they should have fired him for inventing Javascript, to be fair

Lmao, how soon people forget that, by the end of the first Trump presidency, there was already talk of starting a massive denaturalization / deportation process.

that way lies madness; the most popular cpu will be the most studied for vulnerabilities

you may as well decide to let the market decide and then...choose the second or third most popular platform

how will people send me word documents containing screenshots of their errors now?

NASA actually planned, a posteriori, a rescue mission; they concluded it was possible although really difficult

https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fspaceflightnow.com%2Fcol...

of course one of the options involves launching _another_ space shuttle to rendezvous with Columbia, which risks either having the same problem as Columbia, or, you know, any kind of problems of both machines colliding

Not sure how many models you can run locally with any utility on a laptop, particularly with how stingy with RAM apple is; base models are still 8gb

I guess they will want to charge premium for this

if you are playing chess or go the domain constraints are the game rules

if you are making a generalist ai, the domain constraints are ultimately physics; that's why the AIs cannot stop hallucinating, there's no way for them to check for correctness, while a chess rule checker is ridic simple

The problem is that IBM is really bad at pushing this technology for some reason; I've worked at places that run OS/400 and the only thing the managers want is to get rid of it all, while at the same time keep it running forever.

So I've seen people build tons of typescript microservices on top of DB2 tables that are also used by ancient, almost untouched RPG/Cobol systems, but these microservices run on AWS or a local kubernetes, but never really utilizing these wonderful technologies you mention.

And in the case one of those old systems need to be changed, they either have an old dinosaur dev come in and do the change, or subcontract a shop that does the same, but that's always a last resort.

So that's why they are reinventing the wheel; nobody wants IBM and people will reinvent or write around them as much as possible, even when they already have the product people need. I don't know what's the deal, since I don't see the contract side, maybe IBM just doesn't promote this, or maybe it's a price problem or a trust issue.

it actually does give them that dictatorship; check what happens whenever a newspaper decides they want Google to pay them for displaying their news in news google com instead of redirecting them to the newspaper site: Google delists the newspaper. Why? Because they don't want to lose that fight

is it legal already?

I don't think it matters whether the NYT tried to prompt the AI to reproduce the text verbatim and that this wouldn't happen normally; the problem is that the AI is capable of doing this _at_all_

GPT4 somehow has in its interior encoded large parts of the archive of the NYTimes and it is using them to generate more text. It's like if I created a program to create-your-own-super-hero, and it included parts of the Spiderman costume that I collage together into new pieces; it does not matter that Marvel had to painstakingly grab all the pieces of spiderman included in my program to recreate Spiderman, it matters that I have a bunch of Marvel comics in my computer program

the detail of how am I encoding these Spiderman bits is not important. Maybe I just store a hundred PNGs and then the program decides how to display them; maybe I store it as weights in a giant matrix a-la GPT4. But those bits are there and they are property of Marvel, or Disney nowadays

the real issue is that GPT4 is laundering these bits and making them into new content they supposedly own

what's more, that these bits of articles are on the web is not an impediment for their ownership by the NYT; afaik openai and the other people working on LLMs are just passing around some massive collection of scrapped files, but that very act is piracy already

the problem with this argument is that there is no guarantee that if someone asks a question on something that the AI learned from the NYT, the AI won't start using paragraphs verbatim, or nearly verbatim

see, for example, what happens in MidJourney when you ask for Italian Videogame; you suddenly receive images of mario bros. Those images are clearly a copyright violation, but the person using the tool did not ask for the AI to violate copyright, it just happened and Midjourney cannot do much other than checking the prompts or the results with _another_ AI to check for those

it's hilarious the SEC is worried whether stablecoins fail or succeed

hilarious but good

they better somehow filter all the spiderman/elsa crap and all the already prevalent generated random bullshit

maybe they can train an AI to flag such content

well the summary says this is an opening to relitigate section 230, so maybe the whole thing is an excuse