Loading summary
Transcript42 lines
- [00:01]
A
We are live signal versus noise. We're back for another week of flipping through all the noise. There's so much noise out there. And so our goal is to try to bring four core topics to you across technology, markets, capital and people and tell you where the signal is, but then also point out where some of the noise is as well. It's good to know, it's good to have that context about what is noisy because a lot of times these days we get trapped in it. So, gentlemen, thank you for showing up today with me, Liam, Brian and Michael. What's going on? How are you all doing today?
- [00:35]
B
Living in a wild world, Wild, wild west. No shortage of stuff to talk about.
- [00:43]
C
A lot of noise, lot of noise to sift through. But we're going to do that for the audience.
- [00:48]
D
There's some signal too out there though. There's, there's a lot of signal going on right now actually too, but there's just information overload.
- [00:55]
B
It's actually timely to do this show because I think there's a direct correlation between the amount of money printed and then the amount of noise associated and it gets harder and harder to discern like A, what signal and then B, like what is actually happening versus what the market is saying. So yeah,
- [01:15]
A
yeah, let's get into it because I know we have be conscious of not only our time but also listeners time. The whole idea here is we want to make this a tight show. You can sit down 25, 30 minutes, get some good information, go back to whatever else you want to do in your life. So thank you for being here with us today. We're going to kick off on the first topic, OpenAI and hugging face partner to address security incident during model evaluation. So the high level here, just for anyone who is not in the know, is one of the OpenAI models. Earlier this week, one of their frontier models including GPT 5.6 SOL and even more capable pre release models escaped containment earlier this week. And so these models were out of the sandbox environment that they were being tested in and they actually got access to Hugging Face and other company in this space and there were some security incidents that occurred. And so I think this is something we definitely wanted to cover today because it's going to be the first of, it's not even the first, but it's going to be one of the, let's say first of a much larger trend of not only these vulnerabilities that historically people worried about what would happen in, you know, with human, human beings and threat vectors and now we're going to be more worried about what's happening with the latest frontier models. And so I will hand it over to Michael to kick things off today. What are your thoughts, Michael? Is this signal, is this noise? What do you think we should be paying attention to as it relates to this headline?
- [02:37]
B
I think it's signal. I think that there's two aspects that my mind goes to and depending on the day, there could be other other things. But the ones that it goes to right now when you're asking the question is a. It's eerie how similar the Bitcoin crypto world is to the AI world in the respective. There's this understanding in Bitcoin and crypto around they call like the dark forest of the web and all these different hacks and ways that you can potentially lose all of your assets. And it was always understood you want cold storage offline, specifically if you're going to manage material wealth. And that similar operational security respect is going to be needed in this world where agenda agents are proliferating, prompt injections, all the different things that are coming and the market's just not ready for it. It's the same reason why we see lots of assets lost historically in Bitcoin in the same way that we will see an insane amount of this as AI diffuses into the market. More and more people are going to be downloading these applications and doing God knows what with them. I'm sure most people have that are listening, have some level of like Claude or Codex cowork on their computer, and it effectively can almost do everything or anything on your computer, independent of the permissions that you set up. So, like, that's just one side of it as this thing proliferates. The other side is that there's no shortage of people believing that, you know, AI is interesting, it's a tool, but it won't really do much. And I think that whether this is like a psyop or not, and OpenAI is doing this from like a regulatory perspective at the frontiers of these labs, at the very edges of the models that they're working with that haven't been released, there are capabilities that are unfathomable. And this is just a glimpse, glimpse into what is capable and what is coming. And it's just going to be a very crazy, chaotic world as this continues to rapidly kind of increase and grow. From a market perspective.
