**Brian** (0:00)
Greetings and salutations. It's July 31st. No, I'm just kidding. Hey, what's going on, everybody? It's The Daily AI Show. I just want to mess with Beth's data about this show and skew the algorithm.
Today is July 31st. It's the last day of July. It's Friday. It's a spill the tea kind of AI Friday today. And we got some tea out there. And it doesn't usually happen. I don't usually use that term. Andy is here, Beth's here, I'm Brian. And yeah, we're going to bring it to you. So guys, I don't know if you want to start off with this. I'm happy with any. I don't need to start off this. But the whole Leo situational awareness, Ken Griffin, Citadel conversation has been a lot. And man, has it hit my Twitter feed. I got all sorts of people talking about it, giving all sorts of different angles. You know, so anyway, it's been a very interesting story.
And kudos to Perplexity, who despite my best tries to get Perplexity to say there was some conspiracy to it, more because I was just poking the bear a little bit. Perplexity kept pushing back and going, no, I don't think we can say that's an exact, there's no data there. And I'm like, yeah, but what about this, this and this? And it's like, still no. So props to Perplexity, who I was toying with, and seeing if I could get it to push and agree with me from a devil's advocate side of it, that there was perhaps more to this story than just a hedge fund taking advantage of a situation. So that is a weird intro to get to, but I don't know, Andy, Beth, does anybody want to jump on this first? Or would you like me to jump into this and explain more?
**Andy** (1:43)
Well, I have only a surface level understanding that Leo's hedge fund, which is largely invested, I think, in AI-related stocks, has taken a beating and had to be rescued by another private equity firm because they had to raise cash in order to protect their positions in some way. I don't know what the actual mechanics were, but I think this is more a story about the risks that are associated with the concentration of our overall market's value into the AI hype bubble.
If it is that, or if it's just been pre-inflated too much, and eventually it'll catch up and everything will resolve. But especially for people who have used margin to invest in AI stocks in a way that gives them a much better return if they're right about their long position. It could be that this hedge fund just extended itself too far without the foundation. I mean, I'm just speculating a lot there, so I didn't read anything about it. I just saw the headline.
**Brian** (3:03)
That's the start of it, for sure. So let's give some background here.
Leo Aschenbrenner, who we've talked about many times in this, is like an early 20s wonder kid, godlike in some AI circles, depending on how it is. He wrote, not only did he come from Columbia, graduated in 19, is valedictorian. He is quite an amazing background at such a young age. Went to OpenAI, actually got fired from OpenAI.
Leo says the reasons were not what OpenAI said, doesn't matter. He then goes around and writes situational awareness. That was what, 160 pages. That was a couple of years ago, right? It was now, yeah. I would have thought for sure there would have been a follow up to it by now.
**Andy** (3:48)
But we study it carefully. That was a paper that came out where, wow, this person really kind of put together the picture of the landscape as it should be known.
**Beth** (4:00)
At some point, maybe soon, we'll be able to tell you exactly what episode that was. Cisco or someone in the chat may already be doing that for us.
But we've only been doing it for three years, so it can't be that long ago.
**Brian** (4:13)
Well, what, AI world, crazy world or really world? Because no, three years by that definition is not long at all.
On this show, eons.
But anyway, he writes this paper, and I've said this many times in the show. I would say two days ago with what I saw and I talked about yesterday in the show with Opus, that one's going to stick in my brain as like, oh, I just saw something different. It was a story about Opus 5 that has nothing to do with this. But the situational awareness, you guys know this because I've said it on the show before, but anybody that's kind of new here, I was delayed in the Austin airport trying to get back. And I'm delayed, I'm in an airport. I can remember what the gate looked like. And I was sitting behind, sometimes you can sit seats depending on the airport behind where the little desk is for the gate, where the gate agents would, that's where I was, right up against the glass but facing backwards. And I got into situational awareness and I think was simultaneously texting you guys or slacking you guys, going like, I'm a few pages in, this is getting interesting. All right, I'm about halfway and I think I'm just going to power through. And then I read the whole thing and I was like many, not just me, not just us, like many. I was blown away by this paper. Okay.
46 more minutes of transcript below
Try it now — copy, paste, done:
curl -H "x-api-key: pt_demo" \
https://spoken.md/transcripts/1000651996090
Works with Claude, ChatGPT, Cursor, and any agent that makes HTTP calls.
From $0.10 per transcript. No subscription. Credits never expire.
Using your own key:
curl -H "x-api-key: YOUR_KEY" \
https://spoken.md/transcripts/YOUR_EPISODE_ID