A (2:59)

We have a promo code. People can get a good deal, I think on a plod. Yeah.

B (3:02)

If you go to P, L, a, u, d, AI plod, AI twist. Use the code twist. Save 10%. Jason. I'm using mine to take ideas for blog posts when I'm out and about.

C (3:12)

Love it.

B (3:13)

And because I had the wristband, my children can't yank it off my body. So not really a clip on guy yet. But give me like five years to have older children and then I'll be able to, you know, safely wear it like that. But yeah, I'm loving it. Lon's loving it, you're loving it. Kind of a hit here at Twist hq.

A (3:28)

Everybody's loving it and thank you to them for supporting the program. If you use that URL they know we sent you and tell your friends about it. Shout it out on your social medias. Big show today. Got a couple of great guests. Why don't we get started?

B (3:41)

Alex, we have Elizabeth Yen from Hustle Fund. If you don't know Hustle Fund, it is one of those venture capital firms that actually will invest in you very, very, very early. Or as they put it, they invest in hilariously early startups. Think founder, university, if you will. Jason. People with an idea but not yet product, market fit or anything that far along. I love their fund. I know her co partner Eric, and some of the nicest people in the entire world. Elizabeth, welcome to the show.

D (4:08)

Thank you. Thanks guys.

A (4:09)

It's been a while. It's been a minute. So good to see you. How has this AI revolution in the last two years or so impacted what you and I do, which I call now Year zero investing from the front lines. How is it changing how startups are created? Who's creating them, how fast they grow, product, market, fit, whatever. And maybe even how you run your firm.

D (4:33)

Oh my gosh, it's been crazy. Well, first off, my own workflow has changed. Like on my computer over here, which you can't see, I have multiple terminals going cloud coding away, press 111 and just playing with all the tools and using them a lot. Openclaw, cloud, cowork, cloud code, et cetera. So internally we built a lot more to help us run our operations, but I think that also has informed a lot of our investment decisions as well. Everybody can be spinning up these apps over a weekend. And so now a. The amount of capital you need for software has gone down dramatically. So there's this question of, you know, why do people even need us? That's the existential question. But I mean, we have noticed actually companies get faster growth and so they need to pay for their compute bill. So we're still needed on that level. I don't know if you've ever seen your own cloud bill or mine, but like, there's money that is needed.

A (5:26)

But what are you up to? What was the biggest day slash month you've had? Have you had $1,000 a month?

D (5:31)

Uh, not quite there, but getting there, yeah.

A (5:33)

Yeah. So 700 bucks a month.

D (5:35)

Yeah.

A (5:35)

You've had a $200 a day.

D (5:37)

We'll probably, we'll probably cross that this month in March. Over a thousand. Yeah, yeah, yeah. And I, I just wrote my CFO note. Like, it'll probably be more than 3,000 by the end of the year. Per month.

A (5:49)

Yeah.

D (5:50)

Crazy.

A (5:50)

It is nuts. And I share too, your observation here. You have to ask, well, what is the role of venture capital? And people have been saying that for a long time.

D (6:01)

Sure.

A (6:02)

It does seem like, you know, the cloud computing revolution, the wework revolution, getting rid of those big ticket items, Alex, that we used to negotiate. 100,000, 200,000 for the lease of your office to your commitment. You know, used to stand up 20 years ago, your own rack of servers in a colocation facility. So that was 100k in servers, probably 50.100k plus 10,000amonth, 5,000amonth. Plus you needed a sysadmin to do all that work. All that got abstracted away. Okay. GitHub, you know, digital AWS, whatever you know, you're going to use. But now this is different because those two or three first developers you're going to hire, maybe, okay, yeah, maybe we don't. Maybe we hire one, but we don't need.

D (6:48)

Yeah, we don't need them. At least not in the beginning.

A (6:51)

Right. So that's a big sea change.

D (6:53)

And then what is the moat that's the other question everyone's asking. Right. So it's like, okay, you back this company, they're doing well. But you know, I've had a couple of companies, Jason, where they've gotten to 10 million ARR and then somebody else comes along and does the same thing and it's like sort of game over.

A (7:08)

Yeah.

D (7:08)

And that never used to happen. So that's, that's sort of the other thing that everyone's thinking about. So we've been looking at a lot of infrastructure rather than apps.

B (7:18)

So you're going one layer down the stack essentially to prevent from being vibe coded out of existence.

D (7:23)

Exactly, exactly. We've been looking at more hardware than ever. We didn't used to look at so much hardware. Now we look at a lot of hardware for the same reason. So. So I don't know, I think in some, some ways that nothing has changed, but everything has changed. Right. Still looking at the same fundamentals. Customer acquisition is important, but you know, how, like, how are you going to kind of maintain your position?

B (7:48)

Do you pay more attention to Churn than you did before? It's always been a very important thing, especially in the SaaS era. People wanted to have net positive retention and so forth. But if someone can show up to your 10 million ARR business and crush it, is Churn now even more important at the earlier stages of vetting a startup?

D (8:05)

Yes and no. So there have been some high flying marquee companies and I won't name names, but we all know them in the sort of Vibe coding space that have had awful Churn numbers and a lot and some VCs have been passed because of that. But I think one of the things that will change is as the AI models get better and that's what everyone is riding on, then they will ride that wave to getting more accurate and being a better product and that will actually help your Churn. So there are some things you can control and other things that I think the LLMs will control and make your product better over time.

B (8:38)

Are you referring to Hateable and Zapwitz?

D (8:42)

Yeah, yeah, maybe.

A (8:44)

Well, you know, I kind of think we've seen this before. People did do a lot of sampling, you know, in mobile when the PC came out, there was a lot of sampling. People would do it as a hobby. They would go to CompUSA and buy a bunch of package software. Internet comes out, people would, you know, pay for websites, apps came out, people would pay for a flashlight app. So there's a lot of sampling coming out, but it will come out in the. I think in the end, the thing that I think is also pretty tremendous is the ability for founders to get better at the job of being a founder by learning new skills very quickly. So if you were a founder, you didn't quite understand the legal issues, or maybe you didn't understand cap table issues, or you weren't good at go to market strategy, you're sucked at design. Whatever you sucked at, you will suck a lot less instantly. And then you might even become good at it quickly. And then if you can be good at, you know, the, the dozens of things it takes to do the chores at a startup. Man, that's a big breakout. Yeah.

D (9:47)

A hundred percent. And anybody. And so that means that everyone's bar is higher. Like everyone has to level up and anybody who is not, then it's just gonna get harder to get funding.