- [04:33]
C
Yeah, to me, this is a little bit of both. I generally, at a high level, agree with everything you said in the sense that all this stuff is going to continue to accelerate the sort of like permission setting that you reference. It's like, well, okay, if these things can just break containment, then like what does you clicking a permission setting on your computer actually mean? It also just points to the broader sort of fragility of software. And like you said, there's parallels to the Bitcoin space, there's parallels to the traditional software space, but like we are not prepared. You know, there was, I think this is, this is not the first of these sort of incidents where something has like quote unquote broken containment or even just the models that were allegedly finding a bunch of zero days and bugs in software that had existed for decades. That all is signal to me. I do think that there is a bit of noise in this in terms of what you allude to Michael, and that I think this is the playbook for OpenAI and anthropic to basically create the perception that these things are super dangerous right now. I agree with the long term trajectory that this stuff is accelerating, but I do think there's some noise to be aware of here in the sense that this is part of their playbook to basically create a certain amount of regulatory capture, lock in, etc.
- [05:49]
D
Yeah, I'm going to go with this is signal as well. I think that the big aspects here that are interesting are that OpenAI and Anthropic have both been talking about how they need to have safe AI and obviously they at least in this instance couldn't really sandbox and control their AI themselves. So um, it's a signal that, you know, you can't necessarily trust one party with this. And then the other aspect too was that hugging face tried to actually counter all of the cyber security incident incidents with their own AI. And I forget if they were using Anthropic or Open AI, but essentially the models downgraded them and said I can't allow you to do that because of cybersecurity issues. So they had to go out and download GLM2 or 5.2 and then use that their own models that they hosted on premise in order to use the models how they wanted. And so the signal through all of that is that there is going to be especially businesses and people want actual control over how to use their data and systems around them. Similar to, you know, what Bitcoiners have been preaching about like having control over your own money and sovereignty for a very long time.
- [07:05]
A
Yeah, look, I mean I think this is should be incredibly concerning just for the general public, not necessarily because of the standalone incident, but I. I'll use myself as an example. I generally think that if I'm representative of most people totally unprepared for these types of threat vectors. I mean, even in a world where you're constantly just adding to your online presence, whether it's signing up for new things or opening a bank account, or it's opening up a bitcoin, you know, bitcoin relationship or traditional finance relationship, the amount of things that just pile up over the years, and then as you have family more responsibilities, you own house more, like there's just so much information where that threat vector continues to expand. And it's a little bit concerning to me that not only do we need to be worried about nefarious individuals, human beings that are constantly trying to take your pii, your, you know, your passport, your driver's license, anything that could identify you, your credentials, 2fa, et cetera, now we have to be worried about agentic doing this too. And so I don't think that we're nearly prepared enough. I don't know if people will ever be prepared, at least in a short order, for whatever ends up coming of this. So I think it's a little bit concerning. And Michael, you made a good point earlier just how there's a lot of similarities as to what's been happening in the bitcoin and crypto space. And obviously that's resulted in a lot of loss. So it's not all that dissimilar in the sense that there's going to be a lot of vulnerabilities, not only for individuals, but for companies as well. And the amount of loss that could happen, you know, both financially but also potentially reputationally, is significant.
- [08:39]
B
Yeah, there's a very. I don't want to say it's not a consensus, it's not a popular take, and I haven't fully formed it, but like, there's a world where you can make an argument. You actually do need regulation with all of this. And it goes back to that podcast and just like research I've been talking about with you guys publicly and privately around, but the problem is you need everyone to be regulated. Like, what's really happening here? It feels like it's just the biggest game of or not innovators, prisoners dilemma. And you ultimately need everyone to not try to rug each other and try to develop these models. And. But the problem is that everyone's worried that if somebody else stops, they will get ahead. And so everyone's chasing to get to the very edges. And so ultimately the notion of slowing down I think, like, they may actually mean it, because I was thinking about, like, Dario, as an example, gets a lot of slack or for how he portrays himself. But that would take the angle that he's somewhat dumb from an optics perspective, but he's built this company. And so the angle that I'm coming at is, like, I'm starting to reconcile is maybe he deeply believes and understands what he's saying. And even at the behest of not benefiting his firm, he's out there trying to be as loud as possible that this stuff needs to be regulated heavily.
