mtlmtlmtlmtl

Renewable doesn't mean unlimited.

Same thing with data centres in Norway. We're planning to waste heaps of our renewable energy on generating profits for a small cabal of assholes in SV. For what? Luckily there's been more political pushback against it here recently, but it's so tame. Just for once I'd like to see some conviction. Pause all ongoing constructions, and launch corruption probes into all the local politicians that pushed this stuff through. I bet there's lots of corruption there. On a national level, we don't have corruption on anything near the systemic scale you see in the US. Some bad apples, and lots of mole hills made into mountains by the press, but nothing like the US where it seems like most of congress is pretty much openly corrupt.

Local politicians though? Oh boy. Seems like every time there's some ongoing project in a smaller municipality involving commercial exploitation of natural resources(like undeveloped land, minerals, hydropower capacity, etc) there's a mayor and/or municipal council members who profit from it financially or otherwise. They have shares/board seats in a holding company owning the land, local contractors, etc. Seems like this is very widespread, and rarely leads to serious consequences, and even when it does, it happens so late that the project is already built, or so far underway it doesn't get stopped. Shit pisses me off.

I'm sure data centres are no different. Because I can't imagine a cost-benefit analysis of them that makes any sense from the point of view of the local community.

I'll never understand this point. It's not as though that slogan is the entirety of OpenBSDs security posture. It's just a damn slogan, but OpenBSD naysayers will never stop talking about it.

And no, OpenBSD is not the most secure OS out there, and very few serious people claim it is. What OpenBSD is, is the most secure option if you want a simple, no bullshit, well documented unix that can still run on obscure legacy hardware using a potato as your RAM. As such, when you actually compare it to other options in that space, I would say it is the most secure. They care about security very deeply, but they do it within that context. So yes, they still use C, because they need to compile on a bunch of dead platforms. And OpenBSD is never gonna be something like Qubes because Qubes needs things like VT-D and a boatload of memory/cores, it's never gonna be GrapheneOS because they can't rely on having MTE, TPM, etc. And that's fine.

They also deserve a lot of credit for contributions to the wider ecosystem as well. Just to give one example, making sure the OS can provide high quality entropy in all contexts with arc4random, which eventually percolated to FreeBSD, Linux, and even MacOS.

But sure, keep talking about the two holes if you want.

No it isn't. It doesn't matter how real the technology is. The bubble will still pop, because when a technology is seen as revolutionary, you get rampant overinvestment. And only a fraction of those investments will pay off(or, indeed, none, in the case where frontier labs are eventually made irrelevant by open models, which seems very likely to me).

The internet and web were revolutionary, no one can deny that. They completely transformed our society. Nevertheless, the bubble popped. The current bubble is no different.

Yeah, I'm also switching away. They're very sluggish on features. I went through all the rigamarole of setting up my own domain for the email service. Specifically with the idea that I could them automatically filter all my mail into folders by service. But, as it turns out, Protonmail has very limited support for sieve extensions. The crucial one is being able to create mailboxes from sieve scripts. So I can sort emails into folders, but I'd have to go create a bunch of folders, and make a new one every time I sign up for something. I'll be switching to fastmail, I think.

Though, if ProtonDrive became fast finally, I might still be convinced to pay a smaller fee for PD plus their VPN, which is completely fine for my purposes anyway(mostly just torrenting).

Last time I tried ProtonDrive(years ago), the speeds were complete dogshit, I could barely surpass 10 MB/s on a half gig fibre line. Anyone used it recently, and if so has it improved?

Yeah, I'm not convinced either. I'm just saying that's what I think the AI labs believe.

If you look at their actions, they're not the actions of someone who both A) believes their technology to be an existential threat and B) doesn't want humanity to go extinct

If they truly believed both of those things, they would just shut down their companies, because being a billionaire is entirely pointless if you're dead.

I can only conclude that they don't believe both A and B. I'm gonna assume they believe B because actually wanting to exterminate humanity is too comically supervillain esque even for Altman. Therefore they must not believe A. And it makes sense. If they believe A, why are they so incredibly sloppy about security? Why did they outsource part of it to some external firm instead of leveraging their own expertise on the technology? I'm not buying it.