A (9:58)

Yeah, everybody became a superhero overnight. It was like somebody got the serum and everybody becomes a mutant and has crazy abilities. So it doesn't matter if you used to be a mutant, now everybody is.

D (10:10)

But I don't know if you're seeing this, Jason. There is a bit of a bifurcation. Like when we see startups, like there are some people who are just exceptional and using all the tools well, and then there are people who still haven't really gotten the memo, whether it's for coding or whether it's for crazy, you know, general business ops or sales and marketing. Like there are a lot of people who have not leveled up. So that that playing field is changing as well.

A (10:34)

Debugging sucks and it takes up time. Your team could be spending on awesome new features and products, but now there's a better way. Sentry's AI powered debugging agent, Seer, isn't just guessing about what might have gone wrong with your system. It's analyzing your actual data. And because it has context, it spots buggy code before that code ruins your entire day. Plus, Sentry works alongside coding agents like Cursor, passing along an idea for a fix that can then be applied to your code base and sent to a human team member for review. End to end automation, all the way from bug detection to pull request. So join the millions of devs and companies like Claude and Disney who use Sentry to move faster. Check them out at S E N T R Y IO twist and use the code twist for $240 in Sentry credits.

B (11:27)

Is that founders that are seeking their first check from your fund, or are you thinking about companies that are more like Series B, Series C and aren't evolving to meet the moment both.

D (11:37)

So we have definitely had conversations within our existing portfolio where the companies are further along, they've got a built out team, and then now it's like, okay, we're kind of in this weird dichotomy where maybe some of their existing engineers are not going to make it. And that's, that's actually really tough. And then there's also this dichotomy of okay, I have these people, but now who do I hire next? Do I hire someone? Do I have to kind of change what this person's doing? So actually in some ways it's a little bit trickier I think to be an existing startup with all of this sort of legacy stuff. And now you're changing versus, oh, I'm starting from scratch as a one person shop.

A (12:14)

This is what we're seeing at the big companies too. You know, if you had an overhang of middle management or too many PMs, product managers, you're kind of like, well, product management was a lot of note taking. And like, yeah, and now it's, we just talked about Plaud, like you don't need a note taker, you just get, literally you spend a hundred dollars on applaud and you're done. Like we're here folks. So that note taking role or just every day checking in on the changes, that is kind of over. And if you're a founder who doesn't embrace this, you're at risk. It would be like being the founder who doesn't embrace cloud computing during that change or you didn't embrace mobile. So just pause for a second, right, Elizabeth, and think about our friends who had an incredible website. They never figured out mobile, they had an incredible, you know, business for SaaS or whatever, but they never figured out cloud. You know, and like Dropbox had to make that change, right? Like, okay, we're not going to stand up our own servers, we're going to use storage, we have to put more money into the application level, yada yada. All right, lots more on the docket. Let's get to work.

B (13:20)

All right, so the biggest startup story of the last couple of days has to deal with a company called Delve. Now if you don't know about SOC2 compliance and other bits of enterprise arcana. This is a little bit esoteric to you, Jason, but the issue is, did a company, did a startup that was backed by both YC and some major VCs essentially cook their entire business out of thin air? As in, did they do some fraud? So we're going to bring on a founder, Ryan Madhavi, he's the founder and CEO from Seal, which is from launch accelerator cohort number 35, Jason. And Seal does automated compliance very similar to what Delve claimed to do. Ryan, welcome to the show.

E (13:56)

Thanks for having me. Good to see you guys.

A (13:57)

Yeah, Ryan doesn't have an axe to grind here. This is your straight up heads up competitor. This story is crazy. And there's also like Vanta, which came before this. So this is a pretty wide space. And they came in with the concept, Alex, of disrupting the Vantas of the world and using AI to make this go faster.

B (14:19)

Yeah, absolutely. The idea was to bring AI and automation to a very boring process. Compliance is kind of like doing your personal taxes, like you have to do it, but everyone absolutely hates it and it always involves more paperwork than you think. So why not use tools to go faster? Well, if you fake it, you can go really, really fast. So Ryan, I want to start with just a very broad point here. You've seen all the materials, you've seen the sub stacks and the leaks and so forth. How credible are the allegations against Delve from your professional standpoint and what are they?

A (14:49)

Let's explain to the audience what the actual granular accusations are, if you have them there.

E (14:55)

Yeah, yeah, there is a lot to unpack there. It's quite in depth. I want to just start off by saying that this is a very genuinely unfortunate situation, not just for Dell's customers, but the entire industry that effectively runs on trust. That said, whoever put this investigation piece together, bravo. Level of forensic detail is impeccable. And like I said, there's a lot to unpack there. We'll try to do the TLDR for the audience, but I'm curious whether you guys think this is a disgruntled customer or a coordinated hit piece. Either way, these are very serious allegations which if proven will likely have criminal implications. So on a high level, there was an incident that happened, right? This incident happened December. There was some leakage since then. The biggest problem here is 500 boilerplate reports, same errors, logos just simply swapped of those 259 Type 2 reports, which is effectively the more rigorous version of SOC2.0 auditor findings across every single client, which is almost statistically impossible. And this has affected brands like Lovable Bland Clulee and then Duo's Edge, which is a public company. And of course subsequently they had that Supabase Bucket League which publicly exposed background checks, stripe tokens. I mean, it's getting messy. So that's the latest. And keep in mind, we're at part one. Part two is set to drop this week, which, you know, we're all unfortunately grabbing our popcorn for. But this is what's happened at a high level.

A (16:24)

Okay, so Alex, just a step back there. Somebody released these allegations in a substack anonymously, am I correct?

F (16:31)

Yes.

B (16:31)

They went by deep delver. There was a leak of customer information. The company said that's all fake. They went through the leaked information and said, hey, this looks incredibly fishy. You know, copied errors across different documents from different companies. A web of auditors that appeared to be, I would say, less than above board, if you will. And as Ryan mentioned, just a lot of indication that these were not personalized reports that indicated how a company actually did, but were instead essentially compliance theater.

A (17:00)

So there was a leak of data and then somebody did a sub stack based on that leak or the, the, the, the sub stack is the leak. Just so I'm clear.

E (17:09)

It's. It's hard to say because the leak definitely exposed a lot of information. But there's so much detail here that, and that's, that's the piece where it's like there has to be a collaborative effort. It seems like that leak alone doesn't necessarily paint the whole picture. Like there's, it's very, very detailed. I mean, you know, look, we, you know, see, this is a broader problem than just delve. We can get into the details of specifically how this goes into delve, but there's a difference between hustling and faking it. And to be fair, there's a lot of over promising and hype across the board in this space and maybe AI in general. But 500 boilerplate reports, nearly half of which are basically identical. Zero incidents. It's, you know, this is not cutting corners. It proven true. This is, you know, gaming system.

