Claude Opus 4.5’s Breakout Moment & Investing in 2026 with Qiao Wang

Okay, here's what I realized over the holiday. It's exactly what you described. The line between general chatbot and coding is blurring. Now when you code, you don't actually write code, you write plain English. So like the two things, the coding assistant and general chatbot, they're merging. The software mode is diminishing very rapidly. And for early-stage startups, there's basically no modes. But if you look at the likes of Facebook or Google or Microsoft, Apple, the modes still exist. The AI coding agents won't kill them.

Nothing said on Empire is a recommendation to buy or sell any investments or products. This podcast is for informational purposes only, and the views expressed by anyone on the show are solely their opinions, not financial advice or necessarily the views of Blockworks. Our hosts, guests, and the Blockworks team may hold positions in the companies, funds, or projects discussed. All right, everyone, welcome back to Empire. Very excited about this one. Got good friend friend of the pod, friend of me, Chow from Alliance. Welcome to the show, Chow.

Great to be back. Or great to be here. Have I been on this?

Come on, come on.

I know I've been on other Blockworks.

Not many of the Blockworks pods. Yeah, yeah. I have an embarrassing thing to admit to you, Chow, which is I have a— I've changed my kind of like a scroll. I still scroll social just like any normal human being, but I've got to— I only use lists now. I don't do the algorithm. And I've got a list and it's just called Brains. And I was like, I was like, oh, this is like really enjoyable. Like, I really enjoy my like consumption behavior now. And I was like, who did I put on this list? It's like you. I think it's like you, this guy, Will Manitas, I think is his name, and Chris Berniski, like 3 people on the list. I was like, oh, that's why I enjoy it. Okay. So I'm just basically, I'm just telling you, I mean, I've consumed a lot of your thoughts and tweets recently.

Appreciate it. I started doing the same about half a year ago just to get rid of all the I don't want to see all the political stuff, all that stuff. Like now it's pure either markets or tech. I have two lists, markets and tech.

Yeah. Yeah. I have health and brains. So, um, it's interesting. I had a crypto top 200 list that Blockworks made and I curated that whole list. I looked at the top 200 list. I made it probably a year ago. It's all outdated. All the people, or no, I probably made it 3 years ago. All the people who were like quote unquote big names, the thought leaders, I'd say like 150 of the 200 have essentially washed out, some version of washed out. It's very interesting. Yeah. I want to talk about Claude and Opus 4.5 to start. You've been just talking about it a lot. I think maybe the best place to start is you— there's a tweet. Speaking of tweets, I'll mention a lot of tweets in this episode. You said the quote unquote holy shit, it's over moments for me in the last few years. One was the release of ChatGPT. Two was the first reasoning model, O-1. Three was the Tesla FSD V13. By the way, I just used that. I think, I think I, I think it's now the V14.

It's on V14 now.

I just tried that for the first time. I was like, I live in a parking garage with very, or I have a parking garage with very tight parking. I was like, oh my God, it got me out of that, drove me all the way to New Jersey. So that was crazy too. And then you said Opus 4.5. Why is Opus 4.5 such a holy shit moment here?

I can't explain to you like what's happening under the hood because I didn't build Opus, but I can tell you what I felt as someone who used it. The last time I wrote any sort of code was maybe a year ago, but I really seriously only coded, it's been like 3 years I haven't really seriously coded.

Give me your background before we go too deep to just contextualize your coding ability.

I would say between 2010 and 2017, right before I got into crypto, I was in quant trading and I spent all day coding both the super low-level C++, the actual trading engine, as well as the higher-level Python data science stuff. And that's my work for like 7 years all day. Is coding. And then at Masaru, I wrote some front-end code, a little bit of front-end code and back-end code. And then ever since I left Masaru, I haven't coded for like seriously coded for 3 years. I did some personal projects here and there, like maybe at the end of the year for every year, the last 3 years. But like Opus 4.5, like everyone has been talking about vibe coding, but the The big difference between Opus 4.5 and before was that before you can sort of put together a demo very quickly, but the last 5% of the project always, like it's really hard for the AI to do it for you like completely. It still requires a really good engineer, like a human in the loop to get that last mile done. Like the last mile meaning like the bugs and the corner cases, all that stuff. And with Opus 4.5, what I felt was I just tell it what I do in plain English. Obviously I have to give it a very clear spec, but as long as my spec is clear and comprehensive, it can just do it. It can one-shot it. And that's how I felt that the biggest difference was. Of course, I think Claude and Opus might be going through this hype cycle on Twitter because my engineering friends told me that, for example, OpenAI, the latest OpenAI GPT 5 Pro is as good or almost as good as Opus.

But I use it. I've used both. And I actually, Opus is the first time I've switched out of ChatGPT because for me, ChatGPT, the memory of ChatGPT had locked me in so intensely that I was like, I guess I'm a chat guy for my life. Like it knows everything about me. And now Opus 4.5 is so, to me, it's so much better. I know the models. I looked at the model ranking, but to me, Opus 4.5 is so much better.

As a chatbot or as a coding assistant?

Um, uh, both actually. Yeah.

Okay. I haven't tried, uh, Claude as a, as a, as a general chatbot, but just on the pure, on the coding part, it's, it's a huge jump.

