oorza

Which of these political entities is in a better situation because of an open phone rather than partnering with Apple and Google? Anyone with the funding to make a Linux phone happen loses money and/or power making it happen. And users do not care, less than 2% of users will ever leave iOS or Android for Linux unless it's a substantially easier and more accessible experience for them, and we all know that will never happen.

They're becoming more popular in Florida, as a poured concrete geodesic dome is basically immune to a hurricane.

Are we at the point where there are external constraints that cash can't solve?

See: Claude Cowork trying to establish an entire new group of people in their ecosystem.

Is it the models themselves or the tools around them? There's that patch[1] that floats around for Claude Code that's supposed to solve a lot of these problems by adjusting its tool-level prompts. Also, if it were the models themselves, wouldn't Cursor users have the same complaints (do they? I haven't heard anything but the only Cursor users I talk to are coworkers)?

I think it's more likely they're trying to optimize the Claude Code prompts to reduce load on their system and have overcorrected at the cost of quality.

1: https://gist.github.com/roman01la/483d1db15043018096ac3babf5...

Also, seeing genuine original creations created with AI assistance is much more interesting to me

The great disappointment about how all of this is marketed is what AI should be good at doing - enhancing a tiny budget - is all but forgotten. I don't want a video of Pikachu fighting Doctor Strange, I want some weirdos fantastical horror movie that he could never get financed, but was able to green screen and use AI to generate everything. I don't want a goofy top 40 country song full of silly lyrics, I want musicians to use AI to generate new sounds as part of composition.

In the same way that there's a difference between vibe coding and using a coding assistant...

the problems it did flag mostly would be caught by minimal testing

Testing is more expensive up front and in maintenance than type annotations. A test suite comprehensive enough to replace type annotations would have an ass load of assertions that just asserted variable type; if you were involved in early pre-TS Node, you remember those test suites and how they basically made code immutable.

(2) it regularly missed deeper problems

This is a skill issue. If your types do not match runtime behavior and you choose to blame the programming language rather than your usage of it, that's on you. There are a lot of unsafe edges to TS, but a diligent and disciplined engineer can keep them isolated and wrapped in safe blocks. Turn off `any`, turn on all the maximal strictness checks, and see if your code still passes, because if what you said about infinite scroll is true, it won't.

TS doesn't "bin off" duck typing, it's a fundamentally structural type system. It's statically analyzed ducks, all the way down - when nominal behavior is preferred, people have to bend over backwards. Either you are using the wrong vocabulary or I don't think you've bothered to actually learn Typescript. In any case, it's the programming language that successfully brought high-level type system concepts like type algebra, conditional types, etc. to their widest audiences, and it deserves a ton of credit for that. The idea that JS and Ruby and Python and PHP developers would be having fairly deep conversations about how best to model data in a type system was laughable not that long ago.

There's a certain amount of variance in the way that people utilize these agents. Put five people in a room and ask them to compose the same prompt and you have five distinct prompts. Couple this with the fact that models respond better/worse to certain prompts depending on the stylistic composition of the prompt itself. And since people tend to write in the same style, you'd get people who have more luck with one model over another, where one model happens to align more readily with their prompt style.

To wit, I have noticed that I tend to prefer Codex's output for planning and review, but Opus for implementation; this is inverted from others at work.

I'm not sure you need AGI to clear that bar; I'm not sure you need more technology than currently exists beyond iterative improvements to things like how expensive it is to train a model.

But let's say it's free-ish to train a model, so you decide that that's how you're going to write the next Marvel movie. You train an LLM specifically on screeplay writing, teaching it to cross reference literary techniques with audience reaction, you teach it the sum total of Marvel canon, you teach it the sum total of the American cinema canon, you train it on all the social media reaction to either, and so on. You teach it to specifically engineer screenplays for Marvel movies that Marvel audiences will do maximum Marvel fanboy shit about. Do you genuinely believe such a dedicated model couldn't output a Marvel movie that everyone would love as much as Endgame?