A (17:56)

Yeah, it's pretty clear from what we saw in it that they, if it's true, because this is all allegedly. If, if this is true and they did those reports and they just changed the logo on them and it's like AI slop or it's cut and pasted. That would be, you know, like definitionally fraud. And you could have gotten your customers into trouble with their customers. So there's like lots of second order impact here. But somebody leaked this. Was it an internal whistleblower? Is it a somebody who is a customer of delves? Those are all the questions. But there is an underlying issue here, I think, Elizabeth, that you and I deal with the area of Y Combinator specifically. And you know, startups generally can attract hackers. In fact, Elizabeth, on the application for yc, they asked this very specific question which is like, tell us sometime you hacked something. And when Paul Graham talks about, you know, Sam Altman, he was like, oh my God, he was such a hustler, a hacker, you know, got through all these things. There is a hacker culture, white hat, black hat, gray hat in between. So I guess we have to also think how much of this has to do with the culture of Silicon Valley, saying, hey, bend the rules and where the line is there, Elizabeth. So take the audience through what you and I see on the back end when we do diligence. How often you see people bending reality and how often are they just straight up lying?

D (19:36)

It's really tricky because the audience we attract fundamentally is trying to think out of the box, which may push the envelope and break some rules. And in fact, actually for some Geminis, they may literally be breaking the law, but then the laws may change over time. Right?

A (19:50)

Sober.

D (19:52)

Yeah, that's biggest companies admire. Yeah. So I think that that's tough. But I think let's take it from the investor perspective. I mean from our perspective, Jason. Like the companies we look at are so early and we don't often know what is happening on the day to day basis. Like do they actually have auditors? Are these accredited auditors? We don't know, like candidly speaking. So we're just kind of going based on what the founders say, etc. And that is actually what every founder wants, like investors who can move fast and make a decision. So there's going to be some of that. But I think that's also why we need checks and balances in the ecosystem. Journalists who can discover these things. Unfortunately it seems like the regulatory bodies haven't discovered this yet, but I'm sure based on this they will come in and take a look. But that, that is why we have these different parties. I don't actually think it's our job as investors to get into the weeds. No one would ever get a check if we were to get into the weeds because you can scrutinize everything ad nauseam. So that's on the investor side. And unfortunately we like, I think there's just going to be a percentage of every cohort, whether you're YC or whoever, who will end up in a problem like this. And that's sort of the price we pay. But on the founder side, like I think how would you have protected yourself if you were a Delft customer. It is really important to ask your vendors, like, for, you know, their credentials. Like, you could ask whether the auditors are accredited. Like, you know, show me that proof. Of course, they could still make it up. I'm sure there are ways they can fabricate paperwork, but that, you know, helps a little bit. But that's the kind of thing I do think that founders need to, you know, ask their startup vendors, like, how they're doing this, et cetera, and to some extent suss out whether it makes sense.

A (21:41)

Hiring can be its own full time job. And hey, guess what? I already have a full time job. I make podcasts and I invest. But when you're running a small company, we both know every hire matters. You don't want to waste any of the seats you have at your company. And the best partner you can have is LinkedIn Hiring Pro. Why? There's a billion people using LinkedIn. All the great talent are there. If you're proud of your work, you. You build a LinkedIn page and you update it. LinkedIn Hiring Pro is gonna streamline and simplify the entire.

B (22:09)

Jason, tell me more about how to find fraud at companies and what are the red or pink flags that you look for when you're talking to a founder for the first or maybe even third time?

A (22:18)

Yeah, so Elizabeth makes some really good points here. We live in a trust based environment now. There might be some people, you know, who are, you know, in the media or just naive. We think, oh, well, you're an investor in the company, so you're sitting at the office every day and you're listening to the conversations and then everybody submits your work and you validate it, like, not how it works. When you're a minority shareholder under 10%, you have, you know, information rights, which you never call. You rely on the trust of the founder to send you a monthly quarterly update every six months, whatever it happens to be. We like to get monthly in the early days and quarterly thereafter. So you're relying on them to tell you the truth. But we do diligence. And when we do diligence, there are some very simple things we will ask the founder. Has there been any legal action taken against the firm? Has anybody threatened legal action? Have you had an email, a letter sent to you about that, you know, threatening legal action or of a legal nature? So, you know, we really expand it now. We're trusting them, you know, to do that. In our accelerator, we take great pains to explain to our founders, and we love a good hacker. We love somebody who can you know, bend the will of the world to their vision. But we tell people when you are doing something for a selfish reason. So for your own gain. Like, you know, there was an issue where Zenefits was helping people hack their insurance licenses, right? And this was like the big battle at that company. And the founder, Conrad Parker. I got that right. Yeah, he got ousted. He did his revenge startup Ripple and all this stuff. You know, he made a mistake. He got an SEC fine. He got a, you know, slap on the wrist, basically. But he also lost control of his company in that example. The reason why helping your insurance brokers, in the eyes of the law, cheat on their insurance test or just not take it properly, that was to benefit them. Now, if you're Airbnb and you're saying, hey, we believe that people should be able to rent a room in their house, that doesn't benefit Airbnb as much as it benefits the person renting the room and the person renting it for having more options. So you have to put it through that lens. Okay, Put that aside. Then. We also do training. Alex, very specifically around and Ryan, you went through this. Don't ever exaggerate reality. So when you have your customers, like people. I don't know if you've seen this a bunch of times, Elizabeth, they have a slide. These are our customers in pipeline. And we're like, okay, two different things. Customer implies they're paying. User implies free. Pipeline means you put them in a database and you sent them an email. So put that on three different slides. So we'll train founders to do that. Or they'll say, like, this is our pipeline or weighted pipeline. All this, like, weasely language that can unintentionally get you in trouble. I will give you an example. We had a company, they said that they had certain customers. They also said they had certain employees in their deck. They pitched an Investor. Investor puts 50k in it. Company's not working out. They're not getting information from the company. This angel investor went, called the people who were listed as employees, and they said, I never work there. Like, I met the guy, but I never worked there. And then they. He called the people who they said had pilots, and they said, we never had a pilot. We took a meeting with them. He then sent a demand letter saying, I want my money back or I'm going to the SEC with this. They gave the money back. The whole thing blew up. The founder was exaggerating, slash lying while raising money. When you do those two things at the same time, Alex, you Make claims and you raise money. Do you know there's a special term for that that the SEC has?