There's a, there's a level of work that I don't know if it's coding or chatbot, but it's somewhere in between. For example, um, let me tell you a thing that I just did with Opus 4.5, which is we have, uh, we have a, we have a TAM of Blockworks, right? And it's always a headache. We have an 8-person sales team at Blockworks and it's always a headache to be like, who's— which accounts should we sell to? Is that a tier 1 or tier 2? Each seller has a determined, like, you know, you get 10 tier 1s, 30 tier 2s, 100 tier 3s, something like that. It's always a total headache. I just put the problem into Claude. It pulled in and I was like, look, if they have an FDV, if the token has an FDV over a billion, that's a, you know, that, that gets extra points. If they're a public company, that gets it. It pulled in everything. It pulled in info from CoinGecko, CoinMarketCap, all this different data, and it ranked the list. And then it looked at every deal the seller— each seller had closed over the last 8 years at Blockworks. It looked at the products that we sell, and it then assigned basically accounts to each seller based on the likelihood that they would close a deal with that account. That was— that is— that is— that's not just chatbot work. It's not— I know it's not coding.

Here's what I realized over the holiday. It's exactly what you described. The line between general chatbot and coding is blurring, right? Because when you talk to a general chatbot, oftentimes it writes some code on the fly for you to do the thing you want to do. Whereas now when you code, you don't actually write code, you write plain English. So like the two things, the coding assistant and general chatbot, they're merging.

It's one. Yeah, it's merging. Yeah, exactly. Why don't you use Replit or Lovable for your vibe coding? Why do you use Claude?

I feel like, I think they're targeting different personas. I haven't tried the latest versions, but I feel like Lovable is generally targeting people who want to build a really good demo or maybe a simple app. I think they're really good for front end. That's my impression. I haven't used it myself. I think they're really good for front end. But the tool I wanted to build for myself was mainly a backend. Thing. And so I used— I started with Cursor plus Opus 4.5. Then I tried Cloudverse with Opus 4.5. I actually couldn't tell the— like in terms of productivity boost, I think they're about the same for me. But that's just my personal experience.

Yeah. Talk to me about how you— so for people who don't know, you've been running Alliance where you were I mean, helped build Messari, then launched Alliance. You guys are on cohort 17, I think, or 16?

We're going to be on 16.

16. You've seen thousands of crypto startups now at this point. How does this impact companies?

Yeah, in a huge way. And this has been happening for the last 3 years, which is that basically almost every cohort I ask people, especially the technical co-founders of every startup, how much productivity improvement have you experienced since ChatGPT in 2022? And basically every cohort I get a number that is bigger than the previous cohort. Like it's really consistent. And the latest number I got from the last cohort was like somewhere between 3 to 4x. Productivity improvement. And of course, I think this probably impacts early-stage startups even more so than later-stage, like, than bigger companies in terms of productivity improvement. Because I think fundamentally the limitations with coding assistants, AI coding assistants, is the context window. So for example, for argument's sake, like, let's say Claude can take in a million tokens. Then what that basically means is it can write a, it can start a project very easily and do it really, really well. But if you ask Claude to change the entire Google codebase, it's not possible, right? So it probably affects early stage startups even more so than the bigger companies.

Yeah.

I think what was happening within the bigger companies is the way they can best leverage coding assistants is they have clear layers of abstraction between different divisions, between different departments of the company so that they can apply, so they can shrink the context window and feed it into the AI system, if that makes sense.

Yeah, it does. I've seen that talking about AI in terms of productivity is the wrong lens to view it from with the earliest stage startups. So I think if you're a company, 100 employees, 200 employees, 1,000 employees. What's your productivity gain? 3x, 4x. Sure. Um, I think that the earliest stage startups don't even understand what— they don't think about it in terms of productivity. They just think about it as like, why would we ever go hire more people? That's what I've seen in the, like, you know, companies I've angel invested in and like friends are running like new startups and stuff like that. They're just like, we, we don't have to hire. It's, it's really bizarre. We, you know, even this thing of, um, this target accounts example that I was doing with sales, the other two things that I just built are a sales commission calculator. Sales teams always like, how much am I making? Like, then you got to ping the finance team. Then finance seems like, oh my God, like I got to calculate the seller's commission. Like, look how much money they're making. Like, that's a pain in the ass. And then the other thing is how much does a— for us, we sell these data dashboards, right? So it's like, how much does the dashboard cost? Then they got to bog down the data team and the data team's like, oh, well, like, are they indexed by Dune or Gold Sky? It's like, no. Okay, well, we got to add a little extra money there. I just built a dashboard calculator based on all the— so, so it's like these little— and so what does that actually mean for us? We don't have to go hire a sales ops person. Yep. Right.

So it's like, it really does feel like a 1 or 2 person unicorn is possible in 2026.

Yeah.

And such a startup probably has already started. It's not at a billion dollar valuation yet, but it has started.

100%. Yeah. I mean, I know multiple people who are running $10 million ARR businesses by themselves right now. Yeah, yeah. They're often like ex-engineers from Meta and Uber who got sick of the bureaucracy, and they're just like, yeah, and they're just cooking by themselves. There's an interesting thing though I've noticed with AI. Some of the best AI start— not AI startups like a ChatGPT, like an OpenAI, but like these little businesses that are leveraging AI— is they don't want to tell people what they're doing. It's one of the first times I've seen— and usually if you have a startup and it's ripping, right? ARR through the roof. You want to shout to the world, you want to go raise the money, you want to go tell everybody, you want to post on Twitter. People are very secretive about what they're building.

Now it's too easy to copy your software.

It's too easy. Yeah, exactly. How do you think about moats then for companies? And does that definition of a moat change?

