I think it’s better to think about what swap is, and the right answer might well be zero. If you try to allocate memory and there isn’t any available, then existing stuff in memory is transferred to the swap file/partition. This is incredibly slow. If there isn’t enough memory or swap available, then at least one process (one hopes the one that made the unfulfillable request for memory) is killed.

If you ever do start swapping memory to disk, your computer will grind to a halt.

Maybe someone will disagree with me, and if someone does I’m curious why, but unless you’re in some sort of very high memory utilization situation, processes being killed is probably easier to deal with than the huge delays caused by swapping.

Edit: Didn’t notice what community this was. Since it’s a webserver, the answer requires some understanding of utilization. You might want to look into swap files rather than swap partitions, since I’m pretty sure they’re easier to resize as conditions change.

I’m not sure this is a level headed take… They say, when someone tells you who they are, believe them. Meta has already made it very clear who they are; I’m not sure skepticism is really in order.

I’m not a Mastodon expert, but I’m pretty sure you can still get their memes if they reply to you (or @ you), or if they post to a tag you’re following.

Well… They are of course right about the fact that these sorts of decentralized systems don’t have a lot of privacy. It’s necessary to make most everything available to most everyone to be able to keep the system synchronized.

So stuff like Meta being able to profile you based on statistical demographic analysis basically can’t be stopped.

It seems to me, the dangers are more like…

Meta will do the usual rage baiting on its own servers, which means that their upvotes will reflect that, and those posts will be pushed to federated instances. This will almost certainly pollute the system with tons of stupid bullshit, and will basically necessitate defederating.

It’ll bring in a ton of, pardon the word, normies. Facebook became unsavory when your racist uncle started posting terrible memes, and his memes will be pushed to your Mastodon feed. This will basically necessitate defederating.

Your posts will be pushed to Meta servers, which means your racist uncle will start commenting on them. This will basically necessitate defederating.

Then yes there’s EEE danger. Hopefully the Mastodon developers will resist that. On the plus side, if Meta does try to invade Lemmy, I’m pretty confident the Lemmy developers won’t give them the time of day.

Here’s a random interesting car fact. The accelerator pedal only controls how much air makes it to the engine; it opens and closes a flap in the air intake called the throttle body. The car has a sensor that records how much air is coming in, the mass airflow sensor, which is just a wire in the airstream. Electrical resistance in metals is proportional to temperature, and the air rushing by cools the wire. The car’s computer is then programmed to inject fuel according to the estimated amount of air coming in, which is double checked with oxygen sensors in the exhaust (which detect uncombusted air, i.e., too little fuel).

You have to enjoy a hobby in itself, if you’re too focused on results then you’ll have problems with the gulf between your ability and your aspiration. Is there anything you’ve tried doing that you just enjoy doing? Like do you just enjoy banging on a piano or drawing or writing, regardless of the output?

Given how unstable and user unfriendly computers are now, just imagine a future where programmers know even less about what they’re doing.

Glow-in-the-dark heating elements…

At the moment I’m trying https://gpt4all.io/index.html, a big part of the problem is that I just haven’t had a lot of time to deal with it, and the model parameter files are large downloads.

I’ve used it a bit to try and work on my Spanish. That is, using it as a sophisticated chatbot. Unfortunately it’s still quite frustrating for that: I figured I’d ask it to play un juego de rol (a roleplaying game), and it kinda sucks at it. I’m gonna give it a go with an open source alternative, hopefully they’re less aggressively calibrated toward being tedious and awful. It’s just, getting an open source language model running takes a decent amount of time and effort, so I’m sorta midway through that.

Don’t forget: congress checking the power of the supreme court is an unacceptable violation of the principle of checks and balances, at least if you ask John Roberts about it. Luckily he’s totally unbiased on this issue.

It’s older, but The Longest Journey is good. Unfortunately, the final game in the series kinda sucks.

While it’s an ensemble, most people would agree that the main character of Final Fantasy VI is a woman—they just might disagree about which woman is the lead.

I also liked the first Xenosaga game, but again, it’s a series that goes pretty badly downhill.

This is the curation effect: generate lots of chaff, and have humans search for the wheat. Thing is, someone’s already gotten in deep shit for trying to use deep learning for legal filings.

I guess the important thing to understand about spurious output (what gets called “hallucinations”) is that it’s neither a bug nor a feature, it’s just the nature of the program. Deep learning language models are just probabilities of co-occurrence of words; there’s no meaning in that. Deep learning can’t be said to generate “true” or “false” information, or rather, it can’t be meaningfully said to generate information at all.

So then people say that deep learning is helping out in this or that industry. I can tell you that it’s pretty useless in my industry, though people are trying. Knowing a lot about the algorithms behind deep learning, and also knowing how fucking gullible people are, I assume that—if someone tells me deep learning has ended up being useful in some field, they’re either buying the hype or witnessing an odd series of coincidences.

I understand that, but the amount of money that gets fed into political campaigns already generates staggering amounts of spurious text. It’s hard to remember what happened the day before yesterday, but “fake news” originally meant sites that were set up to vaguely look like news sites, all for the purpose of pushing one or two entirely made-up propaganda pieces. Yes, deep learning can partly automate this, but automation isn’t necessary in this case.

There comes a point of diminishing returns with spurious text, and I feel like we’re already past that point.

The only thing deep learning has done is make forgery more accessible. But Stalin was airbrushing unpersons out of photos sixty years ago, so in principle this is nothing new.

When it comes to politics, there’s already enough money floating around that you don’t need deep learning to clog the internet with shit. So personally I’m not expecting anything different.

I owned an Alfa Romeo Spider for a while, which was Italy’s answer to a question nobody asked: where else could I get an MG?

Plus being small, low to the ground, and with even weight distribution, it’s like driving a slow moving cloud.

But isn’t it such a weird coincidence that “apolitical” always happens to be the same as “whatever is best for moneyed interests?” Like being able to take free software and repackage it for sale?

I don’t really like driving, but it is necessary. My (main) car is a 1993 Mazda Miata, which is currently being repainted bright yellow, and I’m gonna put a new top on it next. It isn’t fast, but it handles extremely well and it’s fun to drive. Or at least, it makes driving as fun as it can be.

I think anyone who’s driven a Miata understands.