• 1 Post
  • 34 Comments
Joined 3 years ago
cake
Cake day: June 30th, 2023

help-circle
  • CEOs are basically just paid to talk. They talk to subordinates to set priorities, they talk to shareholders/investors to attract capital, they talk to media to juice the share price. If your job is to talk all day and you get paid a metric fuckton of money to do it, then naturally you’re going to assume you must be very good at talking and that only the very smartest people can talk as well as you do.

    Then AI comes along, and behold! It spouts endless bullshit better and faster than anyone in the history of the world. Of course CEOs are going to see it as a revolution in intelligence because that’s their kind of “intelligence” — the only kind of intelligence they can conceive of.


  • Whether or not she was technically a Chinese spy honestly seems irrelevant to me… What happened here wasn’t espionage in the conventional sense, it was just a regular criminal conspiracy. Jensen was loudly advocating against export controls on GPUs back in 2022, before he supposedly even met Alice Huang for the first time. He was looking for a way to circumvent export controls and keep revenues growing, and her company had the perfect setup to do it. Who cares if China set up and provided all of the capital for that company? Even a little bit of diligence on the part of Nvidia would have exposed Megaspeed as a front. They knew, and they sold them the GPUs anyway — no spying required.

    The fact Jensen probably fucked her is spicy gossip and certainly a dark smear on his character as a husband and father, but it wasn’t some honeypot. He already wanted to sell the most GPUs possible. The point is that Jensen conspired with her to circumvent US export controls and that he knew what he was doing.





  • Sadly I can’t give any specific technical details without potentially doxxing myself to my employer, but I’ve reviewed 5 or 6 AI-generated PRs in the last year that have contained some pretty severe bugs — the kind of stuff that would have caused serious service outages and security breaches if allowed to go to production.

    Those 5 or 6 were among dozens of PRs, and the amount of code I’m being asked to review is only increasing. Even being diligent, I’m sure it’s only a matter of time before something nasty slips by me, especially now that the process of reviewing code has become so fucking exhausting.





  • Schiff, maybe?

    He’s not as progressive as Bernie, but he’s widely respected by Democrats and has more experience investigating the Trump administration than just about any other Congress critter.

    Schiff would be a good pick if Dems want to demonstrate a serious shift of priorities and how they’re serious about exercising their investigative/oversight mandate rather than more “business as usual”.

    EDIT: While we’re at it, let’s make Raskin house speaker and really get serious about investigating the president.



  • 10 to remove a president after being impeached by the house.

    That’s still a huge longshot. Only 7 Republican senators voted to convict Trump, the first time around and of those, only 2 3 are still in Congress (Romney, Murkowski, Cassidy). The rest of the remaining 44 Republican seats have only gotten more red since then, plus you still have Fetterman who’s spoken out against impeachment in the past.

    EDIT: forgot Romney didn’t run again



  • An under-oath deposition by the alleged victim in which she affirmed Altman sexually abused her starting when she was age 3 and lasting until she was age 12.

    There are also countless former associates of Altman who have gone on record stating that he’s a compulsive liar and narcissist. That’s obviously not direct evidence of anything, but it establishes a clear pattern of untrustworthy and potentially abusive behavior from Altman.

    The civil case was ultimately dismissed because the alleged victim refused to submit to an independent medical examination requested by the Judge… but such refusals are extremely common among people who have been traumatized by sexual abuse and don’t want some stranger poking around their private parts.







  • I agree that they’re lying, but I wouldn’t be surprised if they really do stop training on “live” internet data.

    The lie is in their stated reasoning. This has nothing to do with AI being “too dangerous” to release into the wild. There’s no reason to involve AI in gathering training data anyway… it’s always been possible to scrape the internet quickly and cheaply with non-AI crawlers and bots.

    The real reason is that the internet just doesn’t have enough useful training data anymore. There’s so much AI slop out there that the frontier labs risk model collapse if they continue to train on it. Large swathes of the internet have also become openly hostile to AI, either blocking the traffic outright or redirecting their crawlers into AI tarpits that intentionally waste their time and feed them endless reems of useless garbage. They’ve poisoned the proverbial well for everyone and now they’re afraid to drink the water.