Skip Navigation

Posts
1
Comments
315
Joined
3 yr. ago

  • Deleted

    Permanently Deleted

    Jump
  • But all the interesting people are in the computer, the same place as the bad stuff is.

  • There's a pretty good reason to think it's not going to improve much. The size of models and amount of compute and training data required to create them is increasing much faster than their performance is increasing, and they're already putting serious strain on the world's ability to build and power computers, and the world's ability to get human-written text into training sets (hence why so many sites are having to deploy things like Anubis to keep themselves functioning). The levers AI companies have access to are already pulled as far as they can go, and so the slowing of improvement can only increase, and the returns can only diminish faster.

  • If LLMs aren't going to reach a point where they outperform a junior developer who needs too much micromanaging to be a net gain to productivity, then AI's not going to be a net gain to productivity, and the only productive way to use it is to fight its adoption, much like the only way to productively use keyboards that had a bunch of the letters missing would be to refuse to use them. It's not worth worrying about obsolescence until such a time as there's some evidence that they're likely to be better, just like how it wasn't worth worrying about obsolescence yet when neural nets were being worked on in the 80s.

  • Usually, having to wrangle a junior developer takes a senior more time than doing the junior's job themselves. The problem grows the more juniors they're responsible for, so having LLMs stimulate a fleet of junior developers will be a massive time sink and not faster than doing everything themselves. With real juniors, though, this can still be worthwhile, as eventually they'll learn, and then require much less supervision and become a net positive. LLMs do not learn once they're deployed, though, so the only way they get better is if a cleverer model is created that can stimulate a mid-level developer, and so far, the diminishing returns of progressively larger and larger models makes it seem pretty likely that something based on LLMs won't be enough.

  • Enzymes are specific to a particular molecule, or class of molecules with a particular pattern. A PEI buildplate is not getting eaten by the proteases in a dishwasher tablet. The reasons you're not supposed to rinse things before putting them in the dishwasher are:

    • most dishwashers have sensors to detect how much material is ending up in the water, and if things have been rinsed, it can mislead them into thinking the load is lighter than it really is.
    • dishwashers replace some of the dirty water part way through the load, and the enzymes are more soluble than the dirt, so if there's not much food residue for them to stick to, they can end up getting rinsed away part way through the cycle.
    • it uses water and your time to rinse the dishes first, which is a waste if it doesn't make them end up any cleaner.
  • I think it was pretty reasonable of them to worry - lots of people who don't like spending unnecessary money also don't like spending not-obviously-necessary money on safety equipment, and there's plenty of material on the internet that would imply resin printing is completely safe as long as you don't drink the stuff. Resin printing with woefully inadequate ventilation/PPE is really common, so it's a pretty safe bet that anyone asking questions is probably also doing something unsafe without realising it, especially as resin not liking the cold is something a lot of people learn about fairly early on (unless they live somewhere where it never gets below 20°C).

  • To be fair, if I had all that money, I'd probably just pay someone to figure out how to make it do the most good, and continue spending at least some of my time shitposting. It's okay to have hobbies, but it's bad to hoard the money or invest it in evil.

  • It's pretty easy to put something on the box like this can make your phone buzz if you forget to brush your teeth, and people who worry they're sometimes forgetting to brush your teeth will see that as an advantage without necessarily realising that they need to give the manufacturer their email and the right to associate it with their brushing telemetry.

  • There are a far fewer pedestrians and walls and lamp posts and motorcycles in the air than on the ground, though, so there's a lot more margin to be awful without endangering anyone other than your own family.

  • CUDA is an Nvidia technology and they've gone out of their way to make it difficult for a competitor to come up with a compatible implementation. With cross-vendor alternatives like OpenCL and compute shaders, they've not put resources into achieving performance parity, so if you write something in both CUDA and OpenCL, and run them both on an Nvidia card, the CUDA-based implementation will go way faster. Most projects prioritise the need to go fast above the need to work on hardware from more than one vendor. Fifteen years ago, an OpenCL-based compute application would run faster on an AMD card than a CUDA-based one would run on an Nvidia card, even if the Nvidia card was a chunk faster in gaming, so it's not that CUDA's inherently loads faster. That didn't give AMD a huge advantage in market share as not very much was going on that cared significantly about GPU compute.

    Also, Nvidia have put a lot of resources over the last fifteen years into adding CUDA support to other people's projects, so when things did start springing up that needed GPU compute, a lot of them already worked on Nvidia cards.

  • Generally, you'll get better results by spending half as much on GPUs twice as often. Games generally aren't made expecting all their players to have a current-gen top-of-the-line card, so you don't benefit much from having a top-of-the-line card at first, and then a couple of generations later, usually there's a card that outperforms the previous top-of-the-line card that costs half as much as it did, so you end up with a better card in the long run.

  • Deleted

    Permanently Deleted

    Jump
  • I've found this is really dependent on placement. If I put my libre a couple of centimeters away from the region I usually use, it'll read low all night, but as long as I stick to the zone I've determined to be fine, it'll agree with a blood test even if I've had pressure on it for ages. Also, the 3 is more forgiving than the 1 or 2 because it's smaller than the older models, so affects how much the skin bends and squishes less.

  • It wouldn't be Rocko's Baselisk if it didn't do things that hurt.

  • Removed

    bri'off

    Jump
  • It's rare that English children who learn Spanish as the first foreign language that they're exposed to. If their parents are immigrants, then it'll likely be their parents' mother tongue(s), and if they're not, they'll likely be taught some French before any Spanish. That can then lead to a habit of saying any foreign word with a French accent.

    Also, England has strong regional variations in accent, so you might be hearing people say exactly the same vowel sounds as they'd use when speaking English, but those vowel sounds might be totally different to how you're expecting that they'd speak English.

  • Deleted

    Permanently Deleted

    Jump
  • It'll heat up the firebox, which is exactly what the firebox wants to happen. It's not like we're using precisely-timed explosives to briefly make the mass much more than critical and counter its desire to blow itself apart for long enough that it blows other things apart, too.

  • Deleted

    Permanently Deleted

    Jump
  • You can boost it by hollowing out the middle and filling it with tritium, but plutonium is dense, so 80 tons will probably fit in the firebox just fine.

  • I've printed kilos and kilos of Geeetech PLA, and that's some of the cheapest on AliExpress (although I used to get it directly from their website before I realised doing that was normally more expensive). It arrives wet, but other than that, there's (nearly) nothing to complain about (although years ago, I had a roll with two lumps of grit in it that caused clogs). I've had mixed success with their other materials - their ABS+ started burning in the nozzle while still being cold enough that layer adhesion was bad and their high-speed PLA has ridiculous oozing that causes ridiculous stringing, but their PETG and TPU seem fine. I'm pretty confident that their basic materials are absolutely worth £7/kg.

  • You get both sizes of gametes with all kinds of bodies. It's only the testes/ovaries that are reliably correlated with gamete size, and anything further away from their production than that has about the same chance of not being the style you'd expect as an atom has of not being hydrogen or helium, just like the original meme alludes to.

  • It'll be maintained for a while, so we might get to 3.14.15 or 3.14.16, which will be a better approximation and better because of more bugfixes.