It's hilarious to me that the owners of these "frontier" LLMs complain about others "illicitly" training on their models, when their models were trained mostly illicitly on the work of others.

You know it's going to happen - someone in charge of replacing/stripping all the names and replacing them with numbers, or whatever, is going to fail at some level. A bunch of PII will go into Gemini, and they won't be able to sue Spirit, because they won't exist, and Google will definitely make sure their contract excuses them from any liability.

I suppose it's too much to hope that the people in charge of watched this insightful series.

I'm sure it's a coincidence and not because Mark told some other C-Suite execs that he did this. lol

Does anyone here know anyone, including themselves, who goes to Polymarket "for information"? I have not, and do not, but I've no idea about the broader ecosystem of data nerds or news junkies.

...a business world dominated by Macs. Oh, wait. Sorry - perfect way to prepare students for a consumerist market dominated by Apple. We know which lesson is more important to those in charge now.

This sort of result from people doing the actual work of reporting is why I now support outlets like 404 Media and Popular Information instead of WaPo and NYT.

If I turn a book into my publisher that reads like a Stephen King rip off, they aren't going to publish it with a disclaimer that says "No really, this isn't a rip-off, it just utilizes similar themes and settings!" Gen Ai can protest all it likes, but saying you aren't ripping someone off doesn't mean you aren't. The only way it can actually not rip off an author it trained on would be to remove all author names from the meta-data *before* training, which they'd never do because then they couldn't sell their plagiarism machine to plagiarizers.

Indeed. The only way to determine whether you can fly is to believe the guy who sells you a pair of wings and drives you to the top of a cliff.

There is no way in hell an Ai robot will be able to interact meaningfully with a classroom of kids under 12 - the youngest ones will act like little kids, constantly interrupting and talking over each other and confusing it because they are excited and have little impulse control. Older kids will mistreat it or try to mess it up because it has zero authority. Why spend $57k a year on another teacher, or on two part-time teaching assistants when you can spend that once on a POS that some sales-idiot convinced an administrator would work?

I think the Ai angle might be an unwelcome distraction from the case that could hurt the employees. If Meta can show they didn't actually use Ai to make any decisions, but merely to assemble statistics that they could have gathered via traditional means, then the case loses credibility. If the employees can actually show that the layoffs disproportionately affected workers engaged in protected activities and those with disabilities, they'd probably have a good enough case, especially if those decisions were made by cold, calculating humans.

Once again, the non-creatives like Belsky completely fail to understand why creatives (and those sympathetic) are upset about Ai. It's not because the prompt-generated garbage is garbage. And if he thinks CEOs and bean counters won't push to use the Ai to save money by cutting staff and creative budgets, he's delusional or willfully stupid. First they came for the storyboard artists, and I said nothing, because who the F*ck cares about storyboards?

Genuine question - I'm not opposed to the idea in principle but I don't know how it would work. I don't really know how sovereign tax funds work, but the press release mentions "natural resources". Ai companies are not a "natural resource" so I don't know how current US tax law would even allow companies to pay with taxes with stock. They aren't even a "natural monopoly" like telephones and utilities used to be, so that argument doesn't really work, but IINAL. It also sounds like it would be easily challenged under the 5th amendment, and the current court would likely rule against the law as an illegal taking. And what about a private company? OpenAI hasn't had their IPO yet, so couldn't they just find some kind of workaround to avoid this?

Hopefully all the new Ai features can be turned off on my Pixel 9, or this might be the first update I pass on.

It's interesting how expectations of privacy can change over time with the technology. I'm pretty sure that before cell phones you couldn't get a land-line installed without the phone company having all your details - especially a physical address - that the government could get pretty easily from them. But then you also had pay phones in many places, which were the old timey version of a burner. Bring back phone booths! lol