It seems unlikely that it’s all that mysterious

OpenAI/Microsoft, Amazon, Meta, and Google

There you go I solved the mystery

Many of their talented engineers have moved on to other companies, some new startups and some already-established ones.

When did this happen? I know some of the leadership departed but I hadn’t heard of it from the rank and file.

I’m not saying necessarily that you’re wrong; definitely it seems like something has changed between the days of GPT-3 and GPT-4 up until the present day. I just hadn’t heard of it.

There are a lot of folks in tech who really just want to build neat things and it feels oppressive to be in a company that’s likely to lock away the things they build if they turn out to be too neat.

I’m not sure this is true for AI. Some of the people who are most worried about AI safety are the AI engineers. I have some impression that OpenAI’s safety focus was why so many people liked working for them, back when they were doing groundbreaking work.

If they closed down, and the people still aligned with safety had to take up the mantle, that would be fine.

If they got desperate for money and started looking for people they could sell their soul to (more than they have already) in exchange for keeping the doors open, that could potentially be pretty fuckin bad.

They don’t care if the cat videos are on a banana sub.

I don’t know why but I can’t stop laughing at this

Your Lemmy posts are already being scraped for AI

The level of effort it would take to prevent would be infeasible to ask of even a non volunteer admin let alone a volunteer let alone literally all of them

The cycle continues:

• Hey you guys can have everything for free
• WTF this is expensive to provide, I think I’m gonna start taking advantage of you guys which someone will pay me to do
• WTF where’s everyone going
• WTF I’m still losing money and always have been
• Screw you guys, screw everybody, I didn’t want y’all anyway
• (fades into irrelevance, gets bought by someone and stripped for parts)

Idk it’s not as pithy as Cory Doctorow’s version I guess

Anyway we’re at step 5 at this point

Oh, it might be that. That actually makes more sense.

I just know that I always heard it as so you wouldn’t electrocute yourself, and that the first time I opened up a monitor for something and actually found one, I was beyond delighted

I, too, miss the days when you could write literally anything in XF86Config and that’s the signal it would send to your monitor. There was a warning in the docs that you could easily fry your monitor by sending a signal that it couldn’t handle that would cause physical damage so please be careful.

Also, the good monitors came with an all-plastic screwdriver attached on the inside of the case, so that you would have one available that you couldn’t electrocute yourself with on the big capacitor since at that point you’d already revealed that you planned to open the thing up and start fuckin with it.

It was wonderful days

Honestly? If you’re still on Twitter at this point, that’s on you.

It’s just looting the corpse at this point.

Mining the goodness that used to be there happened long ago; it got so bad that they killed the patient about a decade ago, and now his decay is getting advanced enough that they know they need to start grabbing his jewelry and hunting through his wallet for any credit cards that haven’t been maxed yet, because it won’t fool anyone to keep toting him around town pretending he’s still the mayor. It was fun while it lasted though.

The amounts of energy we use is never ever going to go down. It just isn’t.

Not voluntarily maybe. But that’s not the only way. The only outcomes of any “realistic” course correction to the current state of the world and human behavior include widespread death and societal collapse once uncontrolled climate change takes hold for real, and at that point, it’ll go down quite a bit.

Yeah. People already sell laptops; this is basically a super expensive laptop with a fancy screen and a janky custom OS. But having this as an app for your phone, that let you pop other apps up into the heads-up virtual display or have “full screen” access to certain functionality while still supporting all your regular stuff, would be pretty different. So it can make your phone “laptop like” any time you wanted to pop the glasses on, or pop little notifications into the corner of your vision, maybe with a couple of little buttons on the glasses for “expand notification” “clear notification” “clear all” “up” “down” “minimize” “maximize”, something like that, would be super neat. And then any time you want to break out the keyboard you can use it like a computer.

(I know the permissions and app compatibility and battery life etc would make that not necessarily trivial to do)

1. “Oh, I see the problem. In order to correct (what went wrong with the last implementation), we can (complete code re-implementation which also doesn’t work)”
2. Goto 1

It’s incredibly useful for learning. ChatGPT was what taught me to unlearn, essentially, writing C in every language, and how to write idiomatic Python and JavaScript.

It is very good for boilerplate code or fleshing out a big module without you having to do the typing. My experience was just like yours; once you’re past a certain (not real high) level of complexity you’re looking at multiple rounds of improvement or else just doing it yourself.

ARE YOU FUCKING KIDDING ME

They took the guy who led the legendary team that made the search not only work instantly at a previously unimaginable scale, but also freakishly well as far as “finding exactly what you want based on almost any query,” back in the late 2000s, if you remember… that guy, when he started pushing back against the people who wanted to fuck up search results to boost imaginary metrics that were theoretically (and, probably, not really) going to make more money from ads, they pushed him out.

This absolutely excellent article goes into detail about the exact moment, if you had to pick one, when Google stopped being a legendary tech company and simply became yet another behemoth coasting on its past successes until the market changes under it and it can’t adapt, fades, and takes its place with all the others, all the way back to IBM and DEC. Nothing’s changed in a big enough way for it to get knocked back into that obscurity yet, but it clearly will at some point.

So, I saw this story and I typed a comment about how it was pretty much guaranteed (given Musk’s cutting of the engineering department and the scale of Twitter’s operation) that this would cause some slight amount of breakage for the forseeable future, and the unfixable and unflattering nature of the ensuing jank would be the nail in the coffin for Twitter (which for some reason still is home to a lot of journalists and primary sources and etc even to this day in its wrecked-up form).

Then I thought, you know what, I don’t actually know that that’s how it’ll happen, and deleted the comment and moved on with my day.

And then just now I just tried to click on a Twitter link, and saw a black page with this:

Something went wrong, but don’t fret — let’s give it another shot.

(Button: “Try Again”)

⚠️ Firefox’s Enhanced Tracking Protection (Strict Mode) is known to cause issues on x.com

Oh shit, it must be Firefox’s fault! Yeah, must be causing issues. My bad man, you’re right; I guess I will need to switch browsers now so I can have the privilege of using Twitter.

Still though, OpenAI and Google, I think, have a legitimate argument that LLMs without limitation may be socially harmful.

I actually do agree with this. Because of the massive potential for harm if it goes wrong, AI development is one of very, very few types of technology where it does actually make some sense to try restrict its development to the big entities so that you can have some vaguely-realistic hope of placing regulation on it so that it can be developed safely and responsibly. (Whether the big entities will develop it safely and responsibly is a separate, though related, issue.)

That said, good fuckin luck, the genie’s pretty much out at this point.

for now

(Mistral is actually open source (code and models), which is very much not true of GPT or Bard.)

(In this case the “wrench” is just breaking into some weak link that isn’t ssh, once your password is strong to not be a weak link)