Obviously, in the economy of 2026, it is cheaper for Disney to hire flesh-and-blood writers instead of doing this madness. But one day it won't be - and this is hardly even the tip of the iceberg. The ability to finely hone models quickly and on-demand (potentially on a per-prompt basis) would unlock another tier of accuracy and performance from LLMs, and for some/most artistic tasks, I think that gets you to "indistinguishable from mass market media."

The "smart but lazy" person in an agentic AI workplace is the dude orchestrating a dozen models with a virtual scrum master. It's much more possible today to get a 40h work week's worth of work done in 4h than it ever has been before, because the gains that are possible with complex AI workflows are so massive, particularly if you craft workflows that match problems specifically. And because it's absolutely insane to do such a thing with modern tools and the lack of abstractions available to you, even insaner to expect people to do it, so you can't set proficiency targets on that rubric. You might have to actually work 40h at the onset, but I definitely work with someone who is considered a super hero for the amount of work they do, but I know they dick around and experiment all day every day, because all they do is churn Cursor credits into PRs through a series of insane agents. They're probably going to get a bonus for delivering an impossible project on time, as a matter of fact.

Anyone you'd interact with in a job in a HN-adjacent field has already cleared several bars of "not actually that lazy in the big picture" to avoid flunking out of high school, college, or quitting their office job to bum around... and so at that point there's not that same black-and-white "it'll help you but hurt you" shortcut classification.

I'm clearly not talking about the _truly_ lazy people. I'm talking about classifications within the group of already successful creative/STEM professionals that are the ones who are going to be maximally impacted by AI. Obviously you're not as lazy as you could be if you manage to have a 20 year software career, but that doesn't mean you aren't fundamentally lazy or have a terrible work ethic, it just means you have a certain minimum standard you manage to hold yourself to. That's the person I'm talking about - the person who works twelve hours a day more isn't going to be able to meaningfully distinguish themselves any more. The quantity of their work becomes immaterial, so what matters is the quality, and the smarter, lazier dude is going to have better AI output because he has smarter inputs.

I don't disagree, but I would argue that the reason people prefer human works over AI works is the dynamic I mentioned in the original comment. I play a lot of idle games and it's not uncommon for one to start becoming popular, it's revealed that it's vibe coded, and the community turns against the developer; one game dev was bullied out the community and deleted his entire online presence because he wasn't ashamed of using Claude. It isn't about anything but the mob mentality of "AI bad."

The point you make about "romance" novels is true: there's tools like SmutFinder which are effectively exactly what you describe. You can be as specific or generic as you like, you can lay out the specific plot points and chapters or let the AI do it for you, or you can build the entire story one paragraph at a time like a super interactive choose your own adventure novel. And it's all modeled on smut, specifically built for the AI to design the user's specific fantasy. Smut books are arguably the lowest and easiest bar to clear in terms of audience-acceptable quality, but this technology existing in this space today assuredly means it'll be available for science fiction novels of acceptable quality in short order and eventually science fiction films, and eventually all media.

But the presence of personalized works in an AI marketplace makes what I said more salient because it serves as yet another bar to clear for someone else's artwork to become relevant to me. Why would I consume your space opera when I can make my own that's "better" according to my judgments? There's obviously reasons for me to buy an author's work even in an AI world because I want to be surprised, but if there are dozens / hundreds of options from equally qualified creators, what makes yours special?

I think there's still a very high chance that someone willing to refine their AI-co-generated output 8-10+ hours a day, for days on end, will have much more success than someone who puts in 1 or 2 hours a day on it and largely takes the one of the first things from one of the first prompt attempts.

That's the truth right now, but that's merely a limitation of the technology. Particularly if you imagine arbitrarily wide context windows such that the LLM can usefully begin to infer your specific preferences and implications over time.

The most successful people I know are in a category you leave out: the people who will put in long hours out of being super-intrinsically-motived but are ALSO naturally gifted creatively/intelligently in some domain.