Instead, I think they believe C) that AI will create enormous economic value, D) that value will be distributed across the whole economy by making everyone more productive. Assuming C and D, you get E) for them to capture this value, they must maintain proprietary control of the technology in order to be able to charge everyone else for the privilege of using it.

Assuming they believe E, open models are an existential threat, not to humanity, but to OpenAI and Anthropic.

Their solution: make everyone else believe A, in order to achieve regulatory capture and somehow stop open models from advancing by banning their development, or something. This is a hail mary pass. I can just about imagine them achieving this within the US and maybe even Europe, but China? That ship has sailed.

Not true. Wiles presented his proof in a series of lectures in 1993. The error was found by a reviewer, and he spent over a year trying to fix it. Once he did fix it, he asked colleagues to review it again, and only once it passed that review did he actually publish it.

I'm not familiar with the paper you mention. But it's also worth pointing out that afaik the n log n algorithm itself isn't particularly practical. It's one of these "galactic algorithms" that is asymptotically more optimal, but is so complicated that it's only a real improvement for comically large n. And that's without even considering the mental overhead of implementing and maintaining the thing.

Of course, that's not to say the research is necessarily useless. It's still theoretically interesting to find "better" algorithms if only to shed some light on lower bounds, and so on. And who knows, maybe the line of research could lead to more practical algorithms later on.

Withdrawing it isn't irresponsible. I don't even think releasing it is, because that is often the best way to find problems with it; expose it to the wider research community. Doing PR victory laps based on unverified, unfinished work, on the other hand, is. It is well known, even scientifically demonstrated, that retractions or refutations get much less attention than the initial exuberant PR announcement. And thus this practice contributes to misinformation.

And just to preempt the kneejerk whataboutism: Yes, all of academia does this to varying extents, and the mainstream media are also complicit. And that is also irresponsible. And no, that is not an excuse for OpenAI. Especially when you consider that OpenAI actively portray themselves as some kind of moral arbiter on AI and "doing good for humanity". They should be held to the extraordinarily high moral standards they purport to hold themselves to.

Would you care to share your research? What was your methodology? Did you write it up? Is it published?

Or is it the kind of research you'd prefer to keep shrouded in mystery?

Agreed. There are just too many unknowns. Certainly about how life started here(if indeed it did start here). But even more so about what's out there. There's so much we don't know even about the planets and moons in our own solar system, let alone other star systems. It's only recently we've become able to detect exoplanets in significant quantities at all, and we've still only scratched the surface really. And what we know about those planets is very limited. For most of them we don't know much except their distance to the star and estimated size. For a handful we have some information about their atmospheres, usually only enough to narrow it down to a couple of equally plausible models. It's hard for me to imagine us being able to conclusively detect life on an exoplanet. I think the best we'll be able to do in the next several decades is say we're pretty sure a planet's atmosphere contains chemical compounds for which there's no known abiotic origin. But there will be so much uncertainty there in terms of possible unknown abiotic origins that it would never feel conclusive.

This is such a common HN trope, and it's tiresome.

Yes, lots of things will at any point be beyond the technical ability of most people as long as they remain niche. But as soon as a sizable market exists(e.g one created by a law, like in this case), products and services that close that gap will appear.

Hell, there's even stuff for doing this already, which is not that hard for non-technical people to set up, like Tailscale. I can easily imagine a Tailscale-like product that easily allows friends to use eachothers networks as a VPN with zero configuration or Wireguard knowledge required. I mean, when you think about it it's just Tailscale but more streamlined for that particular purpose.

On what planet is it similar? The labs are taking the copyrighted material and attempting to make billions of dollars from it. And this, in your mind, is "similar" to me downloading a TV show simply to watch it?

Sure, those are the ingredients for evolution to occur. Replication and variation affecting selection.

But it's also pretty reductive to say it's the entire thing. You'd be hard pressed to describe the history of life on earth using only those three concepts. You could do it, but it would be a bit like describing the workings of a human cell exclusively using quantum field theory.

There are many emergent concepts, the understanding of which brings a lot more explanatory power. Things like selective pressure and evolvability. Our understanding of evolutionary biology and evolution in general indeed continues to evolve nearly 2 centuries after On the Origin of Species.