It doesn't change. I mean, the software moat is diminishing very rapidly. And for early-stage startups, there's basically no moats. But if you look at the likes of Facebook or Google or Microsoft, Apple, the moats still exist. The AI coding agents won't kill them. And so the moats for these companies, for example, it's developer ecosystem for Apple and maybe PC like Microsoft. It's switching cost for cloud, AWS, Azure, Google GCP. It's proprietary data, right? Like YouTube, YouTube owns a lot of video data. That's how they're able to build a really good video model in the first place. Like enterprise integration and like super mission-critical software, like that's Microsoft. Sure, you can copy Office, but you know, are you, are enterprises really going to switch from Office to a new software. It's too mission-critical for them. Actually, I'll give you one example I've been looking at recently. It might be one of the biggest inefficiencies or mispricing in the public market right now. It's Adobe. It's Photoshop and the Creative Suite of Adobe. The market currently thinks that the latest video models and image models are going to kill Adobe. I think that's completely wrong. Because Adobe's moat is the enterprise integration. Like all the enterprise users of Adobe Creative Suite software, they store their images and their videos on Adobe's cloud. And they're creative professional, like the switching cost of moving from Adobe Cloud to somewhere else is very high. And the fact that they've been using Photoshop for so many years, like it's muscle memory, right? It's very hard for these people to switch. And so Adobe right now is trading at like 12x forward PE. It's crazy for such a high quality of a company.

Yeah. Is Scott Belsky next in line to run that company? Have you seen, 'cause that would be, that would, he, phenomenal entrepreneur. I'm not sure if he's still online, but that would be very interesting. How do you think companies should like— what's the— let's say you're, you know, so there's a lot of like, you know, very maybe young entrepreneurs who are like, this is kind of second nature to them already. What if you were at a company where it's— you're listening to this and it's not second nature? Like maybe you work at a, I don't know, 500-person startup, been around for like a decade. And you're just like, yeah, no one's really— maybe people are like putting things into ChatGPT and sending like automated emails, but like, how do you push a company onto this like next wave of what obviously needs to happen?

I don't think you can force them to adopt a new technology for the sake of it. I think they have to just figure out where a technology like AI can help them immediately and make a huge impact. By the way, that's what has been happening within Alliance for the last 3 years. We didn't, you know, we knew AI 3 years ago, we knew AI was going to be a big thing. We said we're going to become AI first, but we didn't like try to force AI into every part of our organization. But instead we used AI, for example, to automate some of our admission process. Like we get like thousands of applications. Like 3 years ago, I used to read every single one of them myself. 5,000 applications per year. It's brutal. It's soul-crushing. But now I'm able to automate about 50% of it using AI.

So how do you, how did you actually do that? Did you, so is that an Opus 4.0? Like, how did you, how do you, how did you build that? How did you put in the inputs? How good is it actually on the outputs? Why is it only 50%? Why isn't it automating 99%?

Yeah. So I didn't write it myself. The software was written by one of our engineers. I think the coding part is no longer the bottleneck. The prompt is where the, I wouldn't necessarily say the secret sauce because everyone knows what a, or for the most part, what a great founder might look like on paper, right? So I just translate my heuristics of what a good, a founder or a good startup might look like and put that into the prompt. And I use that to filter. And right now, what that piece of software tries to do is filter out the obviously bad ones rather than tell me which ones are obviously good. Because I think that part, like, it still requires some taste from the human interviewer. I think the AI isn't quite there yet, but I think you could get there like by the end of the year. Like, yeah, by the end of this year, I think AI and AI VCs is going to do a better job than I do.

Yeah. Yeah. That's interesting. That's, um, I think the Revolut founder launched a venture firm and that's what they're doing. I mean, this was Chamath's idea with Social Capital, if you remember, right? It was, and then it didn't, didn't work. I think maybe AI was too early.

It was too early. The tech wasn't ready.

Yeah. Yeah, very interesting. Well, how much, how much value do you think you get out of these? Like, how much would you pay? How much? So to me, it's shocking how cheap these are. And it's, you know, it's kind of like taking a $5 Uber in New York from like, you know, Fidei to the Upper East Side. When Uber launched, you're like, okay, obviously this is mispriced. But they're just throwing venture dollars at it to onboard users. How much more? What do you think that mispricing is? Like, how much more What would you pay for Opus 4.5 right now?

Opus, I'm not quite sure yet. Depends on what I build with it, but I'm not quite sure yet. But for Gemini, I think it's underpriced by 2 orders of magnitude. 'Cause right now I'm just paying $20 for a Pro. I'm not even using the Pro Plus or—

You would pay $2,000 a month for Gemini?

Yeah, because it's a research, like I actually tweeted about this, it's a research assistant. It's a junior research assistant, junior coding assistant, a pretty good health advisor, like medical advisor to double check whatever my doctor tells me, a pretty good paralegal and so on, and a therapist. Like if you combine all these things into one, $2,000, I think is underpriced. I think Google knows this. OpenAI knows this. They know it's being mispriced.

Yeah. How does your wife feel about you putting every single thing into, into Gemini?

We both put a lot of our things into Gemini. Separate accounts, of course. Yeah.

How much time do you spend a day on— how much time do you spend a day on some combo of Gemini and Claude, these kind of things? And has it— where has that time taken and been taken out of? Is it out of Zoom calls? Is it out of social? Is it out of Slack?

Yeah, I actually analyzed this maybe half a year ago, which led me to the Google trade, basically at the bottom.

Unbelievable call, by the way. Like, you were the only one longing Google at the time, and now it's, you know.

I looked at my iPhone usage. The top 3 apps I was using, Chrome, YouTube, Gemini, they're all owned by Google. Yeah. And of course I was worried that ChatGPT was going to kill the search business. So I talked to my wife and my wife gave me an insight where she said she does all her Google, the biggest use case of Google for her is shopping. And that part was, at least in the short term, was not gonna get killed by ChatGPT, at least at that point in time. So then I did some research, I realized, actually more than half of Google's search revenue comes from shopping. I was like, this, this castle is, is intact. Yeah, no one's gonna kill it anytime soon. And then on top of that, there's the GCP, there's TPU, all these things that people realize later. Um, I thought that was a no-brainer. So Google was my biggest trade. It was like the only big trade I took last year.