B (26:18)

Could it be securities fraud perhaps?

A (26:21)

Correct.

B (26:22)

I've heard of that. People go to jail for that. Prison. That's a serious. Okay. But the line, Jason, between exaggeration and lying, it seems fuzzy based on what I've heard from both you and Elizabeth. Can you boil that down for founders so that way they know.

A (26:39)

Go ahead, Elizabeth.

B (26:39)

How to stand the right side, maybe

A (26:41)

with some practical and examples that you obscure a little bit like I did.

D (26:45)

So this last example that Jason gave is probably the number way, one way that we ding founders around, sort of this fraud thing because usually the business is not far enough along such that, you know, they may not even be doing compliance actively at that point when we're investing. So. But the way that we do suss this out is you need to be crystal clear about the, the state of everything. So if somebody is not actually your customer, like you're talking with them, etc. And you say they're your customer, like and we find out about that difference, as much as it may be they may be your customer tomorrow, if they are currently not your customer, we will ding for that because that suggests future behavior in exaggerating in other ways.

B (27:28)

Is that a lethal, is that a lethal infraction? Like if a company does that in the, say one case, would you not fund them based on that alone?

D (27:35)

Yeah, and I've been very upfront and direct with people. It's like you told me on this slide, they're customers. Turns out they're not customers. You're talking with them and then people usually say something like, well, we're so close. They have the contract in hand. It's not totally out of thin air, but they have the contract in hands. Very different from they are a customer.

E (27:54)

Yeah. I have a question for you guys. You guys invest quite broadly and you do this from a diligent standpoint and you have the investor angle much better than us. The way I view it, and I mean for us was pretty clear. A simple 30 minute demo I think would have exposed a lot of so called features that don't even exist. This does not require domain expertise and bringing in someone to do further diligence. So I'm just wondering like what's the. Because we're not talking about just a seed stage or pre seed. We're talking about inside partners and $32 million checks being written here. So I'm just curious like if that

A (28:25)

changes things for you guys, if it was an accelerator that wrote the check, right? You'd be like, okay, it's an accelerator. The product's not even done, whatever. The product's just in market. Hey, you know, and this is like a really good question, you would understand that now. How does a big company make this mistake? How does a big company make this mistake? Like, and how did they make the mistake in ftx? As the stakes get higher, the diligence should get deeper. So if you were writing a 25k angel check amongst 20 angels, you know, Elizabeth and I are doing this. In the old days, there would be a lead who put in 250, and they would do the diligence. But at the earliest stage, again, back to a trust based ecosystem that we live in now. You're ftx. Nobody did any diligence. And the reason was that those founders, specifically that founder, I believe, this is what I believe. And then he went to jail, was found guilty. So the court of law also believed this. I believe he used other people's reputation and the momentum behind crypto to dissuade people from doing diligence and to put pressure on them. So we as investors will deal with this high pressure. Hey, you know, this person's leading. I have this many days, I'm oversubscribed, need the money now, yada, yada, yada. You have to rely on their diligence. And then diligence by proxy is the issue here. People will say, oh, these other people are in the deal for delve for, you know, Theranos. I can use their diligence as a proxy. That's the cardinal sin. People have to do their own. And then the founder will come back to you and say, listen, I don't want you calling our customers. They've already been called by this person. And we say, great, can that person send me their notes? And that's the high level of diligence. So when founders say that to us, hey, we don't want you to talk to the customers. We don't want to burn them out. We're like, okay, that's reasonable. Can you have, you know, Inside Venture partner, Sequoia, Pear vc, whoever's doing an Elizabeth Hustle fund, can you just share with us the document? And Elizabeth and I have, I think probably shared this in the past.

D (30:36)

Yeah, Elizabeth, I think so. For some of these really high flying companies, you may not have that opportunity. They may say you're either in or you're out. Because I have so many people who want in and there's so many late stage investors now that you either want to be in this and just take a risk that okay, the traction's amazing, let's go. Or you say, okay, I'm out. And the late stage investors don't want to miss out because there are very few deals that will really make their fund.

A (31:05)

It's easier than ever to build a great new product, and launching a startup, even as a solopreneur, is getting easier and easier. But as my experienced founders already know, there's a lot more to starting a company than just putting up a website or even building a product. If you're serious about going into business, you need a Delaware C corp that's going to give you a serious leg up on your competition and you can be taken seriously by investors. That's where Northwest Registered Agent comes in. Just a few clicks. They're going to give you the perfect start to your new enterprise. A real identity, a domain, a custom website, a business email, or all the public filings done, and even a phone number. And you'll complete this process in under 10 minutes. Plus you're going to get all sorts of free tools and resources from step to step guides, compliance reminders, and an online account that's going to keep everything organized even as your business grows. So you get all the advantages of a Delaware C Corp, regardless of where in the US you're operating out of. Northwestregisteredagent.com twist okay, so that's the generalist,

B (32:09)

non domain expert take on this. Ryan, you are a domain expert. How early did you think there was an issue over at Delve?

E (32:17)

I mean, we saw the signs like six months ago and look, we haven't gone and I'm actually curious, I want to go back to this point. We haven't spent a lot of money on marketing. I would say basically nothing on marketing. It's all been referrals and inbounds and focusing on the product itself. So that said, obviously we're in the industry, we're talking to other founders, we're seeing customers. I mean, over half of our customers came from incumbent players like Vanta, Drayda, etc. We started to see like bold claims are fine. Marketing stuff is fine. Everyone's kind of, you know, over posturing and doing whatever they need to do to be competitive. But I'm telling you, it was like five or 10 minutes into seeing the platform and saying like we were actually joking around, like this is screenshot native rather than AI native. And you don't even have to be an expert. Like you could just say, hey, These are the features you say you have. Where are they? Show them to me. I legitimately think a 10 to 30 minute demo with anyone who has a baseline understanding of tech would probably catch this. Now, for me, I'm always also skeptical when a company's early growth tactics consist of dishing out donuts and doormats and then just going extremely heavy on growth tactics. I don't know if that's a red flag for you guys. Very early on, I understand aggressive there

A (33:33)

they were sending donuts to people to get them on a exploratory call for their product.

E (33:38)

Yeah, these were what was considered growth hacks. Right. Which, I mean, I can dish you guys out some donuts and doormats today. I don't know

A (33:48)

that hack. I understand the send a pizza. Can we do a call? We'll send a pizza.

