jjjalljs @ jjjalljs @ttrpg.network

Posts

1

Comments

2,818

Joined

2 yr. ago

Many people have found that using LLMs for coding is a net negative. You end up with sloppy, vulnerable, code that you don't understand. I'm not sure if there have been any rigorous studies about it yet, but it seems very plausible. LLMs are prone to hallucinating, so you're going to get it telling you to import libraries that don't exist, or use parts of the standard library that don't exist.

It also opens up a whole new security threat vector of squatting. If LLMs routinely try to install a library from pypi that doesn't exist, you can create that library and have it do whatever you want. Vibe coders will then run it, and that's game over for them.

So yeah, you could "rigorously check" it but a. all of us are lazy and aren't going to do that routinely (like, have you used snapshot tests?), b. it's going to anchor you around whatever it produced, making it harder to think about other approaches, and c. it's often slower overall than just doing a good job from the start.

I imagine there are similar problems with analyzing large amounts of text. It doesn't really understand anything. To verify it's correct, you would have to read the whole thing yourself anyway.

There are probably specialized use cases that are good- I'm told AI is useful for like protein folding and cancer detection- but that still has experts (I hope) looking at the results.

To your point, I think people are trying to use these LLMs for things with definite answers, too. Like if I go to google and type in "largest state in the US" it uses AI. This is not a good use case.

That's really not the same thing at all.

For one, no one knows what the weather will be like tomorrow. We have sophisticated models that do their best. We know the capital of New Jersey. We don't need a guessing machine to tell us that.

You shouldn't trust anything the LLM tells you though, because it's a guessing machine. It is not credible. Maybe if you're just using it for translation into your native language? I'm not sure if it's good at that.

If you have access to the internet, there are many resources available that are more credible. Many of them free.

You don't need AI for people to learn. I'm not sure what's left of your point without that assertion.

Oh, I didn't know about this one. Thanks for the link. Adding it to my mental dossier on police bullshit.

Oh my gosh I didn't realize the tweet was from Donald jr. I thought it was just some random internet idiot. This is worse somehow.

Other people have some really good responses in here.

I'm going to echo that AI is highlighting the problems of capitalism. The ownership class wants to fire a bunch of people and replace them with AI, and keep all that profit for themselves. Not good.

I really don't think creating for real artificial intelligence is a good idea. I mean that's peak "don't invent the torment Nexus"

Are you going to give it equal rights? How is voting going to work when the AI can create an arbitrary number of itself and vote as a bloc?

Creating an intelligent being to be your slave is fucked up, too.

Just... We don't need that right now. We have other more pressing problems with fewer ethical land mines

Had me eating the onion until the third paragraph

I mean, technically true in a "picture a spherical frictionless cow" kind of sense, but not something you're going to find in real life. Most people are susceptible to corruption. Cops have a lot of power, and we all know the saying about power corrupting.

On top of that, I'm pretty sure much of the training they're getting to be cops is bad. They don't spend nearly enough time on deescalation, I think , for example.

The new daredevil tv show goes into this

I'm pretty sure anyone who becomes a cop to "reform it from the inside" will either be corrupted or pushed out. There've been a couple high profile cases of cops trying to reform, and they were rewarded with prizes like being shot in the head or having dead rats left in their mailboxes.

I don't think I know anyone who's cheated. But I also hang out with a lot of polyamorous people, where cheating is possible but I feel like is less common.

Huh. I don't think I know anyone who cheats so I guess I got lucky. Your post is plausible

Is there a name for the thing where you'll make an argument with like 3 distinct points supporting it, and the other person will attack only one, and claim the whole thing is in their favor?

Like, "You can't cast two leveled spells in a turn, and you're silenced, and you're out of spell slots, so you can't cast another fireball"

"No, I have another spell slot from my ring. Fireball time!"

I mean, it's win-win for the assholes.

People don't fight back: they get deported. Win for nazis

People fight back: Nazis use it as justification for further escalation

At least defending yourself / the victims has a chance of saving them. Doing nothing won't

I think a lot of people don't get good information and then don't connect the dots when they do.

There's a lot of slop and mis/disinformation out there

• crawl stone soup. A classic rogue like.
• elden ring + dlc. Big masterpiece of the genre.
• the binding of Isaac (+ all the dlc). Huge rogue lite. Lots of stuff I haven't unlocked yet.
• monster hunter (maybe world? I liked rise too though).
• if I could have online, guild wars 2. Otherwise, maybe the original doom. Especially if it comes with a map editor, fan made maps, or Oblige to randomly make maps.

I bought a couple games on epic when they were cheaper. I don't think I'd do so again.

• the client isn't as good. It's slower, the way it paginates your games (I got a lot of free ones) is annoying. It really wants to show you store stuff
• less (zero?) Linux support
• don't think it does the game recording steam does
• I don't think it has the remote play together steam does

There's probably other stuff I'm not thinking of. It's just not as good a service.

If this joke was a person, it would be in 2nd grade and know how to read and write.