Meta staff torrented nearly 82TB of pirated books for AI training.

[El Heggunte]

There is bad piracy and there is good piracy alright.

Meta staff torrented nearly 82TB of pirated books for AI training — court records reveal copyright violations.

https://www.tomshardware.com/tech-industry/artificial-intelligence/meta-staff-torrente...ght-violations

[usually_quiet]

They are probably not the only ones guilty of this sort of thing. ...but there is little or no chance that any of these multi-billion dollar tech companies will ultimately be held to account by the USA at present. Unfortunately, any discussion of the biggest reason why this is the case cannot be discussed in this forum.

[noemi7]

Most if not all such training is a copyright violation at it crawls through Internet and grabs data which doesn't belong to tech companies. The recent shitstorm about DeepSeek has clearly demonstrated this.

[thecoalman]

Forums are big targets. You can have thousands of requests in short period coming from botnet and it can go on for hours. It's bunch of random IP's with browser user agents so difficult to stop. It's not the intent but it can effectively become a DDOS attack. Best solution is Cloudflare.

[lordsmurf]

Forums are big targets.

Which is also why most AI is so completely ignorant and clueless, constantly barfing out BS wrong answers. The general population is not smart, and a huge % of them believe in myth/misinformation. Us real people often understand who among us is a clueless moron, but AI doesn't have that ability. Very obviously so.

Certain topics, especially video topics, give out some of the most ridiculous answers. Why? The information did not come from reference-grade sources. For example, you can discuss stock details with AI, and it's very accurate, as it tends to source from SEC documents. But if you ask it about the best VCRs to use for VHS transfer, it gives nonsense answers (including not just crap VCRs, but DVD players, and with no concept of PAL vs. NTSC). That's because it pulled information from places like Reddit, but a lot of those people pulled the information out of their ass. So it's just telephone gamed BS, no basis in facts.

Just yesterday, I was searching Google for info on a rare 1990s JVC VCR, and the AI answer stated that it was a Hitachi DVD player from the 1980s, and it hallucinated BS about "most people didn't know that DVD existed back then" (WRONG!).

[thecoalman]

As I often point out to people AI is not infallible and can be wrong. As it consumes more and more data the accuracy and knowledge increases.

I've tossed it a bunch of questions related to burning coal and it's been fairly accurate with it's general information. For example the use of barometric and manual pipe dampers. However I just noticed it gives positive answer for using barometric damper on wood stove which is a big safety hazard.