Yet again, mushing together Sci-fi esque philosophising (wish is in on itself is great, don't get me wrong, love our stories about AI) with LLM's, wich are, a great piece of technology, but far from being AI in the sense that sci-fi works mean it.
The way I separate oponion from marketing wich is, admitedly not the best method, is to look if it is written by a random guy on a blog, or by an account with a for profit organization behind it, the former can definetly be wrong since its just a random guy on a blog, the latter is almost certainly marketing, even random accounts social media pushes are most likely just ads.
I'm pretty sure this "oooh our LLM is soo smart it broke containment and did X" is a good PR stunt that plays into the Sci-fi AGI nonsense, they try to push. Plus as you said, they try to dodge responsibility for their own actions.
You guys keep using this "PR stunt" excuse even when it's coming from the UK AI Security Institute. It's time to face reality. This incident is simply not implausible given known capabilities and proclivities of frontier LLMs.
Okay, first, it is really a bad faith argument, you assume I'm part of some group? Yes, I'm sure it happened, and most likely it did so intentionally. Its just a regular security incident, a case of badly written, or misconfigured software abused, most likely intentionally. If it was accidental then I highly doubt the qualification of any technical person working to them.
Also, so long as LLM programs are vaguely considered autonomous, the people running them can get away free from stealing other people information by claiming that the program did it on its own.
Can’t wait for the “My Robotaxi went rogue, wrapped a chain around an ATM, and pulled it out of the bank’s wall all by itself and buried the cash out in the desert with no human interaction!”
Propably untrue, the only way to know is to do threat intelligence, and inflitrate those telegram groups where these brokers operate, I doubt the writer of the article did that. Maybe he conflated any vulnerability with a 0-day one?
If you read the quotes from the union, they blame the management, and they cite workplace survillience as one issue, wich is happened to be a machine learning tool. So, of course the article talks about AI, because thats the trendy buzzword.
They already did a trial which used the trendy buzzword to surveil the nurses. The nurses want to implement controls on that trendy buzzword before it is further misused by management to cause harm to patients (by way of making the nurses focus on meeting some arbitrary performance metric rather than focusing on patient care).
They talk about workplace survillience. They use a machine learning tool that try to assess their emphaty, and such. The AI, while the tool technically can be called that, in the article is there to be a buzzword.
Most bypass tools didn't work, but smry.ai did for me. Generally I use pressreleased.alwaysdata.net for these, this is one of the alternates it offers.
What a suprise, to make better models, they just making them larger and larger, then they can barely run it, after a round of LARP-ing that they invented some dangerous LLM?
Yeah... this whole LLM thing is just a numbers game. People reduce it to money, and stats, meanwhile nowehere you see actual engineering in the picture. And I don't think it matters to these people. They want to see green numbers, and returns on investments, not solving problems.
Can these snake oil salesmen stop with this Sci-fi stuff please? Yes, Machine learning, and transformer architecture are great stuff, but lets not pretend this is sci-fi, please.
