armchairhacker

They can guarantee that certain small changes won't create a big change in performance (provably, verified through tests, or trivially). An LLM can't.

AI could be deterministic, only it requires sacrificing performance and nobody cares enough. Although algorithms are more predictable, and you can follow an algorithm's steps (even a complex algorithm can be broken into large sequences of steps, which may be broken further).

In a production-grade compiler like LLVM they are rare enough that it compiles giant projects like Linux and Chromium without detectable issues.

Maybe eventually Claude can, since it can write tests, but compilers / (deductive) algorithms have other advantages like efficiency and predictability.

Those are the exact concepts I was looking for.

A compiler is an algorithm and Claude isn't*. A compiler almost never produces a wrong output, even when compiling an extremely complicated program. But a compiler must be clearly defined and is limited to input/commands it's defined for. A compiler will "correctly" process input into unintuitive output, but that's not always what the user intended (e.g. omitting large sections of code that are undefined behavior). Claude makes mistakes, but can process any input/command expressible in training data, and has more potential to infer when something is unintentional.

Tools like Claude are great at building algorithms, then tests to verify them. The algorithms they build aren't just compilers, but any software that can be verified. Even though Claude may not be a compiler, it has the potential to (at least if it improves) create production-grade software that can be verified, like other compilers. But maybe not good UX without human input (expanded to any subjective experience, like video games).

* Not a "source language to target language" deductive algorithm. Technically Claude is an algorithm to predict the next token, but acting as a compiler or anything else it's an inductive heuristic, because it guesses (https://stackoverflow.com/q/2334225)

Many Democrats support surveillance bills, like the KIDS act (https://www.govtrack.us/congress/votes/119-2026/h228) 104 Democrats supported and only 85 opposed (and 23 abstained).

Combine with liquid glass and add a hover effect and raytraced lighting for, finally, a decent UI.

Actually, I benefited from a school-issued Chromebook. I mostly used it for note-taking (I'm bad at handwriting) and homework. Even though it allowed some simple browser games, I suspect it was less distracting, and the browser games could've been banned if the school used a whitelist to make it even less so. Even with paper students can be distracted by doodling and daydreaming.

But I wouldn't oppose banning those as long as the schools provided an AlphaSmart (because again, I'm really bad at handwriting). I think e-readers would be best but those are expensive.

I think banning tools for teachers, like projectors, is unnecessary. I suspect I learned easier from projected computer screens vs. staring at a whiteboard, but anything projected on the screen (e.g. powerpoint presentations) could be passed to students on paper, except movies but those may rarely be necessary.

They’re limiting freedom of children in school, who already don’t have the freedom to ignore other school policies.

Studies show phone bans help school performance, albeit not by much immediately, but there’s not much upside.

Seriously, why isn’t this implemented everywhere?

You're right. I don't see why a device would be age restricted by default, when practically always an adult has first access (too expensive)

What about your family? Are they ok with that?

I'm single without kids so yes

And we both know that the involved museum employees do "not het enough pay", this is 2026 we're talking about

Context from the parent comment

There is no real reason why a museum can't be open at 2 A.M., it's just that it's economically impossible, unless someone pays enough money.

I’d work a museum at 2 A.M. for enough pay and time off.

An analogous good implementation of age restriction is that one must show their ID to buy an unrestricted device if they look too young.

They can't be tracked, as long as the devices are in randomly sorted identical boxes. Of course someone can buy a device and give it to a kid, but that's already possible with alcohol (and legal if it's their kid).

"Age restriction" can be implemented where stock devices are completely unrestricted. For example, just with better parental controls and education: children don't have access to property, so the parent takes their device and enables the parental controls, but you own your device so leave them disabled.

And this wouldn't affect Linux or FOSS: on a child's device their parent installs either a proprietary OS or a FOSS with parental controls, but again, on your device you install whatever you want.

This conflates 1.0 and 2.0.

1.0 (passed) means just don't expect encryption on big tech cloud and messaging apps, which you shouldn't expect anyway. Signal is fine.

2.0 hasn't been passed.

"Banana" and "Extra Fancy Unsalted Mixed Nuts" only occurred once? They go together.

The graph starts falling shortly after 2020. AI certainly contributed but Stack Overflow was dying without it.

Hackathons were unfair long before AI. See https://news.ycombinator.com/item?id=48468766

The solution is to host and join hackathons without prizes. The point isn't to win, but to create and present something cool and have fun.

If anything, AI's assistance making a fast prototype means hackathons should be better.

I think the appropriate solution here is a generous, well-advertised refund policy.

