dividedbyzero

I've been trying to find an answer to that, too. The best I've got right now is this:

Moat-like things going for them:

- access to the US government (for huge inflated government contracts as a source of free capital, having the competition banned, having their products legally mandated, getting other preferential treatment, having foreign governments bullied into granting similar advantages)

Too unsustainable, too fragile, too small:

- Strong brands: Brand loyalty seems like a pretty insignificant factor in AI right now

- Semi-exclusive hardware access, RAM buyouts: Unsustainable even in the short term

- Lock-in effects: Their models aren't very stable across versions, or very unique, and switching providers is about as easy as it can be in a highly competitive market.

- Exclusive technology that is important, advanced and hard to replicate to a degree that would materially set them apart and justify potentially much higher cost. Maybe I'm missing something here, but a lot of the claims they make to that end fall apart even on light scrutiny (due to things like test sets that are small, hand-picked and clearly overfitted on, or broad general claims that cannot be substantiated). I'm getting the feeling they don't have anything that materially sets them apart at all, but that's seems impossible to measure in any meaningful way, so who knows, they might have something here.

They'd still have the >99% or so of their customers who prefer Windows or will put up with it for the smartphone-grade battery life.

The tech and some of the infrastructure was, but the cultural phenomenon that people are nostalgic for was not, that sprang up organically.

The skull needs to be encrusted with the specific rare gem that can only be used in temples to the deity of death. That'll be impossible to misunderstand.

And Russia is winning, they already have neutralized if not subjugated the US, but most of Europe is also heading towards that outcome, and fast. Russia is not this near omnipotent 5D chess mastermind behind all of the scenes in all of politics all across the West, but they are terrifyingly good at exploiting existing trends and influencing trajectories just so and they have been at it with great continuity over many decades. It's starting to really pay off. At this point, it's hard to imagine how the next German federal government could be anything else than either a completely dead-on-arrival coalition of polar opposites that only barely hangs on to keep the AfD out of government but will be unable to do anything at all, or an extremely Russia-friendly government the AfD is part of if not leads that is pretty much certain to continue within the EU what Trump started within NATO. Both outcomes serve Russia well, and the establishment has pretty much handed them that opportunity without meaningful resistance.

I can't see that, I don't have an account with the US site. Could you paraphrase?

Could you be more specific? I just skimmed through the reviews on the one I use, seems fine?

yq can do a lot more than just yaml:

> yq is a portable command-line YAML, JSON, XML, CSV, TOML, HCL and properties processor

Giving out control of such things to LLMs is already complete madness, so once the first pleasure bot powered by Grok has dismembered a few thousand users, they'll get sophisticated safety mechanisms.

Though like as not you're still going to be right, after all, Stuxnet happened.

I don't relate to it myself, but I see it in people around me. It appears that's something that may not be limited to but is very prevalent in adults with very high functioning autism.

We're not that far into this new era and I already see people go back to hand-writing the important bits and manually designing their architecture because it turns out that unless you have them spit out yet another boring web app, they aren't very good engineers at all, their errors and bad decisions are very hard to predict and accumulate over time, and the resulting mess becomes unmaintainable at some point, and it appears all the harness engineering in the world can only stave that off somewhat, not prevent it from happening. So perhaps IT won't become a monoculture of agent prompters after all.

The problem is not that subtle, the only somewhat subtle point is that people don't differentiate between the datacenters of yore that were built sort of well and sustainably and the ones built in the current mad gold rush AI construction frenzy. The former aren't very problematic, often pretty good neighbors even, while the latter are nightmare fuel and will purposely not be built to be quiet and safe and stay within available energy capacity and not deplete struggling water tables, and they have the financial and political clout behind them to get away with pretty much anything. Small wonder no one wants them anywhere near them. Traditional DCs might be able to not be a nuisance despite using evaporative cooling, but no one would build that anywhere where there is a shortage of water, and the amount of heat to be dissipated by such a DC is way smaller.

But giving up a local water table might even be seen as an acceptable sacrifice, it's done for agriculture all the time, if only it wasn't done by and for the immense profit of corporations that aim to destroy all entry level jobs and profess to be scared by what havoc their own creation will wreak, all while creating an unprecedented financing bubble.

Like, that's not very subtle at all. You can pick out some detail and try to make it about subtleties, but the larger context is very much not about subtleties.

Upgrades also take time and may not be a good idea seeing how the current AI buildout is financed. There is a certain amount of capacity right now and AI behemoths have the financial muscle to take what's there. If that leaves too little for everyone else, prices will reflect that as soon as existing contracts allow.

That's not always a thing and doesn't help with fires in the rest of the complex. Backup generators, fuel depots, power distribution, large-scale cooling infrastructure all pose fire hazards even when facilities are built with actual regulation. The current AI buildout in the US happens pretty much without any meaningful regulation and gives the impression they'd build unshielded nuclear reactors serviced by children while running if they felt they could get it running cheaper and faster.

The Star Trek computer doesn't feel uncannily human, that would be a good starting point.

If that loud sound is roughly boom-like in shape and texture, and caused by an object moving faster through its medium than the speed of sound of its medium, then they are, in fact, the same thing.

There are 28 000 passenger flights in the US every day, imagine the noise under major corridors if a large percentage of them was supersonic.

> Of course, but if you haven't read them you also shouldn't cite them.

But if you build on them you should have read them. I don't know about the specifics and I don't know if Schmidhuber is out of line or not, and citations and impact factors are a terrible mess, but generally speaking, you are responsible for finding and reading and citing any related work that needs to be cited, and if you work on neural networks in an academic context you probably have been forced to read that particular one at some point. Citation obligations don't just disappear because you don't want to do the research.

