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.

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?

People have been driven to delusion and suicide because of interactions with LLMs, and although you might be able to argue that they would have killed themselves regardless and were in need of treatment, no one can possibly know that, because that's not what actually took place.

But I think the more serious question is the framing of the premise: why is it that the point at which we ought to hold people accountable is if someone is killed? Who decided that the tipping point is when these AI agents start killing people, and why? Not only do most legal systems seek to prevent and punish abusive, antisocial, and outright criminal activity, a just society has a moral and ethical responsibility to apply rules and norms fairly. While we can see that the fair application of the law largely remains an idealistic fantasy and that those with wealth and status often escape accountability, the fact remains that adding AI to that privileged status certainly is not doing anything to help reduce that inequity.

So again I ask: why should agentic AI and their operators be permitted to do things that we would not allow under other circumstances? Just because some people find these tools useful does not excuse the requirement that they be safe and respectful.

If you look at Anthropic's terms of service and privacy policy, it clearly states that they can look at your content in order to enforce their usage policy. You have to agree to this when you set up an account or use their service. Given that this is an AI company, you can bet your content will be screened for violations more efficiently than if humans were involved at every step.

If she really didn't mean for anyone else to view it, she would have scribbled it down in a notebook instead of uploading into a cloud owned and policed by strangers and their robots.

It's a paper copy of underground utility records and a backhoe.

Lights on around the house seems to deter the POS junkies that we no longer keep in jail.

Probably the next code that it outputs will be:

My Walmart is worse than that. They didn't have any price tags at all. Just video screens with AI avatars of a guy in baggy pants sitting cross-legged on a rug.

When I walk by, he says "Look at my cotton swabs! I offer you the best cotton swabs in the land!"

I say: "Well, how much to they cost?"

He squints at me, looks me up and down for a long while, then says: "For you........ I give to you seven hundred and fifty swabs for two pieces of silver only. Believe me when I say you will search this entire kingdom and not find a better deal!"

Please do let us in on the joke and share your explanation of what you find so funny about this phrase.

The conversation only seems unnatural because they didn't use the right avatar.

If you choose the "Max Headroom" avatar, then everything will look and feel exactly as you would expect. Moreover, the dystopian vibe he gives off will be a perfect fit with the current state of the world.

Wrong again.

Moreover, many of the big complex systems I've worked on were big and complex mainly because OOP nitwits had designed them. They often didn't understand how computers actually work, lost sight of the fact that they were writing a computer program, and spewed out frameworks that filled countless gigabytes of resident memory with endlessly deep layers of nothingburger abstractions. Maybe that didn't matter so much when memory was cheap, but that's no longer the case.

If you think that there's only one way to deal with large data sets, you're a just noob that needs more experience before spouting your theories.

C++ is just as much of a sieve as C, only more obtuse.

"Object models" are an anti-feature.

The ideal for memory safety is to have an air-tight pressure vessel.

With C, you're starting that task with a colander. You've managed to fill two of the colander's holes with silicone caulk. Congratulations.

A space weapons arms race plus one million data center satellites.

What could possibly go wrong?

And fission plants definitely aren't cheap.

Especially if you include the unsolved issues such as waste barrels stranded on the beach.

"But but no true Scotsman would put barrels..."

Just shut up.