One of the most interesting things to me is that the process of evolution itself evolves. Variation, replication and selection evolve over time. Evolution seems to accelerate. Evolution can even create entirely new substrates for itself on top of the old ones. If you look at all the traits that distinguish humans from our closest relatives, the brain sizes, the slow maturation of offspring, our vocal chords, complex language, etc, there's a single thread running through all of it, namely an entirely new evolutionary substrate: culture. Once culture appeared, evolution accelerated massively. And cultural evolution itself seems to accelerate. It took several hundreds of thousands of years to get to the agricultural revolution, but then it only took several thousand years to reach the industrial revolution/enlightenment. From there it took centuries to invent the computer, about half a century for the internet and now society changes dramatically within a single decade, some might even argue a single year. In certain specific cases, we can even do evolution in the span of hours or minutes (seconds?) now, with genetic algorithms and the like.

And yes, biological evolution, cultural evolution and genetic algorithms all operate via variation, selection and replication, but it should be pretty clear by now how important all the little details and emergent phenomena are for understanding the differences between them.

I actually tried something similar, but failed. I'm curious, how often did you dilute and by how much?

This was my setup: I mixed up 100ml of e-juice in a graduated bottle at whatever strength I was using at the time, somewhere between 15 and 20 mg/ml, can't remember exactly. From that bottle, I would fill a separate 10ml bottle, which I'd use to refill my tank. After refilling the 10ml bottle, I would top up the big bottle with VG/PG back up to 100ml, so diluting by 10%. I was vaping about half a 5ml tank per day at this point, so the 10ml one lasted about 4 days, and I'd top up the tank every morning. The theory being that the stuff in the tank would also get somewhat diluted every time I topped it up, since it would gradually get more of the stuff from the most recent filling of the 10ml bottle.

But somehow I just ended up unconsciously vaping more and with longer puff until I was at one tank per day, and then more than one tank per day.

There is probably some truth to what you're saying, but also keep in mind that tobacco contains more active compounds than nicotine alone, including MAO inhibitors in the form of harmala alkaloids. These can synergise with nicotine and make it more addictive.

Also, I'm dubious about your claim that most people need breaks, or at least any more than vapers do. I was a smoker for over a decade. I don't remember ever feeling like "that's enough, my throat hurts". It's just that smoking was triggered by the onset of withdrawal, which begins after about an hour, and becomes urgent after maybe 2-3 hours. So I'd typically smoke at an interval of 1-3 hours depending on how convenient it was to do so. And cigarettes are expensive, so that factors into it too.

Sometimes though, especially when drinking, I could chain smoke for a couple hours straight while hanging out in a park with friends, especially in the summer.

After I switched to vaping though, I have at times ended up puffing so enthusiastically that I've made myself nauseous from the nicotine. So there's certainly a limit there too.

Edit: another interesting thing I've noticed. Nicotine withdrawal seems much less severe after I switched to vaping. It takes several hours longer to become noticeable, and also feels less intense. When smoking, it took me 1 hour to start craving a cigarette, by 3 hours I'm becoming irritable with strong craving, and by 5 hours I'm extremely irritable, light-headed, and the only thing on my mind is smoking, completely incapable of focusing on anything except how to get some cigarettes and how to get to somewhere I can smoke em.

For vaping, it takes me maybe 7-8 hours to get to the 3 hour mark for smoking. I certainly get cravings long before that, maybe starting at 2-3 hours, but I'm able to push it out of my mind much more easily, for much longer. I don't know why this is. My theory is something to do with the lack of MAO inhibitors. It's also possible that it's a pharmacokinetic phenomenon, but that seems counterintuitive. If I'm ingesting higher doses of nicotine compared to smoking, blood levels will take longer to go to zero. But then I should also have higher tolerance, and therefore a higher threshold for the onset of withdrawal, since tolerance and chemical dependence are more or less the same thing.

Yep, good observation. And it's so exhausting not to take a side. Because people on both sides invariably misinterpret you as being on the other side and strawman you accordingly.

Personally, I agree with the author on some things. The copyright laundering, the PR theatrics in search of regulatory capture, the concentration of power in a handful of gargantuan corporations in a country currently headed for totalitarianism at an alarming pace. The proliferation of slop drowning out all the good stuff. These are genuine issues, and they worry me, a great deal.

