Katie Haun (Haun Ventures) at the Machine Earning AI Summit

Well, I'm glad you mentioned my background because the lens at which I look at these opportunities through is not the same as a lot of other VCs. As you noted, Eric, the entire first chapter of my career was in law, was in government, and I learned how to work within the government through that chapter. And you know, now we have a fresh billion dollars raised of capital to pursue the opportunities that we're calling the machine economy and the future of finance.

Yeah. So, here we are. And what are those opportunities? We see kind of three buckets. First of all is new financial rails, new financial infrastructure. You know, let's be honest, banks, the stock market, everything was built for an analog economy. It was not built for the global 24/7 digital always-on economy. And so companies like in our portfolio include companies like Arabore. I don't know how many in this room have heard of Arabore. Put up a...

Yeah. So, we've invested in Arabore and we're the second largest shareholder in that company. And that's a bank, but it was opened on a Sunday. It opened on a Sunday. It launched on a Sunday. Gone are things like banker's hours. I mean, that's actually a thing, right? A banker's holiday, a banker's hours. Wire cut-offs, things of the past. Why in this new global always-on machine economy would wires be cut off at 2 o'clock for those of us on the West Coast? So it's new financial rails like that. The second category of focus for the new fund I would say is let's call it new assets and new markets. And when I'm talking about new assets, you can think of things you've heard of like stablecoins. You know, stablecoins now the volume of stablecoins exceeds the transaction volume of Visa. And that's incredible considering that stablecoins have only been around for what, seven years now. I was on the board of Coinbase when stablecoins were literally invented. And now they exceed the transaction volume of Visa. And if you can digitize a dollar, it turns out you can digitize a lot of other financial products and instruments. So like the stock market, equities, and just last week we had the SEC proffer an innovation exemption. So we're going to have tokenized equities now. And again, stock market hours gone. Now it's 24/7. And importantly, anyone with a smartphone around the globe can now access the US equities market. And that's not true today. So that's the second category. And then let me just quickly get to the third category. The third category is what we're calling the agentic economy. And that is quite simply what does the world look like when the end user of a financial product or a financial service is not a human but is a machine. And I think it looks really different,

Well, first of all, let me talk about the parallels between AI and crypto. I think a lot of people think it's one or the other, but the reality is they're very synergistic. They're very synergistic also just in the talent pool. I mean it's highly technical talent, distributed systems, PhDs, machine learning. So cryptography and AI actually in many ways go hand in hand. And in fact, like, look, we had David Sachs who's the crypto and AI czar. They've put that role together and I think they've done so for a reason. But in terms of parallels, I see a number of them and a number of lessons learned. Cautionary tales if you will. We'll get to those. But what are the parallels? First of all, people talk about now what's the right frontier model for the task or do you even need a frontier model for a task? And it reminds me a little bit of the debate on what L1 blockchain would be good for this transaction. You know, if you were going to send millions and millions of dollars around the globe, you might have been content to pay the relatively high transaction fees or put up with a longer block time settlement of the L1 that was the Bitcoin blockchain or even the L1 that was the Ethereum blockchain. But if you were going to, for example, trade meme coins or NFTs, you might not have been willing to pay those high transaction costs and wait that long of time for settlement. So that gave rise to other architectures. Things like the Solana blockchain came along.

Or you talk about efficiencies and we think about the efficiencies for a basic task. Do you need the new cutting-edge frontier model? Is the compute there really efficient? And so what do you have right now? Well, like we had with the explosion of alt L1 blockchains, now we're seeing new lightweight frontier models. We're seeing the emergence of Neolabs, which by the way brings me to my second parallel between... there are four that just kind of are top of mind. So that's the first is the technical architecture and improvements in scalability, efficiency, throughput. The second though is valuations and the valuations right now that we're seeing some of these... I don't want to pick on... but I mean, you have very highly technical, very well-respected technical, amazing teams that basically still don't have a product. And it's reminiscent to me of the times in 2017 in the digital asset marketplaces where we had these... do you remember the ICO craze? ICOs...

Right. Teams like... have you heard of them anymore?

No, actually it was we won't launch the coin. We'll just raise billions of dollars first for a foundation. And so we had teams like, not to pick on folks, but like EOS or Tezos, you know, raise billions and billions of dollars and like you said then launch a coin, initial coin offering. And really what was the need for that alternative L1? What was the differentiating feature? And by the way, Eric, I mean we did invest in some of those alt ones and some of them have performed remarkably well and some of them haven't. The point is that there will be some winners, but there will be a lot of losers.

I'm not. Right now, we're just taking an early look. I think we're in the earliest innings. I'm just saying that I have no doubt that some of the Neolabs will be huge businesses, but I also have no doubt that some of the Neolabs will not be huge businesses. Much like you needed some alternative layer 1 blockchains. We saw Solana. No one had heard of that. Now, it's still around. It's still a thing. A lot of transaction volume still there. But I could give you for every one Solana, I could give you probably 30 other alternative blockchains that you've never heard of and that are kind of zombie chains.

The first two. The other is the public backlash. Now the public backlash I think with AI is even greater because there's more eyeballs on it, more people are paying attention to it. The public backlash that the crypto markets faced... I think not as many people were paying attention to it in like 2013 to 2017, but if you recall back to 2013 to 2017, it was all about Bitcoin is bad for the environment. It is using up all our energy and Bitcoin mining is really going to be a real problem for our planet. And now we're hearing the same kind of public backlash with respect to, for example, data centers. I mean, not even about the job loss. I'm talking just about data centers. What is the argument? It is energy expensive. It's going to make our energy costs go up and this use of electricity and computers really bad. So that's another parallel. And then I'll end on government engagement or lack thereof. And I think the AI industry broadly speaking is taking some lessons learned from crypto. From crypto you had this kind of... well because it was started as almost like a crypto anarchist movement, there was not a lot of government engagement in the early days of crypto. In fact, some of you might know I was actually asked in 2012 to shut down Bitcoin for the government. Obviously that wasn't possible nor do I think it would have been desirable. Instead we thought it was digital breadcrumbs...

I was.

By the powers that be at the Justice Department.

Not initially. I had to learn about it first. And what I learned about it was the Bitcoin blockchain was a remarkable tool for the government for digital breadcrumbs. Like there's no...