E (33:52)

Yeah, yeah. It's like a branded doormat for their offices or prospective clients, you know, offices, something of that nature.

A (33:58)

I, I have seen those techniques from sales departments. They started in the pharmaceutical industry where what they would do is they would call the nurses station and they say, hey, would you guys, we're going to be coming over from Pfizer. We wanted to buy the nurse's lunch. You guys want, you know, Italian or you want Chinese or you want Mexican? And they'd say whatever. And they spend 200 bucks bringing some food or, you know, in the morning some donuts. And then they would come and give them samples and talk to them about the drug. Totally fine. That uses the effect of reciprocity. I give you something, you give me back your time. It's not. There's nothing illegal about it. There's nothing untard about it. But if you are spending too much on that and not enough on the product, that would be where you'd have breakage.

B (34:44)

So Dell went around SF handing out donuts to different people, Jason. And also they use them, per the substack report, as a way to smooth over things. The report says when clients ask hard questions, delve dodges. They demand calls where founders charm, promise and name drop. And when that fails, donuts arrive. So it appears to be a donuts on the front end and donuts on the back end strategy.

A (35:03)

Well, Elizabeth, the did you notice that they used other brands? So this brand washing strategy is what a lot of criminals use. A lot of the most notorious. There was that movie with Leonardo DiCaprio, catch me if you can, Catch me if you can. And he was trying to figure out like, how do I cash a check? There's like this great scene in the movie and it's like, well, who's the most trusted person in the world? Like, we would, you know, and the most trusted industry at that time it was airlines, right? Like Pan America was like. And a pilot for Pan America was like, wow, that's like God coming in and cashing a check. And sure enough, they would cash his check. So that's what happened with ftx. They said, hey, we have all these investors and they bought all those celebrities. And apparently, allegedly, perhaps. Allegedly, Allegedly. Allegedly. Insert that all over the place, Elizabeth, that maybe that's what happened here is they were using those names to try to get people to stand down on diligence.

D (36:01)

That may be true. At the same time, there are so many great, reputable companies also, you know, using social proof of other brands. So I wouldn't use that as like a bad signal. Like, oh, you know, Salesforce is a customer. That's not a bad thing. That's a great thing.

E (36:14)

I think the question is not necessarily the donuts. I think it's like, I mean, Jason, you were here last week, right? Was it last week? I don't know. You guys walk around sf, you see Vanta everywhere.

A (36:26)

Yeah.

E (36:26)

Now you see Delve everywhere. Now Vanta has raised something to the tune of half a billion dollars. These guys are in their, what, series? A 33 million whatever it was, you guys as investors, is that something that you're happy with? Is that something that raises any red flag? Do you walk around a city and you see stuff very fitting like this, I would say, all over the place.

A (36:46)

Yeah. So now let's move to. Elizabeth and I are on the board of the company. Right. And this is another place where things can break down is governance. A lot of folks, including yc, they don't want people. They don't want the investors to have ball control. They would rather see you do smaller rounds. They kind of advise against this. You know, don't have any major investors, don't give anybody a board seat. I think when you raise over $3 million or you've got over a million or 2 million in revenue, you need to have board meetings. Why? There's something at stake here. And you will make a mistake and you might flip the car and you might flip the car unnecessarily because you just don't know how much speed to take on that turn. If we were using Alex, one of your beloved F1 or whatever racing analogies, and that's where, like a good coach and a good team would inform you, like so if you told me, ryan, I'm in a board meeting and they're saying, hey, we've got a million dollars in revenue. We have a $4 million ad campaign, I'd say, let's slow down one second here. Where did we get our. How did we get to a million in revenue? Like, how many customers are they. Oh, it's 50 times 20k. Great. How do we get those 50? And like, we used email and we got word of mouth and we have an outbound sales team. It's okay. Let's double the spend on the outbound sales team and let's see if that works. Pouring like that kind of gasoline that early seems very dangerous. Very dangerous. It is a red flag. But what do you think, Elizabeth? Go ahead.

D (38:11)

That is a red flag. But I would say as a performance marketer, sometimes you can get amazing buys on those billboards, obviously depending on the supply and demand situation. And I don't keep a pulse on that. But I'm not critical immediately if one of my companies wants to take out a billboard. I don't know what. But I think, like, understanding how much are you paying and how much you're getting out of it is. Is something that I think.

A (38:31)

Which is what a board does, right, Elizabeth, you've taken exactly. It's before, I assume, and.

D (38:35)

But I imagine they would have a board.

A (38:38)

Like, I think they had a board at 30. Somebody will check that right now. A Series A, almost always somebody joins the board. What we do in our firm is we say, if we have over 5 or 10%, we want a board observer or board seat, we're not going to take it because we can't possibly take in all those. But we tell the founders, like, Ryan, we probably have this conversation with you, and your attorneys might have said, like, why are you giving them a board seat this early? And we say, no, no, it's an option to have one. And it's only if we have, like, a certain large amount of the. A large amount of equity. And you want us engaged, you want your large investors engaged. Unless you pick the wrong investors or you're doing something where, you know, oversight would be not good. And this is where the whole industry has fallen down with governance. And it's because in the early days, too many founders got ousted from their own companies. And then the pendulum swung to, hey, founders get 100 to 1 shares, you know, at the seed round. And the. That was like maybe six months of that nonsense. But yeah, that would be. Board would come in and review spending and have a Budget and a plan and the accountability. So the CMO would say, yeah, we're going to do this $4 million thing. It's going to generate $12 million in pipeline. We are closing 50% now. Even if we close just 33%, it's going to break even. And that's what we want to do.

E (39:59)

Yeah, I guess the argument there is like, with tech being so commoditized, is distribution something that people are going to start spending more heavily on? I think this is something you guys brought up earlier as well. So to some degree justified. On the other hand, it's like, what are we competing on? Like, is this like marketing agency with a software layer? And of course that leaves room for us smaller players to compete because usually they're falling behind. There's marginal differentiators between different companies. I guess it's a balancing act of how to do that. Right.

B (40:29)

So it's interesting how many things can be high trust and low trust and good and bad. Because if you think about the billboard point that Elizabeth made, Twilio's billboards are like historic SF milestones. You know, Twilio ask your developer. And then if you think about C3AI, they had tons of billboards and that company is an absolute dumpster fire. They, you know, Dell have used a lot of IRL advertising, not so good. But then again, Brex did the same thing because they leverage the arbitrage that Elizabeth was mentioning. It's amazing how many things are just not exactly black and white in this case. And it really boils down, context matters.