And yet, it's become impossible to seriously contend that this technology isn't tremendously useful for certain tasks. I've used it myself with good results. But I don't buy the maximally bullish view that an LLM will eventually surpass humans in all areas. There are severe flaws with the technology that are inherent to its architecture and simply cannot be scaled away. This is as incontrovertible to me as its usefulness. The trough of disillusionment is coming for all of us, it always does. And much like the internet past the trough, we'll have a lot of cool new shit as a result of this technology, and we'll have a lot of horrifying new problems as well. This is the case for any paradigm shifting technology.

That's my view when trying my best to be objective.

Subjectively though, I'm deeply depressed about it all. LLMs have taken everything I love about computing and turned it on its head. I got into computers because they were so different from people. Mathematical, rigorous, efficient, unambigious(yes, bugs can be very confusing but there's always a satisfying and logical answer to dig out with enough effort). Now, apparently, to do the job I (used to) love, I have to spend most my time spoonfeeding instructions to a pretentious, know-it-all asshole who repeatedly fails to follow simple instruction, and yet is somehow still better than me at my job. I hate it. I'm 10 times more productive, and 100% of the joy is gone. I've quit, and I'm now considering what the fuck to do with my life. Something with minimal computer involvement. I will continue to code as a hobby and tool for my own purposes, LLM free, but I won't even bother uploading my code anywhere, because what's the point.

> Absolutely no human ever could be good at chess just by reading a few books, or even every book on chess.

Maybe not, but you'd be surprised how little it takes.

A six year old child can learn the rules of chess well enough to be able to play legal moves only in a single day. And they can improve their game at a pace which is almost frightening to behold. I have taught children, and I've witnessed significant improvement materialise in a single game. LLMs have probably thousands of chess books, games, videos, etc in their training data, yet they are unable to even follow the rules.

This is, at the very least, interesting. It illustrates many of the things brains can do, which current ML systems in general, and LLMs in particular, can't.

Interesting that you say you had no visuals from LSD. How high a dose?

It's surprising to me, because as far as I understand it, psychedelic visuals aren't really akin to daydreams or internally generated imagery, but abnormal processing of external imagery. At least for relatively mild experiences. Something like a DMT breakthrough is a whole other story from what I hear, though I've never been there myself.

First of all, there is no password, unless you somehow have the actual encrypted file of private keys that Satoshi used way back when, at which point you probably know who Satoshi is and could just use the $5 wrench attack anyway, assuming they're alive.

The only information you'd get from the blockchain is the public key. So you'd have to break elliptic curve cryptography to derive the private key in order to sign transactions from the Satoshi wallet. To do that you'd need novel mathematics. Which is possible, maybe, or maybe not. But keep in mind that elliptic curve cryptography has had our smartest minds trying to break it for years, unsuccessfully, as opposed to a single enciphered sentence from an obscure source which hasn't seen nearly as much academic attention.

As for some bug in the blockchain protocol or code implementing it, allowing an attacker to sign transactions without the necessary private key... It's possible that this exists, but I'd expect it to have been found by now considering whoever finds such a thing could stand to earn trillions of dollars from it. That's quite the "bug bounty".

Here in Norway we already have issues with volatile and high electricity prices, because:

1) We're not expanding hydropower generation very much. We can upgrade existing dams for improved efficiency, but no new dams are being planned, because it's politically and environmentally arduous to do so. There's some wind buildout, but that's also a political quagmire due to nimbyism. Nuclear isn't even close to having a political consensus behind it, and even if it did, it would take years. And our geography is uniquely ill suited for solar.

2) Winters have less and less snow, summers have less and less rain, leading to issues filling reservoirs. This will only get worse in the future.

3) The north has a surplus, but the infrastructure for transferring it to the south is underdeveloped

4) lots of the electricity we generate is being sold to central Europe to compensate for Germany's idiotic decision to shut down their nuclear reactors and Trump's idiotic decision to send oil prices over $100

Unfortunately data centres are currently being built, which depresses me. Because who will use them? Giant American tech companies will use them, to lobotomise our youth, poison our political discourse and steal our personal data. Fuck that.

Data centres also create very few jobs beyond the initial construction phase. Overall it's a very, very bad deal for Norway as a nation.