Did you size it well?

Um, I put like maybe 10% of my network into it.

That's great.

Yeah, that's great.

Um, yeah, I got dinner, um, on New Year's Eve with someone from Google and we're, you know, we started talking about these things and people were talking about what they use and the Google person's like, what all you guys misunderstand is we've got the shopping data. We have the shopping data and no, no, the market does not, still does not understand that yet.

So, um, I think Adobe might be the Google of this year because it's so cheap. Like it's experiencing the same fad that Google experienced.

I don't, I just don't, but like, when's the last time you've used Adobe?

Never. But so I'm a lot of—

no new people will use Adobe. That's my problem with it. Like Google, I am the same. My top apps are like you, I'm on YouTube, like YouTube. I use YouTube so much. I use Gemini, all these things. I don't use Adobe. So all the new people don't use Adobe, but all the new people use Google.

I think what the market misunderstands is that Adobe is not a consumer product, it's enterprise product.

And enterprise, my enterprise, like quote unquote my enterprise would never adopt Adobe.

So what would they use? Figma? Figma. Yeah. The Figma and Adobe, they're actually targeting two different market segments. The war is actually over for that segment of developing like new demos and, you know, website Interesting. Like that war is over. Adobe lost that war. Figma won. Adobe is more for—

So then what does Adobe win?

Like the marketing people.

But Canva?

Canva is for lower end segment of the market. It's more amateur versus like true high-end enterprise. So I agree with you. Adobe's net new users has not grown for a very long time. But their pricing power is so insane that they're able to charge more and more every year.

Do you put on the trade yet?

Yep. Small one because I don't use it personally. I don't have the same amount of conviction as Google last year.

Yeah. Well, I want to get into investments in a little bit, but I want to talk about this idea of like where you should be spending time.

Yeah.

How do you think about how AI changes what you should spend time on?

I think everyone can and should write code. And if they don't, they're going to be left behind. And by writing code, I mean specifically automating some part of their own life or their own work. I think what's going to happen is, or what might, I have no idea, but what might happen is there's still going to be these like B2B SaaS software that everyone purchases and every company uses, right? So for example, we use, I don't know, Gmail or whatever, or Zoom, and we pay for these things. But there's always this long tail of workflow that is unique to yourself and to myself that it doesn't make sense for third-party software providers to build it because there's not enough money for them to make. But you and I can build this software.

I have a commission calculator at Blockworks. That's a good example. No one, there's not, there's not a SaaS business to go calculate commissions, but yeah.

And when I say like everyone should be coding, like it shouldn't sound scary anymore because no one's going to write any code anymore. You just talk to the thing in pure natural language.

I would highly recommend anyone listening to this, don't even use Claude Opus because there's, it can give it, it's too much.

Just use Inet10 or something.

Use, use Replit for the love of God. Go, go on Replit.

It is.

That was to add to your, you know, FSD v13, Opus 4. For me it was Replit. Replit was my moment where I was like, oh my God, AI is not just a chatbot where I talk to ChatGPT and it gives me some nice answers. I was like, you can build anything. The whole world is gonna completely change. I would really recommend people try Replit. It's amazing. It's very cool. Why? How do you think like the people who use AI in the workplace, like, so I'm kind of envisioning it like this, people aren't watching on YouTube, so they can't see, but it's like, you know, people who use AI, people who don't use AI, and then it's just going to go like this.

Yeah.

Like, how will that show up? There will be a K-shape of the workforce. Maybe just talk to me about how you see that happening.

Um, I feel like it's going to be similar to to the internet, like these things just make the high agency smart people smarter and the low agency people dumber. It's just an incredible tool that's out there available for you to use and it makes you more productive. So it's really a matter of how much you want to use it.

This episode of Empire is brought to you by VanEck's On-Chain Economy ETF, NODE. If you feel like you are constantly missing alpha, never have time to scroll through crypto Twitter, Node is for you. Node is actively managed by an experienced team with years in both tradfi and crypto. That team at Node looks to capture the most powerful trends across all of crypto while actively managing volatility as markets evolve. Super important in today's environment. Check out Node's track record so far at vanek.com/empire/node. That's vanek.com/empire/node. Www.vanek.com/empire/node.

You can also—

a little easier— just click the link in the show notes. Thanks to Vanek and Vanek's On-Chain Economy ETF Node for making today's episode of Empire possible. Mantle is launching the Global Hackathon 2025 to accelerate the future of real-world assets with a $150,000 prize pool backing from a $4 billion treasury and direct access to Bybit's 7 million+ users, this is the ultimate ecosystem for builders. Mantle has entered a new phase as the distribution layer connecting TradFi and on-chain liquidity. To accelerate this vision, the Mantle Global Hackathon 2025 is inviting developers to build scalable real-world assets and DeFi products. By building on Mantle, you get direct access to Bybit's 7 million+ users for potential listing exposure, support from the Mantle Treasury, and mentorship from top VCs like Spartan and Animoca Brands. With 6 tracks prioritizing real-world assets and real fi, and a $150,000 prize pool plus grants, this is your chance to deploy on a high-performance modular Layer 2. To register, click the link in today's show notes. Thanks to Mantle for making today's episode possible. As always, investments in blockchain technology involve risk. Terms and conditions apply. Is there some sense that it's a, um, to play devil's advocate, is there any sense that like you built your Warren Buffett stock tracker?