A (41:01)

And you know where context comes from. Adult supervision. I'm going to sound like a grandpa here. I'm going to sound like Unk J Gal, but there's a reason why, you know, we have a structure in America that made us the most competitive entrepreneurial force in the history of the universe. Not since the Greeks has there been so much innovation. And you know, listen, us Greeks, you know, it's just in our blood. But Persians, Persians as well, which are, you know, kind of. The Greeks, you know, kind of expanded from the Greek peninsula all the way down. We share a lot of blood, the Greeks, we claim everything.

E (41:40)

Yes.

A (41:40)

We believe everything came, went to Turkey and Italy. Roman Empire came from the Greek empire. That's our story. We're sticking to believe, we're sticking to it. But you know, if we were looking at this just first principles, you need to have some adult supervision in the corporate structure in America. Which is why when you file as a Delaware corp, they have in your charter board Meetings, board members, you know, reports, audits, whatever. You have to define these things. And that's what you work with your attorney on. Hey, we're going to define how many board meetings we're supposed to have. Do we have them or not? These rules came out because of previous lawsuits, previous issues in the marketplace. So the fact that they're bubbling up again is because sometimes in a hype cycle, which we are in peak hype cycle right now, people let their guard down because they're greedy. So the greed index is off the charts right now. The fear index is very low after 2008. Fear index all time high. And you know, greed index very low. And that just keeps happening back and forth, back and forth. And you get weird behaviors during this. But Ryan, just to wrap up here, your company uses AI and tools to automate this. But explain why we can trust Seal when delve. Maybe we shouldn't. How do you make sure that your AI is not producing slop or making errors? And how should customers look at Cl Vanta, whoever they choose to use to make sure that the product is good and it's tight?

E (43:19)

Yeah, great question. I think, look, everyone is shitting on Delft now because they basically got accused of doing something that's fundamentally wrong. The idea was great. I mean, it's something similar that we're doing. We want to make it easy. Soc 2 is not a one size fits all. These guys have a chokehold on the market. And look, you know, board meetings and meeting minutes for a two person startup. And you guys just talked about it. We're going to see companies that are going to be sub five people doing potentially billions of dollars in revenue if we're not headed there already. It doesn't make sense. There needs to be changes. There's a huge need for this. So we have AI tools that help. We make sure that guardrails are in place. We're not cutting any corners. So you have to be competitive on the marketing side, but you have to deliver. So people come to us and they're often a little bit disgruntled. Like why is this taking so long? This is taking longer than what we we were promised. Xyz. Well, this is the perfect opportunity. So we have a huge influx of people already reaching out and we're going to take this one step at a time and not create a leaky bucket like they did because it's going to backfire on you. So for anyone affected, please feel free to reach out to me directly. It's Ryan Seal IO.

A (44:29)

That's my guy. That's my guy, Ryan. That's one of the great market take care of you.

B (44:34)

That's my segue.

E (44:36)

I haven't even started yet. Wait till they give me more money. I'm gonna have posters all over the place.

A (44:40)

Here's what you have to do. Remember, if we give you the first money, you have to prove a second person can get you, give you money, and then we'll come in after them. That's, that's how the game works. We can't be the never ending source of revenue. Although I have done that a dozen times in startups, it never works out well. You need founders to be able to raise money. All right, great job, Brian. We'll drop you off.

E (45:02)

Good to see you.

A (45:03)

Everybody go to Seal IO SEAL IO C E E L. We have our syndicate. We like to share deals we're doing. So you know, we're going to interview a company here that we're considering for our syndicate and would love to have you. Yeah. Give some thoughts. You're awesome. And you can evaluate companies. So let's have our next founder on.

B (45:28)

All right, coming up next is Seb Sheng, the founder and CEO of Brick. Jason, I love this company because we're talking endlessly about power consumption and how to make the overall grid more stable as we add data centers. Well, the idea behind Brick is bringing a piece of technology into a building or a data center, plugging it into key systems and then learn about how that building or data center uses energy to save up to 15, 30% step I think you told me right up to 35% which is an enormous amount of energy, an enormous amount of money. And if this works the way so 7 company think it's going to, well, maybe our grid won't collapse by the time OpenAI finishes Project Stargate. Welcome to the show.

G (46:05)

Awesome. Thanks Jason. Alex, obviously, pleasure to be on. Look, Brick is a AI native energy tech company. What that effectively means is we are effectively building an energy saving agent that as Alex mentioned puts all these energy consuming system from our everyday H vac systems to industrial level chillers or heat pumps on autopilot. We all have that silly thermostat in our everyday households. Doesn't work, makes tons of mistakes. Obviously Cloudbot of latest we all have probably come across Andrew Kaparthi's podcast, all these new smart home optimization leveraging these new tools. Once that bot as such can find every single device on their local network, things can be automated. But really in the enterprise world things are much more complex. So what we do is we have a hardware piece that gets connected to these complex systems in reference and effectively build a software learning the data, everything from system data to environmental data to power data training this AI agent which again we all know is overused notion nowadays, but we have an AI agent that does real work in our case scenario, coming up with all these energy saving rules, strategies introducing temperature thresholds, pre cooling, preheating based on power prices, power prices, so on and so forth. And the end result as Alex, you alluded to 15 to 20% in energy savings on average up to 35%. And decarbonization is just a natural by part of such. So that's really the 20,000 foot view.

B (47:40)

Seb, just to go one level deeper, tell us more about the hardware device that you have to bring on site and then also move past the temperature element of this and tell me about data center because I think everyone is freaking out about how we're going to fuel essentially our AI revolution at the

G (47:54)

end of the day, you know, look, there are different type of devices here as we all know, right? The underlying principle being garbage and garbage out. You know, this agent does not only need system data but environmental data, power data. What we have proprietually internally a brick is a universal gateway. So think of this universal gateway devices, this black box size of this, width of this, it gets connected to a master unit of the H Vac system. Again it's a silly notion of master, you know, slave type notions kind of within the H Vac world. But effectively what that means is the master unit is already connected to these life events. So very much via a one to manning relationship, simple gateway gets connected to the master unit will be able to cover the entire building. So all of a sudden you have a $2,500 device that can cover the entire building footprint that gets all the system data from there like we discussed, right? Whether it's a data center, commercial office building, hotel, we effectively are getting the needy greedy data such as coil temperature, flow refrigerant. So all of a sudden we are not just controlling things via on and off, we're doing much kind of a deeper level analytics. In terms of data center scenario, guys think as follows. It is getting oversaturated, it is overcrowded space. In terms of development, in terms of solutions providers, it's just another data aggregation platform. So it's another dashboard plate. What we need is taking the relevant data set from the existing operating system. In this case a DCIM be able to drive intelligent decision making. On top of that so, so you

A (49:19)

have a hardware device connects to the system, you get AI to analyze it and then give actionable real time decisions that will save money. Who are the customers and how do they pay for it? Are you charging for by square feet? Are you charging for the number of H VAC units? Are you charging by outcomes a percentage of what's saved? How do you think about pricing this product? And Elizabeth, I'm sure you have some questions. Hopefully I didn't steal yours. But we always like to know the business model as investors.

