Thank you for this comment. I have backups I tested on implementation and rummaged through two years ago after a weird corruption issue, but not once since. I still get alerts about them, so I just assume they’re fine, but first thing Monday I’m gonna test them. I feel stupid for not having implemented regular checks already, but will do so now.

Yea, I mentioned in my comment that there was a confluence of issues, but the article does point out that the AI translation made the statement more definitive.

Edit to add:

As part of our post-mortem on this article’s evolution, PCWelt’s executive editor pointed out that the translation makes the article sound more definitive than its native German. He says that in the context of the article, the German word “soll” signals a rumored expectation, but the English translation used “will” instead of something more akin to “is rumored to.”

I thought this was a very well written, transparent article that took accountability as seriously as it should. I am still not sure why people are using AI for translation when translation software already existed. People mention that AI is more context aware, but I feel like when you saw those friction points in old translation software it prompted you to look further into the context, whereas AI will just make an executive decision and people feel like it must be right because it’s AI. I guess it’s possible old language software, or even a translator, would have done the same thing, but I still think people would have less inherent trust in the old software alone. I do want to point out that this AI issue was just a small part of the problem and they addressed plenty of other issues and how they plan to remedy those.

Yes, I saw that, I just didn’t see them say that’s what happened to them. If that’s what happened then this should be an open and shut case. Like I said initially, Google is a bad company doing bad things and this change was an objectively greedy and evil thing.

Google is a bad company with bad policies, but I’d love to have them explain what caused the compromise. They dispute that it was uploaded publicly to GitHub, but don’t seem to provide any information as to what happened. They also didn’t have 2fa on, which is strange to hear because AWS (they’re using Google) required 2fa on all accounts at least a year ago, regardless of permissions if memory serves. Really sorry to hear this happened to them, and the fact you can’t set a hard cap on spend makes Google the party ultimately responsible here, but I’d appreciate having more information on the actual cause.

I get where you’re coming from, but I think it’s important that ars has held this person accountable. They have a journalistic standard they are sticking to, which is that there should be no AI use, and there are repercussions for people who don’t abide. There’s not an extremely large cohort that is willing to spend more to avoid AI, but I am certainly part of it, and seeing ars hold this person accountable helps me know that I can trust and patronize them ethically. There are businesses out there unwilling to acquiesce to an AI first narrative, and I’m just worried that elements of doomerism are going to make people unwilling to believe those companies when they have every reason to believe them.

Extremely sorry to hear that is happening. For what it’s worth, reports like this are not uncommon now:

https://futurism.com/future-society/deloitte-government-ai-hallucinations

https://www.theguardian.com/australia-news/2025/oct/06/deloitte-to-pay-money-back-to-albanese-government-after-using-ai-in-440000-report

There’s an assumption that there has been an increased workload requested of them that I don’t have a reason to believe. That person has been a writer for them for years and since they don’t use AI as a rule, I don’t know why they would have increased expected output from their staff. I’m not saying that never happens, I just don’t believe that’s what happened in this case as there is no evidence to suggest that. I appreciate you explaining that comment though.

What was the damned if you don’t in this scenario? Seems more like damned if do, best if you don’t in this situation.

I don’t work at Ars, and maybe you know something I don’t, but I have seen nothing to suggest that they’re one of the companies doing that. It seems like they are pretty open about how they do not allow AI to be used in the process. Have they said something to indicate otherwise and I just misssed it?

I feel like that may be worse. Kind of like how if you have certain security measures while browsing the web it’s almost easier to fingerprint you. It’ll get a good idea of your age and that’ll be enough rather than sticking to a specific lie. Just always be 3 years older with one additional sibling or a sibling of the opposite sex. If the sex of your sibling is relevant just describe them as a close family friend or close cousin in that instance. I can’t say for sure, but if I had to guess having a static lie is maybe more obfuscation than a variable one. Though even posting on this thread is bad opsec.

AI is suicidal because it was trained on the internet and we’re all depressed here.

Really wish people would get off twitter, but if this tweet from the link is accurate then I think it explains a bit:

I don’t think this is the correct read of the chart, the NYT literally stopped offering a News-only subscription for new customers — the News product is the tentpole feature of all the bundles.

Reddit had a lot of really friendly “femme leaning” communities. Especially the smaller ones. If you were only going to Reddit for nail painting and wedding inspiration it was actually really wholesome. Those communities tended to be 1) very well modded 2) “easy” to mod 3) not fun to troll. There’s a little grey area on if someone is offering good faith critique, but if you’ve commented twice and neither have been positive you lose the privilege to comment. It can create a bit of a hugbox, but it’s much preferred to the opposite.

I really like my experience with the fediverse so far, but I really miss the experience of those positive “femme” spaces. It’s a very different feeling and I haven’t gotten it from the fediverse yet. Not that we’re not empathetic, just that it’s a different space.

Is the idea that they’re somehow reviewing the code on their phone during their commute? Or are they just pushing to prod without even glancing at it? Why bother with the middle man. Just have the AI push it. What a stupid admission.