Yeah.

Took you a lot of time probably. Um, was that a good use of time?

Yes. The, the, the end result, uh, was, was okay, but the process of building this tool helped me understand this technology better.

What did you build? Tell us about it.

So it's a digital clone of Buffett and Munger and Howard Marks and Druckenmiller and all these people that scans on a regular basis thousands of tickers and it conducts deep research on every single one of them. And it tries to emulate the mental models of Warren Buffett and Munger and so on. And so the coding itself was very easy, but the prompt took me several months of iterations. So I actually have a very detailed prompt to emulate how Buffett and Munger think about evaluating a potential investment. So it's a 6-step thing. And so it uses Deep Research to gather information for every single one of these 6 things that they would potentially care about. And then I run another API call that uses the reasoning model, which is separate from the Deep Research. Deep Research is really good. And by the way, there's one thing I learned is Deep Research is really good for gathering facts and data. And reasoning models is, it can hallucinate a lot, but once you give it the facts and data, it can reason much better than Deep Research. So then my second round is a, is a reasoning model call to make a decision on whether or not the investment committee consisting of the digital Buffett and Munger would actually make what investments they would make. And it gives me the output. That's the whole—

Have you heard the theory that Renaissance basically discovered LLMs before everyone else didn't, they just kept it a secret and that's why their returns have been so high?

It's possible. And the reason why I did Buffett and Munger is precisely to not compete with Renaissance because those guys are really good at short-term trading with the horizon of maybe daily or weekly time horizon. And at that time horizon, there's zero chance the latest AI models are going to be right.

You're not competing against CHUMP or Susquehanna or SIG.

But at a multi-year time horizon, no one has the patience to invest and hold it for more than 5 minutes these days. So you can leverage an LLM to—

So then whose model do you pick? Do you go with what Howard Marks is saying to do? Because what Howard Marks is saying to do is very different than what Druck is saying to do, which is very different than what Buffett is saying to do.

Exactly.

So who do you go with?

I average them out.

Ah, they all say Adobe and you long Adobe.

Exactly. And here's the other thing. If you run the same prompt multiple times, it'll give you different answers. So if you run the same prompt multiple times, you average them out. If every single one of 5 runs tells you that you should buy it, then you have pretty good conviction that it's a good idea.

What's another stock that it says to buy?

I think it output about 10 stocks and 4 of them are actually in the current portfolio of Berkshire Hathaway. So it has—

but isn't that obvious because it's pulling in Buffett's—

I specifically tell the prompt, do not use in any circumstance the current or past portfolio information of Berkshire Hathaway.

Ah, okay. So what's the stock?

Chubb, which is an insurance company.

Chubb, interesting.

Yeah. Google.

Yeah.

And there's one more I forget.

How are you uploading the data? How do you like, how do you, because I built, I built something. So I built, I have these things called cluster headaches and they're these like gnarly. Yeah, it's a head thing. And there's all this new stuff. There's always new studies and research that come out about it, but I don't know how to read like studies that well. And so I always upload my studies and then I was like, all right, I'm just building a site. So I built a site, it's like clusterheadaches.co, I think it is. Um, and it pulls every single study. I put the study in and any article I want in, and then it uploads it and starts to build a database and it, and it starts to actually build a, like a really nice user interface where it's like, oh, it'll pull in like, oh, this medication. Oh, this medication was mentioned 7 years ago here. Oh, that's tied to like this thing in the brain. Oh, this thing in the brain is tied to that because like I was trying to do all that basically like that meme of like, you know, the wires crossing on the whiteboard type of thing. I was trying to do that in my head and like on all these notebooks and now this website does it, but I have to manually upload the studies. How do you have it pull in automatically like the studies and research and papers and things like that?

I tell the Deep Research agent in my prompt that you need— you can only use the following sources, and it'll do the search. It'll do a Google search on these sources. Okay. However, I think that the output isn't perfect. I think most of the time it'll follow my prompt, but occasionally it'll go to some random websites like Seeking Alpha, like, which I, you know, that kind of stuff.

Do you, have you found a way to pull in YouTube videos?

YouTube videos, I think there is a setting. It doesn't work for all of YouTube videos. But it works. It could work.

I've been manually— so YouTube now gives you the transcript for everything and you can— so I just— there's a button that says like copy this transcript. So then I manually upload the transcript, but I'm like, there's gotta be a better way. So, um, how does this change?

The most important thing is that the insane thing is one of the latest Gemini models can actually watch the video itself, not just looking at the transcript. You can, you can watch, you can analyze your, your body language or facial expression, that kind of stuff. It's really insane.

It can analyze the body language and facial expression of the person in the video?

Yes.

Interesting.

As opposed to looking at the transcript, which is like very—

Interesting. Oh my God. Oh, that's interesting. That's very interesting.

The world has completely changed in the last 3 years. It's like I feel behind every day.

Yeah, me too. How does this change your job at Alliance and how does this change how you think about where you should be spending your time?

So this goes back to the earlier point that I mentioned. I think we always want to start with the perspective of how we can help our founders, how we can make the investment process easy, and then reason from there what technology we can use, rather than instead of trying to shove everything, force AI into all parts of our organization.

So how does it, your day-to-day, how do you think about what Chow's job looked like a couple years ago and what Chow's job looks like in a couple years from now? Is it similar or different?

It's going to be very different. A couple years ago, I spent a lot of, just way too much time reading the applications myself. And 3 years from now, I probably don't need to spend any time on it. I think AI will do a better job than me on the due diligence front. But I think founders still want to have someone to talk to. The human aspect is very important. You feel lonely and you feel desperate sometimes and you need an accountability partner. And I think AI might do that eventually, but I feel like in the short term, there needs to be a human aspect.