G (49:50)

Good question Jason. The end outcome being it's all tied to the actual energy savings we can generate for our clients. So it's a rapture of energy saving as the core of the business model. We certainly have two bifurcated payment models that we work with clients which are commercial and industrial. So again taking a step back we engage with commercial office buildings essentially managed to hotels. On the industrial side of things obviously we have our internal packing order but for industrial it's really working with industrial manufacturing clients such as Lucid Motors. Tesla will be next in line. Where we've been we've been pushing for a lot of the kind of a sales motion conversations on the front end to the data centers on the industrial too.

A (50:32)

And who are the current customers? You have current customers using it.

G (50:35)

So we've worked with the likes of JLL on the commercial real estate side of things to hotels Paradox Hilton to Sheraton Grant for data centers DTCT Inspir as well as industrial manufacturing such as Lucid.

A (50:47)

So you said worked with. Are you currently working with all those just to be clear or are there previous like one off engagements?

G (50:53)

They're all existing customers to Brick.

A (50:55)

Okay great. Just wanted to parse that language. You can see Elizabeth and I were just talking on a previous segment about making sure we have that dialed in your your questions Elizabeth, just to understand

D (51:05)

where you are life stage wise like where are you? It sounds like you're pretty far along and you have a lot of customers like how much are they typically paying you?

G (51:14)

We're still very much early stage. So look we you know the company was formally incorporated back in March 24th so over the past two years really the year and a half timeframe is on the our actual product development R and D side of things. We kick off commercialization just about a month ago. So within past eight months. Right now our ARR at this stage is about 450 live AR to 400 contracted AR. The difference in that definition being the contractor er, the folks have signed the contract, but yet the full deployment hasn't started. That's the key distinction.

D (51:51)

What does it take to do a full deployment? Because it seems like the hard part is getting these brand names which you have. And they have many buildings.

G (51:59)

Exactly right. So at the end it's a portfolio play. Right. We obviously have to land expand in a sector that's relatively stale, slow moving. It's not only about, you know, adhering to continuous of a direct DTC type of sales motion. It's actually working with the channel partners, which can be The H vac OEMs, which can be, you know, energy efficiency, existing energy efficiency players that are already plugged into these environments that need more of a smarter controller on top of their existing offering. So look for us, at the end of the day, the ACV on average will stand at minimum for 30 to $50,000 per facility on average for each client we'll be working with at least two to three facilities as starter.

D (52:39)

Yeah, I think very directly I've seen a number of companies do this exact same thing and everybody seems to get stuck around the same thing. Like the big logos are willing to kind of try or talk with you or, or whatever. But then somehow there's this hurdle of how do you get JLL to roll this out to all of their buildings. And so I think that's, that's really what my question is. Like you're suggesting you have to go through another third party to get this onboarding or how do you get everyone to move faster or prioritize this, or is this candidly not a priority because they're swimming in money and they don't actually care about saving that much?

G (53:13)

Good question. Let's, I mean here are the two aspects of the, of the, of the fact of the matter on hand, right? On one hand, like you said, slow moving counterparts we're working with, they have been approached by tens or 20 dozens of these energy efficiency players. Really the existing barrier to entry is always boiling down to two things. It takes extremely long to get a deployment started and it takes a lot of money to get deployment started. So it's always going to boiling down to the time and cost of the equation. And for us, because of the hardware capability that we have, effectively all of a sudden you have a $2,000 to $2,500 device that can cover the entire building front end and the deployment timeline is exactly the 46 week pilot that can prove the ROI right outside of that time frame. And we extrapolate that to determine the actual pricing point of our acv. That's how it works. So all of a sudden again the other beauty thing with our underlying software model, it's the scalable nature of it. We have a foundation model for each of the sectors that we tackle. So all of a sudden you have an agent that can repeat, rinse and repeat for a second or third building underlying the same section.

A (54:17)

Lightning round. How long have you been working on the company for? When were you incorporated? When did you have your first customer?

G (54:23)

March 24th we had our first customer. Paying customer. Not just pilot customer, first paying customer in July 2025.

A (54:31)

Perfect. So two year old startup, how many people work at it? Where are you based?

G (54:36)

We have four full time employees and two part time and we are based in sf.

A (54:41)

Have you gone through any of the accelerators yet? Have you raised capital before? Tell us a little bit about the history of the company.

G (54:48)

Yeah, so you know, obviously great to walking through this current round with you guys as we're wrapping our second capital race which is the precede round that we have. So we only have one prior capital race which is the accelerator round that we did with Forum Ventures and Rising.

A (55:04)

Forum Ventures has an accelerator in San Francisco. Awesome to know. We'll make a note of that. So we can trade deal flow. And what makes you, last question, uniquely qualified to pursue this vision for the next decade?

G (55:20)

At the end of day it takes a blend of not just technical aptitude kind of a taking a refreshed lens on the same problem that's been around two or two or three decades. You need the commercial leadership in terms of network and channel partnerships as well as the domain expertise being deeply embedded in the space that we operate in. You really need to think not just from an operator standpoint but but also from the building engineers and owner standpoint. Each one of these has different set of incentives on the line. So it's really about getting everybody on the same page understanding okay, for building owners, they care more about bottom line, OPEX decrease for building engineers, about checking box for technical side of things. And again it's really the blend of those three things that can make Brick outcompete everybody else. And really the rest of the roadmap going forward is about scalability.

A (56:05)

Okay, great Seb, stay on the line. We've got a little surprise for you. I've been obsessed with Tao and Open Claw for the last 50 days. I've gotten very involved in it. We've got somebody running one of the subnets. A lot of people have been asking me, Alex, about, you know, my tweets this weekend and then just the last couple of weeks of having a lot of subnets and bittensor topics here on the program. We'll get into that in a moment after we talk to. Yeah, here it is. Not financial advice. Not financial advice. Family member. What's the style you keep talking about in the pod? Me, sell half your BTC and buy some tao. I don't give financial advice, but maybe to family members, I might give a little bit of advice. Sure, sure. And, you know, it'll become apparent why I'm so bullish about Tao. And I'll tell you about my specific trades right after we have our next guest on. And before we have our next guest on. Alex, 30 days of skiing. Look how svelte I look, Elizabeth. You notice I've lost £40. Thanks for noticing, Elizabeth.

