People don't read a lot of things that they agree to, especially if it's 20,000 words in a postage stamp-sized text box on the new toy they just turned on. However, anybody who's not an idiot realizes that most everything in an agreement that they haven't read is against their interests, and plans accordingly. I personally always look for binding arbitration agreements and make sure to opt out of those whenever possible.

Oddly enough, the Anthropic TOS and privacy policy actually aren't all that obtuse and can be skimmed in a few minutes. They make it clear that they'll be looking for problems with what you upload and what kinds of problems those might be. Like I said, you can easily avoid all that by using a paper notebook for your diary.

If all the negative publicity accomplishes is getting these small towners better deals and maybe some reasonable restrictions on noise and resource use then it's a pretty big win compared to the status quo of fuck you, I'm eating.

That depends. Do you mean a useful LLM vs a Linotronic? Because you can't reasonably make either thing at home out of nothing. You could make either thing out of parts, e.g. fine-tuning a LLM, or assembling a printer out of parts, or you could make a simpler example of either thing at home easily. You can't even afford the power bill to train a useful LLM that can, for example, explain things to you.

I had to tell a co-worker in his mid-20s to put his phone down while pissing in the office bathroom because it was rude.

There just is no common sense. What is so damn important to scroll through for the minute you're standing there pissing in the urinal?

A false positive rate of 10% _among the population of people you arrest_ is acceptable.

I highly disagree. I can understand if Flock's data provided a 10% false positive rate, it's then up the police to do their actual job and gather other evidence to reduce that to near zero before taking any action on that data.

A 10% false arrest rate is criminal. Arrests stick with people even when their is no conviction. Maybe 1 false arrest in 1000 is acceptable in a civilized country without corruption.

I hate when someone gets away with a serious crime, but I'm not going to balance that by lowering the bar for arresting someone.

I'd have to know a lot more before I'd call for conviction, but it's not surprising if arrest was warranted (ha ha ha) because she wasn't researching a book or whatever. It's pretty hard to be non-suspicious about something like that.

I'm not saying stupidity is a crime here, I'm saying it's concerning coupled with the potential threat.

It's especially irritating considering that [IMO obvs] AmiMojo is right and improving Firefox would be a better use of these tokens, so if the guy is being irritating around a Firefox fork, then it's extra ironic — and therefore extra-stupid.

I am not against all AI use, but I do see serious copyright problems, and I also see serious quality problems. I think you need serious chops to address the latter, and who knows where the chips will fall with the former?

If it could reasonably be construed as a threat, then it has to be investigated as a threat to be responsible. Is this a reason why some people might want to run models locally? Yes, and it's a good thing you can do that with any decent game machine, or with many other relatively inexpensive computers. But you can't just run a service and monitor it and then ignore what people are using it for, that's not a thing. If people are going to consume SaaS then they need to be aware that the owners of the platform are going to spend some effort attempting not to be accessories to any crimes.

This is absolutely correct, and it's extremely important to understand that this behavior is baked into the model. It's not even so much that it's "baked into" the model as it IS the model.

Ever since the earliest days of LLMs and generative AI, we have always, always seen what people call "hallucinations." But the only thing that qualifies such output as a hallucination is that we, the human observer, has judged it as such. By construction, the model DOES NOT KNOW THE DIFFERENCE. If it knew or "understood," it wouldn't have produced the spurious output in the first place.

So every now and then, we see some amusing audio/video clip of some human managing to fool or "break" an AI agent, and it makes the rounds of social media for people to point at and laugh and say "tee hee, how funny!" And then these same people, without thinking about what these hallucinations mean and what they imply about the intrinsic nature of the technology, immediately turn around and ask Claude or ChatGPT what they should make for dinner, or who to vote for, or who to hire and fire; or worse, give them access to their personal data and critical systems. That is a level of uncritical thinking and cognitive dissonance that ought to be profoundly distressing.

From a functional perspective, there is quite literally zero distinction between an AI generated pretty woman breaking and suddenly speaking in Cantonese during an English-language interview, and an AI agent that violates constraints imposed upon it and causes damage. They are both "hallucinations" in the sense that they are unintended behaviors/outputs, and as we have already established, the only meaningful arbiter of "intention" is our own human judgment. AI proponents try very hard to make excuses and minimize the severity of the problem. But it is intrinsic to all models.

Whether through laziness or ignorance, when humans cede their authority to interrogate and decide the truth for themselves and let machines do their thinking for them, who or what ultimately becomes the authority for that truth, and who is or is not accountable for the consequences? Who has a vested interest in controlling the source of truth and knowledge, and what are their motives?

(I failed at writing that comment organically)

Yeah, it turns out that actually IS what Claude told him to put into the comment. He's literally using it to write a shitty browser and AND to write shitty comments. What a failed human.

The fact is that she was arrested for making credible threats in the context of a managed and monitored service, proving that she is an idiot at best, and therefore potentially stupid enough to follow through on such a threat. This is what responsibility looks like all around, which is the only reason it's news.

"No was was harmed, no one was threatened, nothing happened.
This is some crazy big brother shit that should not happen"

Are you on the bad drugs? You just said nothing happened, then you said it shouldn't have happened. Well, it didn't. Why aren't you happy?

Yes, they are both bad and addictive.

Never too soon to point out that someone who doesn't respect choices in one area surely doesn't respect them in others.