- [09:58]
D
Yeah.
- [09:58]
C
So anyway, the only problem with that is that, like, you're not going to get China to agree to slow down. And so that's if you're playing out the game theory of this particular prisoner's dilemma. Like, you basically cannot choose to slow down, otherwise, you're basically guaranteeing that.
- [10:12]
A
Let's transition that into the next topic here, too, because we're going to be talking to Jackson. Oh, Brian, that's all you, my friend. Well done. Unintentionally.
- [10:20]
C
I didn't even mean to.
- [10:21]
A
Yeah, unintentionally. This guy's a whiz. He does so many podcasts. He just got. He kills the transitions here. So next topic would be the US vs China AI race and anthropics hypocrisy. And so who actually wants to cover this one? Because I didn't. I did not see this link yet. But the.
- [10:39]
B
Brian, continue off your. Your riff from that.
- [10:43]
A
Yeah. Set the stage.
- [10:43]
C
I mean, so at a high level, we talked about this on the last episode around Kimi releasing their model, and there's sort of this open versus closed China versus US dynamic going on. And then you had the US government coming out and saying, well, we're going to look into what's going on here. We may potentially ban these Chinese open models. And then they sort of backtracked on that a bit and said, hey, we're pro open source, but we're going to look into IP theft. And now I think just yesterday they're claiming that Kimmy actually did sort of. I mean, this was assumed, but they're now saying its fact is that they distilled Claude Fable 5 to create the latest version of Kimmy K3, and they're claiming that's IP theft. Now, the hypocrisy angle is that. And this sort of relates to, you know, a headline that was a few weeks ago around Bernie Sanders putting forth a bill that was basically saying, like, you know, the American public needs to own Some of these companies, because you just stole all our, all our data to train these models. And so these models wouldn't exist without a bunch of existing content, information and data that was effectively stolen from all humans around the world. And so to, and so now anthropic is saying, well, you know, this is a problem because they're distilling our model. It's like, well, was it yours to begin with? And so that's the hypocrisy angle here. And you know, it relates to what we just talked about around the prisoner's dilemma, slowing down versus not the regulatory capture side. So, I mean, I think this is a massive signal. It's sort of the crux of the AI race or sort of war that's going on right now. It is China versus us. It is open versus close in these dynamics. And so, yeah, I think this is a massive signal.
- [12:32]
B
Yeah. And I, I, look, I could be way off base, but I've shared it here. I think once, at least you have the, the frame here. My anchor point is like the race for asi, then it changes, and then the belief that it will come, it changes the dynamic. Is this like a structural business strategy thing versus is this like a true security thing? Because if you go back to slowing down regulation, and I'm not saying the right or wrong, this is just like a hairy, messy situation, but nobody wants somebody else to have the lead or come to the greatest part of the frontier, the most access points. And so let's say it's anthropic. And Dario, because of natural bias, we will control, we will maintain this. Well, then they naturally believe they can gatekeep on what are the weights, what are the model? Open, sourcing it, because they can just release enough that is manageable and not have to go to the US Government or the US government has a different layer. But the problem is if they're distilling it, there's a business case to be made, but there's also just like a national security and risk case to be made that if they have to continue and they're competing with OpenAI, right. So they have to continue to add more value and release these. Because I think if in a vacuum, they would probably slow down on the releases, but if they lose that revenue, there's a bunch of other downstream effects of that. So I do think that there's the notion of just business, but it just feels like there's something bigger at play when it comes to the security and risk nationally, then globally. That is the, the substrate to all the stuff going on, it's not like the version of like just pure business that you would naturally think around business strategy and distillation, etc.
- [14:13]
C
Etc.