They can be pretty great, but not for just servers. Hetzner and others serve that market segment much better.

It's just good business. They know intellectual property is only meaningfully enforced outside China against entities outside China, why wouldn't they use that competitive advantage? I don't buy they are clueless about that, BambuLabs is built for global distribution, they know what they're doing. They may play dumb about the issue (because that's good PR practice), but they'll have decided they can ignore that license and they'll be right in the long run.

I wonder, is there an appetite left to remain a democracy, or more like an appetite for an autocrat who pays a little more attention to the façade and doesn't go out of his way to offend even his most loyal followers?

The model provider would be like a union, at least if unions had absolute control over their members, could take them all away at any time forever with no substantial negative consequences to itself, and spend billions on employer lock-in so switching to the competition is worse than paying the 12% model salary raise.

I suppose if it all works out it'll end up way more expensive than the employees the models displaced ever were. These kinds of technologies usually end up as an oligopoly at best, and those players will have a wide moat by then, and the things these models build will be tweaked such that no other model or human being can realistically work on them anymore, and then they can price gouge everyone to the brink of unprofitability.

I suppose people did learn that it isn't that bad or costly after all and the risk and the bandaids are still better than the cost of being the first to try and fix software supply chains for good. As things stand, I don't know how that might even be done if it's supposed to not be a better bandaid and someone has to do the legwork and it can't be so costly or impractical to overworked IT teams that everyone is just going to ignore it.

I just finished building a Core One+. It has a number of printed parts, but it also does have a bunch of injection molded ones, and they've just replaced another printed one with injection molding. Most of it is metal though, with the printed parts mostly used as relatively simple brackets to hold stuff in place that doesn't need great precision, and replacing those probably wouldn't save much on cost. I think these days they do the printed parts thing mostly to dogfood their print farm solution and I wouldn't be surprised if the next generation had only one or two printed parts for bragging rights. I wasn't a big fan of that either, the Mini I got in 2000 had a few critical parts printed and that did impact performance somewhat, but the Core One+ is fine in that regard.

From a hobbyist perspective, I find it's a much better designed machine than a friend's Bambu that recently broke down and turned out pretty much unfixable. Performance is at least on par, but the entire Prusa can be taken apart with basic hex and torx keys, it's highly serviceable and repairable, lots of fairly standard parts, not very highly integrated. I consider that a feature, but that will cause higher sourcing and assembly costs. It's built like a tank, lots of attention to detail, I expect it to last for a long time with minimal servicing.

That also means it's not targeting the same niche as Bambu's printers. That's not a personal insult, that's just a consequence of how things are right now. No European company is going to undercut a ultra high scale Chinese market dominiation vehicle, that's just not happening. Prusa is doing lots of R&D on much lower sales, they don't have the kind of access to Chinese industry that Bambu has, obviously the Bambu will be cheaper even if Prusa tried to compete in the same segment. But once the market domination thing is far enough along I expect Bambu will disallow non-chipped filament, lock everything into their cloud and jack up their prices. That's how these schemes usually end if they work out, but if they did that now, companies like Prusa would see record sales, so they don't do that just yet.

I'm pretty happy we still have some trace amounts of viable B2C tech industry in Europe. Companies like Prusa provide insane social value too by keeping skills and production in the EU. That's something we sorely need more of (not that companies are to blame, but we still do). Not sure how things will play out, and I'm not too optimistic, but perhaps with everyone else going all-in on dark patterns and pumping out disposable low cost crap, there is an emerging niche for reasonably open high-quality products that serve the owner first and don't data mine them for every last private detail.

But why would I want those directories visible in my home dir?

They don't even really do that IME. If I ask Claude or ChatGPT to generate terraform for non-trivial but by no means obscure or highly unusual setups, they almost invariably hallucinate part of the answer even if a documented solution exists that isn't even that difficult. Maybe vibe coding JavaScript is that much better, or I'm just hopeless at prompting, but I feel a few dozen lines of fairly straightforward terraform config shouldn't require elaborate prompt setups, or I can just save some brain cycles by writing it myself.

That sounds like just what I want, that ought to have great battery life as well. Very promising, if someone were to build a nice light laptop around it...

Not saying I'm not considering it given the current political climate, but I'm spoiled by my Macbook Air. The Thinkpad I've been issued for work costs about the same, runs hot like crazy, always has fans running, is cheap-feeling plastic, thicker, heavier, garbage touchpad, weird keyboard layout (printscreen right next to the arrow keys, what were they thinking?), mushy keys, barely serviceable display ... what do I buy if I want something as sleek and well-built and polished as Apple?

The "eat raw" part seems at least partially misguided, since our ancestors apparently started cooking the heck out of their environment pretty early, didn't consume much unprocessed dairy until very late, and the raw food they did consume tended to carry less pathogens than modern mass-produced food.

The greatest part of the rest, however, appears to be true. I find I'm feeling much better overall, not worse, if I take the bike somewhere even in uncomfortable weather, and it turns out it's more fun as well, more often than not. Low-processed food makes my digestive system measurably happier, walking lots makes me unreasonably healthier, being among trees and mountains calms me to a crazy degree.

But then we did spend like 98% of our evolutionary history since the last big speciation event as hunter-gatherers, and we gotta be as adapted to that as any critter is to their lifestyle.

At this point I kind of expect to find perversions the social patterns and structures of hunter-gatherer groups embedded in the dark patterns that make social media so insidious, much like exploiting our built-in craving for scarce energy-dense nutrition made Coca Cola etc. the economic giants they are. I just don't know enough about the social structures of the deep past to spot these things yet. There doesn't seem to be a lot of literature on that either, so I'm not sure how I'll get there, but I'd like to.