What a terrible fucking headline. Even if they kept a record of every single interaction on hand for the chatbot to reference (which is preposterous), the notion that it could reiterate it verbatim to a degree that is sufficient to hold up as evidence of anything is ludicrous. It's also wildly inaccurate to the actual story. Bad job all around.
Actual AI would be more than "just math", but LLMs aren't AI, so the comparison is moot.
Now we’ve built a collection of simulated neurons, at a scale close to that of the human brain, and trained it on the entirety of the human language
We are not even close to anything of the sort. We've got a probability machine that's mostly decent at previous collections of human language. The other two are much farther down the road (if they're even possible) than you or the rest of the tech bros are trying to convince everyone else of.
Though commonly reported, Google doesn't consider it a security problem when models make things up
To be clear, all llms "make things up" with every use - that's their singular function. We need to stop imparting any level of sentience or knowledge onto these programs. At best, it's a waste of time. At worst, it will get somebody killed.
Also, querying the program on why it fabricated something as if it won't fabricate that answer as well is peak ignorance. "Surely it will output factual information this time!"
Should start playing into it. Nothing would piss her boss off more than people treating someone else as his equal. Would probably fire her after the first news headline.
That's exactly it. LLMs and their image counterparts have no innate or burgeoning knowledge as people tend to assume. Their singular, core function is to generate their output from literal random noise, like the static you used to see on TV. So the response to the same question will change because the random noise changed, not because the algorithm learned or reconsidered anything. And if you used the same noise, the answer would be identical. No knowledgeable or self-sufficient AI will ever evolve from that.
Even humans who are driving normally and sober can be held responsible if the law decides that they were at fault. Fatal accidents are investigated by people whose job is specifically to investigate vehicle accidents.
That's a good question, but it's important to keep in mind that a lot of public judgement here is based off a single headline, not a thorough investigation. The world is messy, and even the best technology will get into accidents due to any number of causes, including children dashing into the road.
In the case where the technology in the car was found to be at fault, I agree that there needs to be a responsible party, but it's not exactly straightforward. Punishing the driver seems counterintuitive, but may be useful in getting owners to be diligent in selecting safer vehicles. And punishing the company is complicated - if the punishment is just a fine, it will simply be rolled up into their expected expenses. If it's too damaging, it will chill development of automated driving entirely. I'm not really sure what the best solution is.
It's like the text predictor on your phone. If you just keep hitting the next suggested word, you'll usually end up in a loop at some point. Same thing here, though admittedly much more advanced.
Your experience with people calling out Democrats in your first post may have given you the wrong impression - Lemmy doesn't have much of a right wing problem, but it does have an overabundance of leftists. You'll hear a lot of similar "fuck liberals" comments from them, but from the other end of the spectrum. You'll also find a lot of tankies here - leftists that unfailingly support violent, authoritarian regimes, mostly because they're a country other then the US or EU.
In short, you'll find the political inclinations of users here to be (possibly confusingly) diverse, if you stick around.
What a terrible fucking headline. Even if they kept a record of every single interaction on hand for the chatbot to reference (which is preposterous), the notion that it could reiterate it verbatim to a degree that is sufficient to hold up as evidence of anything is ludicrous. It's also wildly inaccurate to the actual story. Bad job all around.