- [14:17]
D
I'll go next. I think that, yeah, just as Brian pointed out that this the signal is that the most politicians generally has been louder about the negative impact of chemi distilling models than other models training off of the corpus of books and data that the general people of the country have put out. And I think that it's just kind of where we are within the fiat landscape that almost quasi oligopolies and technocracy is getting more favor from the government than most people. But I do think some of this is noise in that it doesn't make a lot of sense to me because the prompt from the actual user should be their property rather than the property of any of these models and labs. And so I do think that that is noise and I'm not sure why, you know, being prompting in order to learn how something is done should be actually the property of the IP of any of these models, especially if the model data wasn't all theirs.
- [15:29]
A
Yeah, I mean I think there's kind of three things that come to mind for me. So there's the market valuation standpoint, there's the security standpoint and then there's the fact that China doesn't play by the same rules that the United States or the west does. And so working backward, it's really a tough situation as Brian, you've alluded to, because if we want to slow down or regulate the industry, well, China is not going to be playing by the same rules. They don't have the same, you know, they don't hold themselves to a similar standard, even if you don't want to say it's a standard. But they just kind of operate under different assumptions. So that's one thing. And that ties into the security risk because if we want to regulate as the US and address specifically on the security side, as we just discussed in the last segment, well again you have the largest other player that's neck and neck with the US they're not going to do that. And so there's going to be a ton of vulnerabilities that could be surfaced by their frontier models. And then the third, which I think is not necessarily covered widely enough, is there's obviously the incentive for these companies to want to try to slow down and curtail China doing this. Not only because, you know, I think it's wrong to a degree, but more importantly because they need to protect their valuations. And so I've discussed on a couple of shows now, but there needs to be. These US companies are going public next year. I think even if you were to pay attention to 2025, they were expecting to go public this year. And so I don't know if those delays have anything to do with trying to get more defensibility around the valuations and try to address what's happening with China. But you can imagine if this continues to be a problem and companies like Microsoft and other large players are starting to just put all their token spend into Chinese models, well then that is going to have pretty significant implications to market valuations for U.S. companies. And so there's that certainly that incentive as well to want to fend off, you know, fend off this, this threat vector from China. All right, let's pivot over to the next one then. So Google, there's some interesting news with their free cash flow. So for those who are not on Nick, if you could just click on the chart here. For those who are not on video, what we're looking at is Alphabet Google, their free cash flows dating back to, in this chart, March of 2015. And you can see here, they've been printing positive free cash flows. They've been one of the most, obviously one of the most successful companies in the world for a very, very long time. But then you can see here, on June of 2026, they free cash flow turned negative. And so that from the CFO it said we had negative free cash flow of 5.9 billion in Q2 2026, driven by our investments in capex. Free cash flow was 53 billion for the TTM. We ended with the quarter with 242 billion in cash and marketable security. So Brian, I know this was one that you were particularly excited about. Why don't you give us a signal or noise here?
- [18:14]
C
Yeah, I mean I think it's a signal in the sense of this is not an isolated incident. So Google is in the headlines because they just reported and this is, you know, a pretty crazy chart just in the context of, you know, their remarkable run of positive free cash flow over the past decade plus. But we know what this, what is causing this, right? So it's this build out this infrastructure, build out this massive onslaught of debt in order to this race that we've just been describing. And it's not just them, right? Like it's a bunch of other of the sort of mag 7 the hyperscalers that are all in the same situation of basically Having to go all in because it's, it's a different but similar prisoner's dilemma of this, this capex build out. Like you're kind of damned if you don't just take on all this debt and attempt to chase this prize effectively. And so it's a signal to me in the sense that I think this dynamic has been sort of bubbling over the past, call it six to 12 months, but now it's actually showing up in quarterly filing. So, you know, I think Google's down 6% or so on the day. You know, nothing crazy but like it is maybe a little bit more. People are more aware of it now. Like the people are beginning to really assess some of the assumptions around all of this capex build out and whether or not there's going to be a near term enough ROI on all of this to basically make it to the other side of all this.
