowhere else. Many of these were valuable and showed new insights you won't find elsewhere else, especially for professionals in the field.

Here's an example. The copies that we have of, say, Homer, are copies of copies. Pretend for a moment that Homer actually wrote in English. We can imagine a line in the "original" (a copy of a copy), that says:

He of the stout quarrel chest said:

It kinda works - stout men quarrel, I guess. But you know what would work even better? He of the stout barrel chest. You can make a case that this was an instance of bad copying and should be corrected in editions going forward, especially if you can cite additional evidence.

Multiple this by the Greek & Roman corpus and all the possible permutations and you have a good reason to turn to those books that earlier writers thought very deeply about.

Re:Did Claude Increase Bugs in Rsync? on Friday June 05, 2026 @02:58PM (#48416729)
If Claude is actually good enough to commit to rsync, of course I'm going to look at that and think "it's good enough for my side project too." And (benefit to companies aside) that is info it is useful to know, if it's true.
Re:Samurai City on Thursday June 04, 2026 @08:01PM (#48406333)
Warfare, or more plainly the deaths of many, seems like a pretty decisive tiebreaker: whatever the flip side of the option that involves the death of thousands is, I will choose that. Of course, without this example, most people would say something like the downside isn't that bad; this case study is instructive because it shows that yes, it is.

It's not a perspective we usually see: the nobility, not as a noble or morally elite class, but as a problem that a successful government can manage and minimize, without violence.

Re:Odysseus – self-hosted AI workspace on Monday June 01, 2026 @03:10AM (#48353536)
That looks good - nice to have an alternative too.
> Personally, I think all funding in California education (other than terminal levels like 4 year bachelors and up) should be a function of the percentage of students that succeed at the next step.

This has the unintended consequence of encouraging schools to eject students who are struggling. For example, if the student has a learning disability, declare that it's too serious for them to handle, and then transfer them to a school that theoretically can.

The system gets gamified and the "top" schools are just ones that reject, socioeconomically, every student who can't pay for tutoring or full-time care, which is a very technical form of "excellence".

He makes good points: I probably would have skated over this if not for the line-by-line reading. After examining it more closely, I think the "population increased by a lot" line seems especially unnecessary.
Re:The Cost of Owning a Home on Tuesday May 26, 2026 @03:18PM (#48284583)
Sure, but from the perspective of the individual renter (who arguably doesn't have the power to change rent control law anyway), their point stands.
Re:Ferrari Luce – Designed with Jony Ive on Tuesday May 26, 2026 @11:56AM (#48281493)
I think they will eventually get to the point of technological differentiation but they've got to start somewhere: they must first have an electric car on the market before they can start experimenting with it to improve the performance 1.5x, 2x, 10x.

At some point, EVs are going to pull ahead of ICE cars not just incrementally, but categorically. Instead of 0-60 in 3 seconds, it'll be 0-60 in under a second: the limits are physics and the human body, not the engine. Full self-driving, native to the architecture. Over-the-air updates that do things like improve the car's range by 5% (Tesla did something like this). And more no one's thought of yet.

So, this is their beachhead. You've got to start somewhere: they're starting here.

Re:The map that keeps Burning Man honest on Thursday May 07, 2026 @03:51PM (#48054016)
Would it though? It seems like it could work, even if people opt to "not comply" aka pay the fine.

Charge $1,000 fee per acre (eyeballing it, that seems reasonable). There are people who will clean an acre to be spotless for $500: not bad for a day of honest, actually contributing to the environment, outdoor work!

If I'm missing something and it actually costs more than I know, raise it to $2,000. If heavy trash needs to be removed also, charge that too, by weight.

And if you don't pay, you're banned.

It's worth a try if you ask me.

Great idea. I starred this on GitHub, where by the way for those are interested, your star will be statistically significant (only about 100 now).
He means "careful that the 9k of revenue doesn't come from ads that the scamming owner placed so that the site could show higher traffic => ad revenue." In other words, paying $2 for ads to send people to your website, to make $1 on the ads that the ad platform now shows to your "audience."

Older »