Have you looked at Delphi? Not Delphi Digital, Delphi the AI clone.

The digital, yeah. I looked into it, I haven't used it.

It's gotten— I used it when it first came out, let's call it like 2 years ago. It's not, it was not very good. It has gotten so much better. You should look up, um, Brian Halligan, who's the founder of HubSpot, created his digital, uh, digital clone because Sequoia is an investor in them and he was doing like a residency at Sequoia or something. It is— not only can you ask it question, like you used to be able to ask it questions and then it would output and it's like, it's like sounds like Jason Yanowitz doing it. Now it's like the video, it's like me and you can't tell, you can't tell that it's me. Like you really can't tell that it's not me. It's a video of me. It's nuts. So you should look at, you should look at Delphi. What is—

which means that the scarcity in 5 years, 5 to 10 years is the in-person relationship.

That's, yeah, that's why I'm actually very bullish on our events business. We cut our news business, which you might have seen, but we kept the events business. I'm very excited about the events business. Like, the multiples are lower, like all the, you know, it's not like a venture bet, like all that kind of stuff. But, um, yeah, the events— I think events businesses are going to get more and more and more valuable. Events will change. Like, this kind of lame conference thing is going to change. But like, yeah, highly— there's more people with more— there's, uh, what's— I don't know a better way to say it than like the rich are getting richer and they want really, really, really amazing in-person experiences. And we want to provide that for them. Where was I going with that? What can you not automate? I think like you mentioned, writing is something where you don't use AI. Maybe talk to me about how you think about that.

You can automate, but you probably shouldn't automate because writing for me is— writing is thinking. Clear thinking, clear writing. So when I have a bunch of thoughts in my head and I want to really clarify what I'm thinking, I put it down. Maybe I do this over a period of multiple weeks. I think about a topic for multiple weeks and whenever I have some thought, I put it down. And at some point I have to put everything together into one sort of post, like one page or two page or even longer. And the process of doing that helps me think better. And I simply can't use AI to do that for me. AI is going to produce some junk that doesn't align with what I have in mind.

Yeah.

And so that's one thing. Writing actually helps me to think. And two is if you're on Twitter these days, everything is slop. But if you're actually writing something human, I think people can tell. So if you look at my Twitter, my tweets are just full of lowercase and full of you know, grammar mistakes and, you know, typos, all that stuff. But it feels more real, right? I think that might be more valuable than—

Yeah, I think there's a misunderstanding of why writing is important. And if you think about writing as a thing, an output that needs to happen, this blog post needs to get live, I need to produce 3 pieces of content a week, you will always use AI if that's your mental model for why to write. But if your mental model for why to write is, I need to hone my thinking and sharpen my thinking, then, then, then your argument makes sense.

Yeah.

Um, I want to talk about investing, um, and, and, and crypto, which we're, yeah, 40 minutes into this, we haven't talked about crypto, but, um, which is maybe a sign of the times. But, um, before we do that, can we talk social? You mentioned your— so I've, I've noticed a very, um, deliberate strategy change in your social media account. In your Twitter, which is dumbing it down, more lowercase, more grammatical errors, more misspellings. And I know you and you're too smart to do that not on purpose. So why are you doing this?

I don't try to put typos on purpose, but typos just happen. When I tweet, I just I just tweet in production. Like I don't, I have something I want to tweet. Like I learned—

But that's a conscious thing. Like you historically might have like, all right, like let me write the tweet and then let me review the tweet. Now you're, now you just fire it. So why is this whole, why is all of this, you spell the word like R, just the letter R, like not AR.

That I've always done that.

Okay.

It's just how I chat with people. But I just feel like some level of authenticity is valuable. And I just tweet my gut. Like, I think of something interesting, I just tweet it out.

I think the person who does this really well is Levels. Do you know Levels.io? Yeah. Yeah, your style reminds me of Levels.

He's the poster child of vibe coding to like $2 million AR.

He's the original vibe coder. Yeah, exactly.

You're the original coder.

Yeah. Um, one thing that I think is going to be interesting, I'd love to get your take on, is how brand and distribution actually changes, uh, over the future. And so my theory is that we're talking about Delphi. I don't know if you saw when OpenAI came out with Sora and you could clone other people and put them in your videos. Yeah. But in your video, like Mark Cuban was the, the main example of this. If you put Mark Cuban in your video, it automatically somewhere in your video said, uh, cost— go to costplusdrugs.com, which is Mark Cuban's website. So how do you think about how like brands and individual personalities will change as, yeah, honestly, people start getting cloned?

I mean, Mark Cuban would probably be very annoyed to see it.

No, Mark Cuban signed up for that. Mark Cuban opted in and said, if you clone me, ChatGPT, I'll let all your users clone me, but somewhere in the video it has to say Cost Plus Drugs. So I did a— I made a video on Sora. It was like me playing Mark Cuban at the US Open.

Yeah.

And the video's like us hitting back and forth, back and forth. Someone, someone wins the shot and he goes, oh, great shot. Go to costplusdrugs.com.

So I thought about this maybe a year ago, well, both during the NYC election and the general election. I figured that at some point every political candidate are going to make use of AI and just mass produce AI content that is individualized, customized to every person. They know exactly or with a fairly high certainty who you are, who you might be interested in, what your bias is, and they would produce the same content that you would want to see. And so this goes back to your point about marketing and branding. I think it's pretty clear that in the near future, marketing and branding is going to be individualized to each person.

