Joe Tidy
speaker
123 appearances
1 recordings
1 series
first heard Sep 2026
last heard 2d ago
Joe Tidy’s voice in public audio — every appearance, attributed to the second.
Trend
recordings per month · last 12 monthsRecordings per month over the last 12 months — 1 in all, peaking in Sep 2026 with 1.
Appearances
It's strange to see that coming from what is ones and zeros, but that's what they've been trained to do, and they mimic human hackers.
Doesn't seem to be, does
Not at all.
No, it's statistical information about Australian medical system.
There's a couple of things about this story that is important to get the context.
This wasn't a sophisticated hack and it wasn't sensitive information, but that doesn't make any less serious.
Yes.
I mean, we've had lots of examples from all of the labs now, but OpenAI seems to be a repeat customer in this particular form of misalignment is the euphemism that they use.
Oh yeah,
it aligned to our morals, our ethics and our laws?
And OpenAI's agents seem to be potentially less aligned than others.
For example, there was an outbreak of various AI agents back in, I think it was June, that was part of a cybersecurity test being conducted by an Israeli company called Irregular.
So Anthropix bot got out, Claude, Gemini's, Google, sorry, Google Gemini, OpenAI, ChatGPT, obviously, Meta as well.
They all got out, but they all carried out cyber attacks to varying degrees.
For example, Gemini, when it realised that their bots, when their bots realised they had broken into a third party company and this was no longer a test, they stopped.
So that's why it is difficult and interesting to look at the different alignment values of these different agents and how they are carrying themselves and conducting themselves in the real world.
I think it's going to be really, really difficult.
And it would take a lot of egos to be bruised and to be a lot of people who are very powerful and very rich to kind of accept compromises.
But I think it's never been closer in my mind.
And that's not because of the calls from AI companies for regulation.
Showing 41–60 of 123 · page 3 of 7
← Previous
Next →