I have some questions as a chess player who barely even understands the rules of go. First these josekis sound like what in chess is called a forced tactical sequence. When you say Shin played out a long complex joseki, how does he do that? Does he have to read/calculate it out over the board(50 moves seems crazy to me unless the search tree is highly constrained by geometry/deduction/very few candidate moves, which does occasionally happen in chess endgames), or is the joseki more of a fixed sequence of moves which he's memorised, only needing to read to "punish" if the opponent diverges?

Second, if it is a fixed sequence, how position independent is it? In chess, tactical sequences end up depending on the entire board state to work when they get sufficiently long. I guess what I'm asking is, could a player with some capacity for stategic thinking recognise this idea and take steps to make the flying knife impossible?

I really should spend some more time learning go, it's such a fascinating game.

Probably a better comparison from the chess world(in reasonably modern times, though perhaps players like Capablanca and Lasker could be mentioned as well. Alas, I don't think FIDE rating existed back then) is Bobby Fischer. In the july 1972 FIDE rating list he held a rating of 2785, the highest in history at the time, with Spassky in second sitting at a "measly" 2660, and only 13 players being above 2600 even.

I'll never understand why people post LLM outputs in forums pretending they wrote it themselves.

You asked Claude or some other agent to make your argument for you, and what you got was a well formatted pile of nothing. Bravo.

I've messed around with pretty much all the fancy new terms to varying extents. Ghostty, wezterm, alacritty, etc. And I'm all in favour of people innovating in this space again. But at the end of the day, I always end up back with xterm. It's reasonably lightweight, runs and is packaged absolutely everywhere, I never have to muck about with termcap files and/or $TERM on remote hosts, and I've never had it break, crash, or exhibit any kind of bug or strange behaviour in over a decade of using it. The only TUI that has some quirks in xterm is Emacs, but Emacs has weird quirks in every term I've tried to run it in, and different quirks in every case somehow, so I just chalk that up to Emacs being weird. I'm sure xterm has some bugs buried in it still, but I bet all of them are in codepaths that are almost never reached nowadays, or at least are extremely niche. It does lack some features like ligature fonts(I just decided not to care), sixels and similar graphics things(again, don't really care) and panes/tabs/yadda yadda(tmux or screen or whatever you like can do all of that anyway).

Never had to dig through the xterm bug tracker to figure something out. I don't even know what version I'm running, or even what the versioning scheme looks like. I haven't changed the configuration since the first time I set it up. It just works.

Kramnik undeniably pushed Naroditsky into developing an addiction which ultimately contributed to his death. Whether it was intentional suicide or accidental overdose(both are possible) seems to me to be completely irrelevant.

Kramnik engaged in a dedicated bullying and a smear campaign against Danya for a prolonged period, without substantive evidence, and never let up. When Danya explicitly stated in multiple interviews that Kramnik was seriously affecting his mental health, Kramnik persisted in his harassment. When Danya's mental health ultimately led to his death, Kramnik tried to put the blame on everyone but himself, including Danya, his friends and family, even the entire chess community. According to Kramnik, it seems like he is the only one who's completely blameless. It's good that he's being punished, but I think the punishment is too mild. In my opinion he should be stripped of all his titles, including his world titles.

The sad fact is that Kramnik is a pathetic narcissist, who because he's a shadow of his former strength, and a slow old man, gets easily beaten in online blitz, and then accuses people of cheating to protect his wounded ego, because he wants to pretend he hasn't declined.

I actually found going from non-split to split surprising easy, simply because none of my old muscle memory worked anymore, and I had never touch typed up until then, so I wasn't able to go back to the old way out of frustration. A few hours of doing touch typing drills on some free online thing, and I could type at 30wpm, and then it only took about a week of doing my usual coding, IRC chatting etc to get back up to my usual 100wpm.

Also surprising was that after I got there, I could also touch type pretty easily on a normal keyboard. But my old ad hoc 5 finger typing had somehow disappeared entirely.

Ceterizine does cross the BBB to some extent, just less so than a lot of furst gen ones. So it can still have some hypnotic effects, sure.

But that's not the important difference here. The important difference is that ceterizine has negligible antimuscarinic effects, unlike DPH, meclizine, cyclizine et al. Antimuscarinics are nasty drugs, and the antimuscarinic activity(and sometimes other non-histaminergic activity as well) is why a lot of first generation antihistamines are so bad for your brain.