Yeah, it already slightly happens. Like if I saw an ad for for me the other day on LinkedIn. It was like, hey Jason, like, as one of the co-founders of Blockworks, like, I really think you'll enjoy Rippling.

Yeah.

Interesting. All right.

Yeah. Um, and this is a bull case for, for Facebook, to be honest, for Meta.

Yeah.

Um, AI is going to make their algo, uh, for advertisements.

The ads are going to get so good. Actually, I saw it the other day. I typed in, um, do you know what HiRox is?

No, what is that?

Like a CrossFit marathon type thing. Anyways, don't go down the rabbit hole. It's cool. But I looked up— I'm training for HYROX, and so I looked up on the App Store HYROX fitness app, and I got these, these apps, these fitness apps that were like fully like HYROX training. And HYROX is a super niche, weird thing, like not, not big enough to make a bunch of apps around it. I'm like, oh, that's weird. Then I was like, oh, I wonder if these apps are HiRox apps, or if they just targeted, if they had built fully customized HiRox ads for people typing in HiRox. Sure enough, these are just normal fitness nutrition apps, but they highly targeted. So I think that will happen much, much, much more at scale.

Yep.

So great. Um, you were, I listened to the podcast with you and Imran a month or two ago. Um, you were talking about the markets and how you're fearful, but also in, you know, the dot-com bubble and someone, something like in 1996, someone called the top and then it went on for 3 more years. I'm botching it, but something like that. Um, you had a tweet the other day, it was talking about, uh, the forward PE ratios versus the subsequent 10-year returns. Um, and then you said, yeah, but here's, here's the— and obviously they were down. Uh, but then you said, look, the counterargument is like A, B, and C. And then you said, I remain, I remain scared. So maybe we can dig into the— those 3 words, I remain scared, and how you're thinking about markets today.

I remain scared because I think for the most part, most people are fairly euphoric. I wouldn't say it's like 100% there, but you see screenshots of P&L, year-end P&L every day for the last week on Twitter. And then that's the sentiment part of it. And then if you look at the quantitative aspect, valuations are near historical highs, but then you can make the argument, argument that companies are a lot better these days. Their profit margins are a lot higher, they're a lot more durable, much bigger moat. And then that, plus on top, on top of that, you have all this money printing, debasement. So I can make a case that the current valuation is okay, but you can't deny the fact that it's near historic high. So I'm not all in on risk assets. I'm in fact probably 40% in cash. I know it's going to be painful if the market keeps going up this year, but I sleep better at night.

Yeah.

I have no predictions on the market. It's really, really tough. It's not obvious to me.

Yeah, yeah. I was listening to a Druckenmiller interview. He said, in my 50 years of investing, this is the hardest time in my entire life. Forecast what's going to happen in the markets. So that made me just say, if Druck can't get it right, I'm not getting it right.

I just trained myself to become an individual stock picker because if you look at the market overall, it's very high. But if you look at, if you study the individual companies one at a time, occasionally you can find attractive opportunities. And if you do, then you're less scared about an overall market downturn, like in 2000, in the, in the dot-com bubble, tech companies went down, uh, for the next 10 years. But the small caps and value companies actually did really well during that decade. Like, they went up like maybe 10% per year. Um, I think this is one of the best times for individual stock picker.

How do you think about buying stocks versus buying crypto tokens right now?

Um, I don't feel that crypto tokens are particularly attractive yet. There will be a time they become really attractive, similar to what we had in 2022. So right now I'm just focusing on— again, I analyze assets individually rather than asset classes as a whole. If you look at asset classes as a whole, probably Nothing is attractive at this point except for the US dollar. Yeah, if you look at assets individually, you might be able to find something here and there, but right now in crypto, I, I mean, I still hold Bitcoin, um, and a bit of like, like tokens here and there, but nothing, nothing really big. Yeah.

Okay, so if we try to break down your portfolio, I don't know how in-depth you want to go here, but you know, 100% portfolio, 40% in cash, that leaves 60%. What of the 60% is stocks and what is crypto? And what if we can bucket it as there's Bitcoin, then all other crypto assets.

Yeah.

And then stocks.

Between stocks and Bitcoin is probably 50-50. Like, I still have some Bitcoin, a lot of Bitcoin, but this is not the ideal portfolio I want to have because if I sell Bitcoin, I would have to pay a lot of taxes. Um, but crypto ex-Bitcoin, it's very little. It's probably less than 1%.

Less than 1% of your portfolio is in crypto tokens?

Yeah.

Wow. What was it like a year ago, or at like at the highs?

Uh, at the highs it was like maybe Bitcoin plus like everything crypto was probably 80%.

Wow. What—

but that's just being financially irresponsible. Like, that's just—

yeah, yeah, I know, I know, I know. What stocks are you in?

Right now, still mainly the biggest one is still Google. I have Tencent, it's a really good company. No one talks about this stuff. Like Tencent is a boring old Chinese company, but it's a really, really good company.

You like Tencent more than Alibaba?

Way more right now. I think Alibaba is ahead in AI, but their retailer business is very competitive right now. I like Amazon right now just because I feel like the valuation isn't super attractive, but I think the sentiment was pretty bad last year 'cause all the MAG7 went up a lot except for Amazon. So I have some Amazon and I like the robotics story. By the way, all these things are not, for, for this year. Like all these things are 10-year hold. Yeah.

Amazon, I think, is leading on a lot of the robotic stuff. They'll be the first big company with more robotics employees than human employees.

Yeah. Their, uh, human headcount over the last 5 years was flat and robot, uh, headcount has been growing like 20, 30%.

Their human headcount was flat. That's nuts. That's nuts.