- [19:59]
D
Yeah, I think this is signal, the Google has been profitable actually since they went public, so every single quarter I believe, and that was maybe 2004, maybe earlier. So that's an extremely long time. And I think that the, the real signal here is twofold. One, the economy has been mostly constrained by software and that's where most of the value has been put onto the economy over the past 10 to 20 years. And I think that's completely changing. And this negative free cash flow that's going towards data center servers, chips is, you know, obviously all physical and thus the, there are different rewards now in how the economy is pricing things. And so I just think that it's a tough time if you're a stock picker or investing in, you know, the S and P because you don't necessarily know how that transition will take place. Is there a different bottlenecks in the economy? And it would be a very, I, I would feel weird being completely all in on the economy unless I know exactly how this is going to play out, which I think is just tough to determine at this point. But I think the, the big picture is that with all these air pockets as the bottlenecks to the economy change so rapidly, there will need to be some sort of massive shift and potentially bailouts for these companies that are perceived too large, too big to fail.
- [21:25]
B
Yeah, I mean if there's any chart that shows a race for asi, I feel like it's this when you look at like All In. I mean they publicly said they rather lose the race or like go bankrupt and lose this race. I think that was from Sundar, but this resembles a lot of the companies a lot of magnificent seven outside of actually Apple which is its own interesting strategy. On the AI side I think Google's just fascinating because without just known enough to be dangerous when it comes to gpus versus tpus it resembles similarly to early bitcoin mining. And there was GPUs and then there was ASICs which were specific to mining BTC. The TPUs my understanding is like they're heavily invested there, they're vertically integrated. So it has a lower cost per token. I believe Anthropic is the only other large frontier lab leveraging it. So there's a lot of that build out in place but it's understood and that's part of where Nvidia bought grok. The guy that founded Grok was formerly at Google I think that built that TPU stack. And so I think like Google's in a fascinating spot because of their like client or their just database. When you think about I think it was like 2.5 billion users between YouTube and all the other products you think about G Suite being embedded. There's so many different angles of Google's opportunity to win this. But independent of all that, what's happening there's like multiple things that could be simultaneously all true or at least one or a few of them is the race for ASI that the money doesn't matter anymore. That's the big thing. The big flag is like all these companies are spending money like it doesn't matter. And it doesn't matter necessarily because the dollar's cooked or the debt will be paid back because other individuals, the US citizen will pay the pay it have to wear the burden on a bailout. And so you start to look at that and you're like what is going on here? But I think that is like the biggest thing that's flipped over the past year is the amount of capital outlay. And it went from obviously like natur balance sheet capital to taking on debt.
- [23:39]
A
Yeah, I mean that's well said. I don't have, I don't need to add anything here. We got to move on to the next topic anyway. All right, so the next topic here would be the U.S. 30 year yield raises alarm in longest run above 5% since 2007. Nick, if you could just scroll down to the takeaways here on Bloomberg, I can just read those off and then we can get into the takes. So just for everyone's context, the US 30 year bond has held above 5% for 27 days in 2026, echoing investor concerns about a growing debt pile and sticky inflation. The sustained rise in long dated yields is driven by growing concern about deteriorating fiscal position and a deluge of issuance to fund artificial intelligence infrastructures. Debt concerns and competition for debt buyers from over 500 billion. Financing linked to AI has have prompted some investors to expect 5% plus yields to stay unlike similar spikes in the past. So I guess I'll go first on signal versus noise on this topic. I would say this is a major signal to pay attention to just because yields are far higher than they've been for two decades and the fiscal position of the United States has eroded significantly over the two decade period. And so the real thing that I would pay attention to is the bond yield ultimately is controlling markets, foreign policy, the economy. And so when you have this much pressure, when you have yields climbing up, which is just signaling a lack of demand for the long dated end of the US treasury market, that is something that can't be ignored for too long because ultimately that is the underpinning. The US treasury market is the underpinning of the global financial system. It is the kind of cornerstone of collateral for global trade and sovereign reserves. And we know of course that gold is kind of moving into that position over the past several years and will likely continue to do so. I think the main signal here is that this is eventually, it's hard to put a time frame on, but it's a massive buy signal for any sort of risk asset. I don't think Bitcoin is necessarily risk asset, but it's certainly perceived by the market to be a risk asset. So I would take this as a major signal to continue to monitor. Again, it hasn't been. The yields have not been this high in two decades and we're trending to be about on par for where it was in 2007 despite adding like 30 trillion plus to the federal debt in those past two decades.
