zlatko

It’s not his tests. It’s the ai writing tests.

Who wrote the tests?

Who will lose money?

Why collaborate with Britain? That’s basically just going back to giving everything to us again, without getting much in return? Talking about that missile program in the article.

Could be both, but that’s a good question. They say memory but mention the space for updates.

I don’t know how that works, is 60k karma good? That’s post karma, not comment karma, right?

Donating hardware that you already own is probably possible, but can be tricky and has security connotations, depending on the arrangement. Unless you’re simply paying for hosting costs for that node and have no operational control at all. There are simpler and better days. If you want to donate, chip in to whatever server you’re on (e.g. lemmy.zip) - send them a few dollars a month, or a one time donation. Or find an instance or a software project (e.g. Lemmy or Gotosocial or Mastodon) and give them those few bucks. Just making the effort is both not that hard, but it’s also great. Even people who don’t have money to spare can try it. I don’t know, find something like a book or a cupboard you’re not using, sell it for a small amount of money, and send it. Of course, lots of people can’t afford to, and that’s why other people are running open servers. That part is super hard not just because of money, but moderation. Running an instance means you commit your time, every day, all the time, to moderate the users, to monitor the software, to fight off DDoS attacks and a million other things. So you can also offer to donate some of your time. Be it moderation or e.g. issue triage or something, it’s usually welcome. So you can look into it. But make sure you can commit - short term “I’ll try it” help is often taking more time to train the volunteer then they’re going to contribute back before giving up. Finally, as someone says, it’s also easy to run your own server. Either buy hosting and operation - I know e.g. masto.host but I’m sure there are others - you pay and they monitor and maintain the software. You just moderate it l like an admin. Invite your buddy there. Some people have the time and skills and operate their own servers, without spending the time for moderation. I’m running a single user instance of Gotosocial. I have control, but I don’t have to deal with moderation. That’s also feasible. Anyway, for your buddy, educate them that it’s not the same as Twitter Reddit or Instagram, but the core ideas are similar. Send them a meme post or an interesting thread from time to time. They’ll want to engage eventually. Upvote or reply. They have to make the account. They’ll then get bored, give up, but that’s fine. Moderate yourself, don’t spam them. If they like it they’ll come back. Or they’ll get frustrated by another ad Block busting scheme by Reddit and come back again. Or not. Which is also fine.

And if they do misdiagnose one in five patients, well, they shouldn’t have gotten infected, it’s their fault.

I would hardly call this dropping a hammer, though. I mean, the article itself says that OpenAI was lobbying for stuff like this. For one, I don’t think it’s addressing any of the most troublesome problems. No mention of environmental impact. No mention of societal impact, or economic one. Nobody is addressing the horrible misuse of LLMs in, say education or biases that are built into the LLMs. For two, what are the consequences? Sam Altman can go around saying whatever the hell he wants, mislead the government, the public, even their own users. No consequences. They can use bait and switch, no problems. The worst thing they can get is a few million dollars of fines. That’s for the long time been just the cost of doing business. It’s a fig leaf law that doesn’t look at any of the worst problems we have but the worst offenders can use it to say “no wait we’re regulated”.

Hell, if you did that, you could make a 100 billion IPO, with today’s market.