That's how they were able to increase their margin. A lot over the last few years, even though the revenue didn't grow.

What about the, um, the GLP? Like, what about in healthcare, like an Eli Lilly or someone like that?

Um, I still hold a bit of Eli Lilly. Uh, I just, I just don't understand. Like, Eli Lilly is a trade for me. It's not a long-term hold because I don't understand it enough. I don't know what's going to happen in 10 years. I know that in 10 years I'll probably still use Amazon, but in 10 years— the, the thing with, um Pharma is— they have too many competitors. Yes, they have a big patent, but you have all these like gray market Chinese peptide, you know. Yeah, very cheap and very effective.

The peptides are hot right now.

They're very hot and they're directly competitive with Eli Lilly. Yeah. Um, it was—

you should listen to, uh, there's a Cheeky Pint episode. I don't know if you've listened to that podcast, but it's, uh, uh, John Collison, one of the founders of Stripe. One of the Collison brothers hosts a podcast with the CEO of Eli Lilly, and they end up talking about the Chinese knockoffs.

And yeah, it's interesting.

Um, all right, two other—

I think one of the biggest, um, uh, overlooked areas in, in the markets right now is AI-enabled biotech. I think everyone is thinking about robots, drones, like AI chatbots, all that stuff. But I think one of the biggest areas where AI can make a huge impact is, is bio.

Yeah.

And that's what DeepMind is working on and so on and so forth. I just don't have a good thesis where—

I was gonna say, I agree. Theoretically, I agree. I just have no clue how to— yeah, I have no clue how to do it. Um, okay. Two head-to-heads I'd be curious to get your take on. One is in the car space and one is in like the fintech space. Um, within cars, Tesla or Rivian? Hold either of those?

No, too expensive. Tesla is too expensive. Rivian, I just don't know enough. But I love my Cybertruck.

You have a Cybertruck?

Yeah, it's by far one of the— it's the most important investment I made for myself because I hate driving in the last 5 years.

It's too big. It's too clunky. Why didn't you just get another Tesla?

When I bought the Cybertruck, I think it was the only one with the hardware for HW4. So that's the thing that enables the best best-in-class self-driving. But now there might be more models available.

Yeah. Okay. Robinhood or Coinbase? Or neither?

Between the three? Robinhood, Meta, or Coinbase?

No, Robinhood or Coinbase.

I think they're both going to do well, but purely from a valuation point of view, Coinbase is more attractive. but neither is super attractive.

Yeah. Do you own either of them?

I own Coinbase. Yeah. Robinhood is too expensive right now.

Yeah, I just, I'd been in the Robinhood trade for a while and I just sold some. Yeah. So, what, how do you think, I saw your tweet about like you finally realized what Coinbase's endgame is.

Yeah.

Maybe tell me about this.

So I tested the Coinbase One product when they launched it. It's like a membership product. You have to pay like, I don't know, $200 per month or I forgot the fees, but you get an Amex card and some other benefits. But the main benefit for me that I enjoyed is the Amex card that you can earn 4% in Bitcoin on every purchase. By the way, I think this feature is uniquely enabled by crypto because with stocks you can't do buy a tiny amount of a stock, right? You can't do fractional shares of a stock, or maybe it's hard to do, but with Bitcoin it's very easy to do. I really enjoy this feature. So I spend a lot of, I use my Amex card for a pretty good chunk of my purchases. And then I realized, oh, by the way, so the Amex card, in order to get the full benefit, you have to deposit $200K or something in Coinbase. I haven't used Coinbase for years, but that feature just got me back into Coinbase. So then I realized, oh, actually Coinbase wants to become a bank. They want to hold the money for you and offer you all sorts of financial products. And this was before their big announcement last month where they expanded into stocks and—

The Everything App, prediction.

Everything App. Yeah.

Do you think that'll work? Do you think someone builds a predict— like, do you think someone builds the everything app in the West?

Everything like messaging is— I think everything financial app is happening. It's happening with Robinhood, but everything like Tencent WeChat, I'm not sure. I don't have a strong thesis. Yeah.

The thing I wanted to end on is a bit of a, just a personal interest in this is you've been on a health journey for the last couple of years. I saw you just hit a VO2 max PR. Yeah. How is it like, what have you found to be the most, where have you gotten the most outsized returns on the health front?

It's the big three. It's diet, sleep, and exercise. I was really into the micro-optimizations in 2021 with all these supplements and your sauna and all that stuff. But then like after 4, 5 years of personal experience and digging into the research and watching a bunch of podcasts and stuff, nothing is even close to just getting 8 hours sleep, eating a relatively healthy diet and then exercise. Like 4, 5 minutes a day.

Yeah, what do you do for diet and what do you do for exercise?

I don't even fully optimize anymore. I think it's too stressful. For diet, I You know, I just, I mostly eat healthy. I just mostly eat—

It's like the guy in the middle is like doing all the food prep, doing all the like taking like 17 supplements in the morning. It's like eat healthy, eat healthy.

What I realized is that if I try to really optimize for everything like Brian Johnson style, I just get really stressful and the cortisol itself is not conducive to longevity. Yeah, yeah, yeah.

I tend to agree with that. You should look into High Rocks. I bet there's a High Rocks in where you live.

Yeah, you should do it.

Um, Chow, this is awesome, man. Congrats on everything. Um, any, like, anything to plug for Alliance right now? Do you guys have a new— are applications open?

Um, applications is always open, and we try to make a decision within a week, so anyone is a bit— is welcome to, to apply anytime. And they should hear back from us very quickly.

Nice. Great. Well, ciao. Awesome, man. Thanks for coming on. Good seeing you.

Yeah, take care.

Bye.