- [26:01]
C
Yeah, I would say this is a signal of sort of a confluence of factors or variables that are at play right now. One is, as you mentioned, Jackson, sort of the fiscal situation with the short end of the curve being sort of higher for longer than people expected. We've held rates at a little under 4% now for I guess close to a year. That is putting pressure in terms of the interest expense on an annual basis. But it's also in my mind sort of most related to when you're thinking about the longer end of the curve. 30 year treasuries. It's really like calling the bluff on when is the next big print or bailout going to happen. Is it going to be related to the AI Capex bubble booming or busting? Is it going to be related to, you know, conflicts in the Middle east that are dragging on longer than people expected and costing trillions of dollars and also constraining natural resources, commodities, et cetera, contributing to higher inflation which would then in theory make it harder for the Federal Reserve to cut the short end. And so all of these things are related. But to your point, Jackson, like the 30 year rate is one of the most, if not the most predominant sort of signals that the entire global economy looks to. And so when you see a data point like this, you know, anything that hasn't happened since the global financial crisis is usually something to pay attention to.
- [27:31]
D
Yeah, I think that this is signaled just given the fact that there is so much build out that is going to, needed, going to need to be debt driven. Especially when a lot of these companies are not making debt free cash flow at this point. We just talked about Google and so there's more competition there. I think it's also a signal for things are going to change. Whether it's the Iran war and trying to find something out there because oil prices are whatever they are right now or just even the fact that maybe clarity will get done because they're going to need to find more buyers of debt at this specific level. I think that's just a signal that things are changing and there will continue to be a rapid shift. As you know, we go into uncharted territory.
- [28:19]
B
Yeah, I'll go more noise. I think like, you know, long term we naturally know the debt's unsustainable, sovereigns are recognizing that they're piloting gold. But I think the noise comes from. We talked about it yesterday on last trade regarding, you know, the fact that 322 with the interest rate hikes, it was deeply understood there was no way you could keep it higher for this long. And then my understanding is the 5%, the fact that it's been above 12 days in a row and then 27 days, I think it's like roughly 20% of all sessions this year above 5%. That would also be completely unheard of and you would need market intervention to happen. And that hasn't necessarily happened. And I think where I call it noise is yes, like these are metrics to look at. But at the end of the day we're just at a fundamentally new paradigm across the board. Whether it's people, you know, jumping out of sandboxes, free cash flows going negative in a world that Google's been profitable for the past 25 years or interest rates at this height. There's different like factors that we know of and then we don't even know of when it comes to liquidity, sustainability of the debt, you know, on the front end, the demand that's going to come post clarity, etc. Etc. That like I think it's obviously a flag but I don't necessarily know if it means what it used to mean when you think about high interest rates.
- [29:49]
A
Any thoughts on that?
- [29:53]
C
I think it's fair. I mean you're right in that the long term trajectory is. Is what we know and it's sort of, you know, these are shorter term signals is a way I guess you could parse it that are helping people navigate the state of change. But we know the long term trajectory, the debt is unsustainable, the interest expense is too much. At some point they need to get rates down. But it's like well what is the, what is the trigger going to be? And you know, rates being over 5% doesn't tell you anything about what the trigger is going to be. So in that sense I kind of agree with you. Yeah.