Those are the people I mention at the end, those that clear the bar into being uniquely special. From what I hear from my friends that have been teaching for about twenty years now, you're lucky if you get more than one or two of those every ten years.

I think the story makes a good point, but I'm not sure it's even the primary point the story was trying to make.

“Writing a book is supposed to be hard,” he said.

“Is it, though?” said the AI. The novelist wasn’t sure, but he thought he detected a touch of exasperation in the machine’s voice.

“Perseverance is half the art,” he said. He hadn’t had much natural talent and had always known it, but he had staying power.

It's this right here. I don't think any LLM-based AI is going to be able to replace raw human creativity any time soon, but I do think it can dramatically reduce the effort it takes to express your creativity. And in that exchange, people whose success in life has been built on top of work ethic and perseverance rather than unique insight or intelligence are going to get left behind. If you accept that, you must also accept its contrapositive: people who have been left behind despite unique insights and intelligence because of a lack of work ethic will be propelled forward.

I think a lot of the Luddite-esque response to AI is actually a response to this realization happening at a subconscious level. From the gifted classes in middle school until I was done with schooling, I can always remember two types of students: those that didn't work very hard but succeeded on their talents and those that were otherwise unexceptional beyond their organizational skills and work ethic. Both groups thought they were superior to the other group, of course, and the latter group has gone on to have more external success in their lives (at least among my student peers I maintain contact with decades later). To wit, the smart lazy people are high-ranking individual contributors, but the milquetoast hard workers are all management who the smart lazy people that report to them bitch about. The inversion of that power dynamic in creative and STEM professions... it's not even worth describing the implications, they're so obvious.

Let's say, just for the sake of argument, that AI can eventually serve to level the playing field for everything. It outputs novels, paintings, screenplays - whatever you ask it for - of such high quality that they can't be discerned from the best human-created works. In this world, the only way an individual human matters in the equation is if they can encode some unique insight or perspective into how they orchestrate their AI; how does my prompt for an epic space opera vary meaningfully from yours? In other words, everything is reduced to an individual's unique perspective of things (and how they encode it into their communication to the AI) because the AI has normalized everything else away (access to vocabulary, access to media, time to create, everything). In that world, the only people who can hope to distinguish themselves are those with the type of specific intelligence and insight that is rarely seen; if you ask a teacher, they will recant the handful of students over their career that clear that bar. Most of us aren't across that bar, less than 1% of people can be by definition, so of course everyone emotionally rejects that reality. No one wants their significance erased.

We can hand wring about whether that reality ever can exist, whether it exists now, whatever, but the truth is that's how AI is being sold and I think that's the reality people are reacting to.

No, but you could do enough of it in secret with Elon Musk resources to prove that it's both planetarily viable and doesn't cause catastrophes by existing and then lend your political weight to having it scaled up globally. By the time the public heard about it, it would already be a done deal.

I think you could prove it out at a scale that people could measure on planetary CO2 sensors for a couple dozen billion dollars, then take that data to a sitting POTUS you're friendly with and work out a multi-trillion dollar commercialization plan, using the USA's global bullying power to immediately establish a global monopoly.

A particularly cynical view would be this CEO buying global laws that dictate carbon neutrality while simultaneously also making it impossible to achieve without his CCS. Then merely canceling a sales contract topples a regime and you've arrived a global corporatocracy.

I mean, if I had Elon Musk money, I'd build some kind of giant carbon capture mechanism. Perhaps I'd buy the largest basalt quarry I could find and start sequestering carbon at a planetary scale. It would cost a ton of money, but I'd do it in secret. If it worked, eventually it would show up on the scales, and I'd emerge from the shadows. This particular method of carbon capture could potentially work at a planetary scale and could potentially be done in secret, at huge cost, but the only blocking factor today is money.

https://eos.org/articles/basalts-turn-carbon-into-stone-for-...

