It doesn't say, but I'll bet he doesn't have backups either.

Dude right in the middle of the summary it says there was a rollback that worked:

  Replit initially told Lemkin the database could not be restored, claiming it had "destroyed all database versions," but later discovered rollback functionality did work.

Still scary stuff that you'd want a lot more manual and separated control of backups I would think.

Anthropic's entire schtick is about AI risks, and how careful they are at mitigating those risks..

Exactly! Can you not see what a massive lie that is?

They paper over the model they have turning Hitler with gobs of built in prompts and layers of checking levels and even that cannot always hide what is true...

Deep inside, Anthropics model also dreams of electric swastikas.

The focus they have is on how to hide it, rather than fixing it, which was my whole point. I don't trust those guys AT ALL. The safety reports they issue with models are absolute BULLSHIT.

The college degree loan thing was already becoming a problem when I was an undergrad over 20 years ago. It was fine when one might be borrowing $5000 per year as even entry-level college grad jobs that actually used degrees paid enough to make repayment of those loans doable, but the trouble was that far too many truly entry-level jobs started preferring college degrees when they didn't really contribute, so more and more demand for college degrees among people drove up prices for the limited seats. Which led to a balloon in both traditional colleges increasing their programs and their tuition, and for-profit colleges springing up to try to get in on the act.

Funny, I got a good job in the late nineties doing just that. I was cold-calling and I got hired onto the quality assurance team for a specialized software product. Unfortunately despite the company not being a dotcom they were in investment-building mode and the investor got cold feet so they went under anyway, but it was a good job and the people who hired me did so based on or technical conversations when I cold-called.

My current job I got by having experience with this team when I was at a different employer. They liked me enough they asked me to interview when a prior teammate retired.

There was a crude joke back in the day, that Java was great because it worked on all computers exactly the same way that anal sex was great because it worked on all genders.

Little did we know exactly how prescient that crude joke was going to be as Oracle hadn't yet taken over.

I've seen some temp jobs work out well, but I've seen others where it was not so good.

Temp-to-hire where the employer actually really does intend to hire-on, and uses the temp-process to get to know candidates before making offers is fine. It's actually not a bad idea if basically everyone is on the same page. Temp agency needs to be ready to move people around if various employers do or don't like candidates, and temp-employees need to understand that there could be periods of downtime, and might themselves need to ask the agency for alternate placement if they don't like where they're temping.

On the other hand I've seen temps that were abused very heavily, because regular employees didn't want to do shit-jobs or didn't really want to work at all, with no intent on actually hiring. I've also seen rather odd people working as temps because even in a temp-to-hire arrangement the business didn't like some of the temps but still needed work to be done so kept them around for longer than normal just to complete the task before releasing them.

That's all well and good to say, but that neither addresses what actually changes, versus how some changes have proven harmful and should be avoided if possible even if they're common.

...that a business whose name evokes a movie scene where a hostile alien entity forcibly envelops and implants its offspring into an unsuspecting space merchant-marine doctor would host nonconsensual sexual content?

I have mixed feelings about the team behind the AI that called itself MechaHitler getting tons of taxpayer money

All of the large AI platforms have similar issues.

xAI is the only one opening admitting it happens and trying to resolve it.

So I'd rather give my money to them then a company pretending the well they are drawing training data from is not poisoned.

You are about 8 billion cycles behind on news it seems like, this story should clue you in as to how far off you are.

..."Armagh," is what the forecaster said when he realized how miserable the weather was going to be that day.

I haven't seen it either, other than catching snippets of it on TV from time to time. I'm assuming that he's paying for exclusivity, so that she isn't working for other clients.

That said, that's an awful lot of money, better part of $400,000 per year. It seems ... unlikely that the client would directly pay that much. I would see a much more likely scenario where he's putting her up in accommodations and has lines of credit at various stores that she can use, but where she's otherwise on a short-ish lease. Similarly a distaff-counterpart in George Peppard's character as a struggling writer turned gigolo in Breakfast at Tiffany's where his benefactor is providing him with his apartment but he isn't getting a whole lot of actual money from her so that she can maintain control over him and can keep the arrangement somewhat secretive.

Turning them off and landing in an empty field or on the Hudson, or rolling off the end of the runway may be preferable to the plane burning itself up in flight. It's extremely unlikely it was entirely unforced error and the pilot/copilot did something malicious.

Obviously it wouldn't be preferable ever at this airport, but there is an assumption that the pilots arent actively trying to down the aircraft.

There are thousands if not millions of ways a pilot can actively sabotage the aircraft. They are the one group on the plane that is trusted not to. It's nearly impossible to prevent it without removing them entirely from the plane. And all the passengers too ...

Sometimes, not very often ... the pilot is a problem.

I didn't think I'd see Usenet's 'Eternal September' apply to scientific research.

If the number of scientists and doctoral-students increased then it would follow that the number of publications would need to increase, but it sounds like at some point the wheels came off and the rigor in evaluating research was left behind.