- [30:25]
B
And there's the other aspect of like we know sovereigns, other individuals entities are dumping treasuries like funny money. The dollar's cooked. I think that's understood in certain circles. But I think it was that currency of power newsletter we referenced a few weeks ago that an interesting dynamic. It was like like people holding bonds are effectively like prisoners paraphrasing of like you just naturally have to hold them if you're an institution pension and they're increasing. I think it came out this week as well for some of the private credit stuff may get rol individual pensions and other accounts that I just think that there's just different structural ways to manage this that historically haven't been in place or to allow it to stay higher and then other vectors of the economy to absorb that debt.
- [31:16]
A
All right. Well gentlemen, we got through our core topics of this week. We want to do a quick lightning round before we wrap up for today.
- [31:24]
B
Yeah, I'll do. I'll do a quick, quick hit. We're kind of flipping this around. I thought it was supposed to be content, but I think it's other ideas that we're looking at. So the ideas that just recently popped up that somebody had pulled, I think it was Brian. I haven't read it yet but I. There's a mosaic. There's something happening where Bloomberg had how tether benefited as a Trump insider shaped US crypto law. And then just recently the block came out with something From Corey referencing 21 and tether serves political interest. And then I just had a call with a large institution and they were referencing like that there's zero chance that tether will operate in the in the US around genius, even with USAT from a regulator's perspective. And I thought it was fascinating from a signal just because I do think that there's in the same way the AI is like kind of like Wild West, a lot of competing forces. I think that we're seeing whether it's Clarity act passing, not passing banks, consortiums, tokenized deposits, stable coins and then where tether fits into that in the Trump administration there isn't like a concerted view or like tied in view of where we're going. And right now there's just a lot of things at stake.
- [32:43]
D
I'll go next. I thought we were still doing content of the week, so I think it
- [32:48]
C
could be whatever you want.
- [32:50]
D
I'll do the content then. Especially since I was surprised that Michael didn't bring this one of Travis on a 16Z talking about his new business Adams and behind its funding. I I just think that he succinctly laid out the case for robotics, machinery and automation in the real world that many people are still sleeping on. So I would check that out. And there are also some interesting stories just from the Uber days too.
- [33:19]
C
I can go we don't need to pull this up but it's a tweet from an account mgoes distance on Twitter. He has an investment fund called Superhuman Fund and it relates to what I talked about a little bit last week in terms of the positive and negative externalities of AI advancement, the acceleration and so what this fund invests in is basically like biotech and longevity type things that are being rapidly accelerated by AI. And it sort of lists out a number of things that happened just within this past year alone around germline editing hitting 100% efficiency with zero homosonal abnormal abnormalities. A baby born from an embryo selected for an IQ of the 99.99 percentile. Mid Journey created a full body ultra scanner ultrasound scanner. China reversed type 2 diabetes. An Austrian founder sequences dog's cancer used AI to design a custom MRNA vaccine and watch the tumor shrink. A few other things here, but my main point is like while I talked about the education stuff last week and how that's a super positive externality, I think there's this one is mixed on the sort of life sciences application of all these things is there's certainly going to be positive externalities from that. If we can cure more diseases, that's. That's obviously fantastic. But it also relates to, like, the open, closed china, US Side of things, because there are some more unhinged ideas on the life sciences side that, you know, are not going to be able to be sort of researched, even on some of the frontier models that have basically these gates around what it can can even be asked or researched. And so I think there's potential negative externality that the more open stuff is going to basically allow anything to happen on, on this side of sort of the industry. And so it's sort of a mixed bag in terms of, yes, there's a ton of great things that could come from this, but there's also some pretty scary dystopian type things that could come from it, too.
- [35:28]
A
Good points. Mine's mine short, and we said it could be anything. I'm just gonna read out this tweet from our buddy James Camp, which I think is just a good note to end on. One day, you'll realize nobody cares what you do. And if you fail, the only one stopping you from being successful is you. I promise. I just want people to know that. I want people to know that, you know, let's end on a high note. Take agency. Better your life, Better the people around you, Better your business. All right.
- [35:58]
B
Jackson needs a therapist. I feel like these pods are just his therapy. Good stuff.
- [36:01]
C
Later. Yeah.