This is the answer to carbon storage by the way, people just do not know about it. There's more than enough reactive mineral sites on the planet. The process is basically just dissolving CO2 into water, heating it, and soaking basalt in it to allow crystals to form. The water becomes heavier than ground water and can simply be poured into the Earth. The unsolved problems are optimization problems: direct air capture of CO2, using saltwater, that sort of thing.

If the world's billionaire class decided to buy carbon sequestering, we could have global CO2 levels returned to 1900 levels within a decade or two. The technology exists, the economic willpower does not.

https://www.bbc.com/news/world-43789527

Potentially, basalt could solve all the world's CO2 problems says Sandra: "The storage capacity is such that, in theory, basalts could permanently hold the entire bulk of CO2 emissions derived from burning all fossil fuel on Earth."

Having said all of that, this is likely the most dystopian option. It's the "tech bails us out, yet again" solution because we could deploy it thoroughly enough that we can solve climate change without addressing any of the existential issues that got us here. The right combination of corporate+government partnership commercializing this technology and making it mandatory is a very plausible way to arrive at "there's 4 corporations on Earth that run the show" a la Aliens.

HTML purists who do things without JS. They are the real web developers.

I don't think using one set of technologies as compared to another one can really be said to make one a "real" web developer. Real web developers are developers who put sites on the web, there is no benefit to anyone to be had claiming one choice is "real" and therefore the other choices are lesser-than.

Put it this way: whatever set of constraints you used to arrive at that decision does not apply to every situation, and when you frame things through the lens where you implicitly disregard that oh-so-obvious truth, it's hard for anyone to interpret your analysis as anything but myopic in the best case and actively self-serving and destructive in the worst case. It's nearly impossible to read through someone speaking this way about a topic and believe their analysis is objective, comprehensive, or without obvious bias, even if it may actually be all of those things.

I remember back in the before times... when escape analysis debuted for the JVM - which allows it to scalar replace or stack allocate small enough short-lived objects that never escape local scope and therefore bypass GC altogether - our Spring servers spent something like 60% less time in garbage collection. Saying enterprise software allocates a ton of short lived objects is quite an understatement.

You don't even need an AppleTV. My Roku and Amazon TVs were both zero-configuration AirPlay targets. And generally speaking, there's not any issue with wifi streaming video - there's a noticeable input lag, but it doesn't desync audio, and videos tend not to be interactive.

I don't think it's necessarily any larger of a leap than any of the other big breakthroughs in the space. Does writing safe C++ with an LLM matter more than choosing Rust? Does writing a jQuery-style gMail with an LLM matter more than choosing a declarative UI tool? Does adding an LLM to Java 6 matter more than letting the devs switch to Kotlin?

Individual developer productivity will be expected to rise. Timelines will shorten. I don't think we've reached Peak Software where the limiting factor on software being written is demand for software, I think the bottlenecks are expense and time. AI tools can decrease both of those, which _should_ increase demand. You might be expected to spend a month outputting a project that would previously have taken four people that month, but I think we'll have more than enough demand increase to cover the difference. How many business models in the last twenty years that weren't viable would've been if the engineering department could have floated the company to series B with only a half dozen employees?

What IS larger than before, IMO, is the talent gap we're creating at the top of the industry funnel. Fewer juniors are getting hired than ever before, so as seniors leave the industry due to standard attrition reasons, there are going to be fewer candidates to replace them. If you're currently a software engineer with 10+ YoE, I don't think there's much to worry about - in fact, I'd be surprised if "was a successful Software Engineer before the AI revolution" doesn't become a key resume bullet point in the next several years. I also think that if you're in a position of leadership and have the creativity and leadership to make it work, juniors and mid-level engineers are going to be incredibly cost effective because most middle managers won't have those things. And companies will absolutely succeed or fail on that in the coming years.

Tailwind might not be the most perfect fit, but it's "just" CSS.

Where was this line of thinking when it was Obama ordering the DEA to not enforce marijuana laws? Where is this line of thinking when it's a city that chooses not to enforce dog breed restrictions?

