Xirdus

The obvious question is how does it work for linked lists, trees, and other non-stack-shaped data structures?

You can encode extra things in IDs no problem. E.g. Polish civil records system uses PESEL numbers which, on top of being unique IDs, also encode date of birth and gender.

As long as you have absolute control over every single ID generated, it will work just fine. The problem is that it's way too easy to accidentally generate UUID in a way that ignores whatever rules you set up for it. Also, clocks are hard.

1. Exile to which country?

2. Who gets them their immigrant visa?

3. Where do they live while waiting for it?

4. You do realize most countries require immigrant visa applicants not to be convicted felons in their home country?

Ironically, one of the leading arguments against capital punishment is how expensive it is. Including all the legal costs of appeals and reappeals, it's approximately 1 to 3 million dollars per convict (in the US).

"Wealthy people" alone is also consistent with "not many people". There aren't that many rich people in general, musicians or otherwise.

And oil shortage is another way of saying that existing oil owners are earning strong income. Why is either of those things the desired state? (Other than self-interest, if you happen to be sitting on a stockpile of oil.)

"Highest earning people who didn't come from affluent backgrounds" is entirely consistent with "not many people".

You can share your work with others while still not caring about the others when working. Others using AI to duplicate your work only matters if you care about people actually using your software, as opposed to getting some likes when you publish your project etc.

I still don't get where the pressure comes from. Ultimately, the only target audience of a hobby programmer is themselves. There is never a reason to rush things, because your only stakeholder - yourself - is fine with whatever your only developer - also yourself - does. You always have all the time you'd ever want for the technical puzzles because there are never any deadlines except self-imposed ones.

Unless you mean your hobby is other people using your software, rather than programming for the sake of programming in the literal sense. In which case, yeah, sucks to be beaten by AI on all fronts at the same time.

But why does someone else delivering faster deprive you of the opportunity to learn these things anyway? Can't you just... not look at what they've done?

Hospital staff have a certain reputation too. I've even heard nursing career being described as the female counterpart of police career, in that both of them attract the exact same type of person for the exact same reason (having people's lives in their hands). Can't say from my own experience if it's true or not, but I heard it often enough that I'm not terribly surprised.

LEOs being a certain type of people is a big part of the problem. Compare to corporate IT, who have huge surveillance powers within their companies, but virtually none of them use it beyond their actual job duties.

Developers are guilty of making reasonable projections based on contemporary industry trends, and they're getting slapped by unexpected revolution in computer science.

I think most of the difficulty in explaining monads is the IO monad specifically. Leave that off and monads seem almost trivial. Add it back and whatever trivial mental model you just built stops working altogether.

Yudkowsky doesn't stand at your place of worship waiting for you to come out. You have to actively reach out, look him up and open his posts to see a single word of what he's saying. At that point, he's the host and you're the guest; if you came to him to read his posts, it's not a fair complaint that you're exposed to his opinions, even if his opinions are critical your own opinions. It's more akin to being a Jehovah's witness, knocking on an atheist's door, asking them to talk about Jesus, and being mad when they say the Gospel was made up. It would be very rude to say at a church, but they're not at a church, they're at his home.

I'm pretty sure Yudkowsky aims at revealing truth through questions and tackling interesting ideas too, at least in his head. Maybe his questions are wrong or his idea of interesting is wrong, and that would be a valid criticism, but it's a completely different argument than that he shouldn't be asking any questions in the first place.

As far as I know, HPMOR is meant to be a parody, not a serious discourse. It's unfair to judge it like a serious discourse. His blog posts, sure, we should judge them by the serious standards, because AFAIK they're absolutely serious.

I've read a very cool article on this exact subject recently. https://permeations.substack.com/p/i-would-have-been-killed-...

Coincidentally, also references Butlerian Jihad.

They're building hype around their product by solving genuinely hard problems? To quote xkcd[1]: Mission. Fucking. Accomplished.

[1] https://xkcd.com/810/

Maybe I'm biased by my young age and lack of knowledge but aren't modern day hackers way more anti-social-norms (piracy, crypto, doxxing) and against "the man" (directly antagonizing big tech, E2EE proselytizing, fighting against electronic surveillance, circumventing online ID requirements) than they were back in the 70s-80s? I always imagined pre-internet hackers to be spending their nights messing around with university computers, not undermining business models of leading IT companies, but maybe I'm wrong.

I meant hackers, not ham operators.

"The Kickflip" was done without a ramp, wasn't it?

If there were no such laws, then what were they rebellious against?

IIRC during the 2022 snowstorm, it was only impassable one way. Jumping over to the opposite side should be doable.

May come in handy if someone blocked your car on a parking and you don't have time for the tow truck to arrive. Or when you're stuck on I-95 during a major snowstorm.

I think it might be because ham radio is still barely legal and the community wants to keep the few freedoms they have so they try not to piss off any groups of interest. Whereas pretty much everything hacking-related is already a felony thanks to DMCA and other similar laws, so they don't have much to lose.

I know nothing about Yudkowsky, I think I read like 10 of his words in total over the course of my life. I was commenting on the general principle of aggressively dismantling every worldview all the time. I think it isn't just a good thing to do, it's the only way to arrive at a worldview that actually holds together.

If you meant Yudkowsky and co. are doing the aggressive dismantling wrong, or aren't dismantling their own views aggressively enough - that's fair, I have zero knowledge on the subject and would love to hear more. But if you mean that aggressive dismantling is wrong in and of itself, I think it's totally fair to bring Socrates and Rousseau into the discussion.

Aggressively dismantling other (and their own!) worldviews is how we've got Greek philosophy and the Enlightenment, among other things.

What's sufficient information is also a huge part of the debate. See: should we start worrying about AI safety or not yet?

Opinions based on insufficient information help immensely in determining whether we have sufficient information, and also raise awareness of important topics that we need more information on. These blog posts serve a very important role in the information exchange ecosystem and the whole world would be worse off if all these people held off until they're 99% certain. Nobody would be talking about climate change, about economy, about quantum gravity, about ancient civilizations, about prehistoric animal species, etc. etc. and so on.

On the contrary - I found rationalist bloggers to be the most willing to change their mind in face of new evidence out of all groups of people I've ever met o the internet. Although HN commenters are a close second.

I thought only as constrained and more constrained is allowed, and less constrained is always an error. At least that's been my experience - if `Q<T: A>`, then I could never get `impl<T> Q<T>` working in any conditions until I added `where T: A`.

Do you happen to know the specific rules of when it's an error and when it's not?

On further reading, the Drop impl is a red herring. The bug is unrelated to Drop. It's there in the repro to show the effects of invalid assignment, but even without impl Drop the bug is still there.