People aren't allowed to (not even defame) write about others? Amazon doesn't allow certain books? No, but I imagine most people who actually buy this kind of book (if anyone does) are expecting higher quality.

And on the off chance they're satisfied, everyone wins. Even the author, since apparently the biography is very flattering of her.

How old did you have in mind?

Somewhere between 70 and 80

All of my colleagues who have died in old age (at least late seventies) did so in full command of their senses, one of them still contributing to international technical committees into their eighties.

Likewise, there are 14-year-olds and 16-year-olds smarter and more mature than most adults. Loss of rights doesn't mean they can't contribute, have others deny opportunities or abuse them; people would still look out for them, they would have more protection than children in this regard. Mainly, it makes it easier and more common to control finances, prevent from driving, and prevent from rotting in front of the TV old people who have lost their senses.

I’m saying that old people should also be differentiated by age: once you reach some age (maybe 70 or 80, the same age for everyone) you lose certain rights. This is separate from the existing system.

I think limited rights for old people are like limited rights for children: justified because there is cognitive decline, and every individual (except children who tragically die young) gets to live some life with full rights.

The biggest problem is that it’s depressing. A child gets to look forward to growing up and having full rights, an old person is already looking forward to declining and dying and the loss of rights reflects that. Another problem is that, just as some old people are scammed, taking away rights will have other old people abused by their “caretakers” (e.g. one relative hurting them and stealing from them even against other relatives, which already happens).

So I want to see it implemented but tied with our culture restoring respect for old people, giving them a sense of purpose, and looking out for those who aren’t our relatives. Specifically including reforming retirement homes, many which take advantage of their residents, and stronger safeguards against abusive relatives who (already are seizing and) abuse POA. I’d like to hear what old people themselves think, because maybe I don’t (empathetically) understand the consequences, but although maybe outside the Overton Window logically it seems reasonable.

We are closer to ending sexism and racism than we were 70 years ago before any affirmative action existed, right?

Correlation does not imply causation

Who’s to say that all the backfiring and unintended consequences that have happened aren’t better than the parallel universe in which maybe slavery still exists?

They are better, but it's a low bar and an even lower one

Would it have happened faster with no quotas or boosts for groups that were historically discriminated against?

Maybe, but even if "no", keeping the quota may cause a backslide.

Artificially balancing preferential treatment by applying it badly in the wrong direction is a duct-tape solution. Sure, it may have been justified after the Civil Rights Act. But it's not ideal and doesn't seem justified anymore, when we can easily positively discriminate for traits that are more obviously and universally negatively discriminated against, like income.

The world that Wikipedia was created in no longer exists. They can’t grow in the new world without changing. They don’t want to change, so they no longer grow.

Probably because the new world seems to be worse at open collaboration. For example, they worry that accepting new maintainers (without careful screening which ensures, deliberately or not, they won’t change the organization), they will accept ideologues and Jia Tans. Wikipedia already has a serious problem with these and dedicates lots of resources to finding and stopping them, it’s why they’re hesitant to accept new contributions; but opening the organization to change would very likely make it worse, because new changes have new exploits.

vlian2088 is also correct that they’ve already been captured by a specific group which has shaped it to their biases. From that group’s perspective, even some good changes are bad. But opening it could introduce more hard-to-reverse bias.

Personally, I think we should try to figure out a new model for a new Wikipedia that thrives in today’s world.

Except preferential treatment backfires.

People will think “if XYZ group has a handicap maybe it’s because XYZ group is genetically inferior?” XYZ members themselves will think that and it will subconsciously affect them. People around them will think that and it will subconsciously affect their opinions towards them.

If you point out that XYZ group is only handicapped because they’re statistically environmentally disadvantaged, then it follows, why not handicap everyone with that disadvantage, or any comparable disadvantage? Why not handicap ABC minority? Some members of ABC will be jealous of XYZ and subtly discriminate against them (for this reason; these members would otherwise).

It creates the background conditions it seeks to destroy. Instead, handicap on things like health and income, which are more obviously fair and necessary (most people can accept that bad health and income are an especially serious disadvantages in today’s world).

I'd say the purpose is usually to get a "career" i.e. good job, possibly as a paid researcher and/or professor.

Not to be confused with Eärendil-1, the experimental satellite to recover solar energy at night, approved to launch this year: https://removepaywalls.com/https://www.nytimes.com/2026/07/1...

generally considered illegitimate

But why? Obviously an unjustified cryptominer is bad, like unnecessarily slow JavaScript, but this one has a good purpose and to the user is no different than PoW.

Still better than AI, it has personality and not paragraphs of empty prose