The enforcement of law being separate from the passage of law is a key plank in a functioning democracy, it's one of the safety valves against tyranny.

And for every video of quality on the platform, there's one that's blatant political propaganda, one that's blatant conspiratorial misinformation, one that's sexualizing children, etc.

It's a mixed bag. It has no more to offer than any other social network. Less, some might argue, because of how easy it is to crosspost to the other video networks.

The only way this is different from the loss of other social networks, Vine most closely, is the government is shutting down the site and collapsing the ecosystem rather than private equity.

I think you'll find most people in leadership positions at most companies are not that forward thinking, proactive, or frankly intelligent. I thought cost-benefit and risk was analyzed on most big company decisions, until I sat in rooms for a Fortune 500 where those decisions were getting made. If you assume that everyone everywhere is doing just barely the minimum to not get fired, you're right more often than not.

C4 still smokes them both, doesn't it?

All the web developers I've worked with in the last four years understand the concept of a bottom type, even if they've never been introduced to the formal phrase, because `never` is the explicit bottom type in TS and appears all the time. You can't go very long consuming libraries written in TS before you run into it, at which point you become familiar with it.

Similarly, a function that accepts (explicitly) `A | B | (C & D)` and then dispatches to functions that accept `A | (C&D)` vs `B` is, you guessed it, type algebra and is a common pattern in hot paths through every TS codebase.

Just because the formal nomenclature is unknown does not mean the concepts are unfamiliar.

There's three buckets of performance in interactive software: so fast it doesn't matter to the user, slow enough the user notices but doesn't lose focus, and slow enough the user has time to lose focus. The lines are obviously different for each person, which is why some people feel that software is "fast enough" well before others do.

The jump from an i9 to an M1 moved a lot of tasks from group 3 into 2, some tasks from group 2 into group 1, and was the biggest perceived single-machine performance leap for me in my professional career. I have an M1 Max or Ultra on my work machine and an M3 Ultra in my personal machine - after two years of the work machine being so visibly faster, I caved and upgraded my personal. The M3 Ultra moves a handful of things from group 2 to group 1 but it's not enough to move anything out of group 3.

I am genuinely curious if those barriers have technical justifications. There's a pretty stark difference (to me, at least) between ignoring standards in order to reinvent better wheels and intentionally diverging from standards to prevent compatibility.

It's a question of whether they're _not_ investing resources to maintain standard behavior or they are actively investing resources to diverge from it. If it's the former, I don't find any fault in it, personally speaking.

Sure, but ten years ago, if I told you the concept of a "bottom type" would be common knowledge with web developers or they'd consider type algebra second nature, you'd have laughed at me. If I were to tell you they pushed a Turing complete type system into the mainstream and then abused its computational power to do things like create value types, you'd ask me where I was buying my LSD.

TS might not be the most robust type system in the world with no qualifiers. And it might not even try to be sound. And it might disappear at compile time. But none of that changes the fact that it took a bunch of concepts that only programming language dorks knew or cared about and turned them into every day utilities that the junior most developers use. And it doesn't change the fact that it is the only mainstream programming language with a fully Turing complete type system.

Give credit where it's due. I say all of this as one of the aforementioned programming language dorks who took a really long time to get on board with TS, as I thought Flow's focus on purity made it a better choice.

I had the same experience working at one in US from ages 15-19.

The thing that stands out about that job that makes it unique in my perspective is decision-making. Every other job, from dishwasher in a sit down restaurant to engineering team lead, works best when I make good decisions and my coworkers all do too. Everything runs most smoothly when everyone is on their game and consistently making the right choices.

At McDonald's, everything ran most smoothly when no one made any decisions at all.

Their system was literally that good. No one else's has ever come close. And you're right that there's a weird freedom and pleasantness to that: your job requires literally zero brain power most of the time and your sole responsibility is to keep up with the person behind you and not overwhelm the person in front of you, which isn't usually something you worry about. So all that energy every other job has got from me, I still had (and so did everyone else) which made it the most social, most fun workplace too.