D (57:07)

I do appreciate it.

A (57:09)

I didn't do it alone. Ro co. Get yourself a GLP if you're so inclined and you want to be fit and you struggle with weight loss like I did for a long time. I used to be a very svelte 166 when I was running marathons. I gained two pounds a year while I was married, having kids, building companies. I woke up one day, I was 213. Now I'm a svelte 172 again with a little bit more muscle on my frame. Skiing 30 days, sleeping great. Feel awesome. I'm going into my 50s in my dad bod era with 40 pounds less of fat on my body and I feel great. Did it for my daughters and to stick around a little bit longer. Plus my energy level is twice as much. If you want to check this out, Ro co because they have an insurance checker, you can just check, hey, does my insurance cover this? They're going to tell you, like, lickety split if your insurance covers it. And if it doesn't, they've got a lot of options for you that are very affordable. Very proud of my association with Roe co. And I think we even have a special URL. Yeah, Alex.

B (58:10)

Yes, we do. It's roe co twist. Very simple. Roe co twist.

A (58:17)

Okay, Elizabeth, I don't think you're on a glp. You've been in incredible shape since I've known you.

D (58:24)

Oh, thank you.

A (58:25)

Yes.

D (58:25)

I don't know. I've fallen off the train a little bit, but thank you.

A (58:28)

All right, well, you know, Alex has been lifting. You know, he likes to wear like, a tank top on the show. I banned him from wearing tank tops. Serious here. But he does have great shoulders and definition Please don't call the HR department. If you want to be healthier, go to rogue.com west all right, so Tao is a crypto project. Have you heard of this Elizabeth Tao?

D (58:51)

No, I haven't. I'm not really a crypto scenester neither.

A (58:55)

Neither am I. Although, you know, we've made some decent crypto bets in the household. I have been waiting for it all to get more legal, and I've been looking for a use case. The basic premise I had, Alex, was like, yeah, bitcoin is like this incredible store of value, but they're burning a hole in the ozone layer running this crazy network, doing math proofs. For what purpose? To the integrity of the network. Sure, in a competition, it gets harder and harder to make the next coin. It's brilliant. But we're using a lot of compute. Why? And do we need to use that much compute? Everybody's kind of realized, like, hey, maybe the infrastructure here is. Could be better utilized. Well, something very important happened on the way to Valhalla, which is the road to Valhalla is we need more data centers, we need more compute, we need more storage, we need more bandwidth, we need more services. Distributed services are more resilient and they grind down the cost of things. What Tao does is it creates subnets. Each of those subnets is run by entrepreneurs. They then try to use the Tao ecosystem to define tokens and to give rewards. They stake reward people for doing various projects. What's an example of that? Storage. So I put up some storage in the cloud. I can earn Tao or the subnets tokens by doing this. And today. And then there was one language model you may have seen on all in Chamath brought it up. I didn't even know he was going to bring it up, but he brought it up with Jensen, that they trained a large language model on a distributed compute network. This is the future. We can't possibly have enough. Compute tokens are expensive, as you know, Elizabeth. So imagine being able to get storage tokens, bandwidth, or other services over a network that just grinds down the price lower and lower and lower. And we've got somebody with an interesting subnet here. There's 128 of them right now. Bittensors Covenant 72B.

B (60:55)

Joining us now on the show are Gavin Zainz and Pranav Ramesh. They are the co founders of lead poet Bittensor subnet number 71, Jason. They met when they were working at Nasdaq and now they're working on this great project that uses Bittensor's economics to help companies find the leads they need to land their next one, 50 or 500 customers. So, Gavin and Pranav.

A (61:17)

Gavin Pranav, welcome to this week in startups. Last couple of weeks I've been talking a little bit about my exposure to bittensor and you maybe heard my description of it. What did I get right in my description? How would you describe this bittensor moment and why did you join it?

F (61:33)

In terms of how we look at bittensors, it's really commodifying digital services and products and what that does is allows you to tap into either a lower price or higher quality output. And that's really why we were gravitated towards the network was to build out sales technology that outperforms all the other products out there.

A (61:54)

Now you had to Pranav somehow get one of the 128 subnets, explain how you get a subnet, how you buy a subnet, how you staked, how to get a subnet, how does it work?

C (62:07)

So the way you actually go ahead and buy a subnet is when a subnet becomes available, you need to have a certain amount of TAO that's being demanded by the network at that point of time to purchase it. And then you stake the TAO to that specific subnet number that you want and then you receive the subnet and you can start building on it. It's a pretty streamlined process, pretty easy to do, but you need to actually have the TAO ready to be able to do that, which a lot of people may not have when they first start getting into the network.

A (62:30)

How many you need? 500 Tao, 1000 Taos. Trading at 275 or something right now? Yeah, it's pretty volatile, by the way. It was 205. Last week it was 300. This thing bounces around 10% a day some days.

C (62:45)

Yeah. So the amount of tau itself varies as well, day by day. It's quite volatile, the amount that's required to buy a subnet. So it could range from 200, it could range to all the way up to 500, 600 Tau really depends on the day that you're looking at.

A (62:58)

Now do you give that tow to somebody or is it just to stake and set up your server in your subnet?

C (63:04)

Yeah, so it depends on if you're buying that subnet from another person or if it's been deregistered and now you're just buying a brand new subnet, essentially because it's been deregistered and there's no other owner for it. So if that's the case, then you would just stick and you could receive your TAO if you get deregistered back. But if you're buying it from somebody else, you actually have to give them the tower directly. And it's a one time purchase. Yep.

A (63:27)

And Gavin, are they going to go from 128 to 256 to 1024? What's the plan here for subnet expansion? Is it like the NBA where they do it once every 20 years, or is it more like, hey, we just have more technical work to do on the infrastructure and then we'll release it and then who are the gatekeepers who approve it?

F (63:46)

The first limit after they kind of opened the gates was 64. That filled out pretty quickly. They upped that then to 128. Once we reached around that level, the foundation kind of decided, let's focus on quality entering the network before we just start having, you know, 10,000 projects and, you know, maybe only a hundred are good anyway. So let's just, let's get a solid number of good ones in that 128. Then they're going to start to expand the network.

A (64:11)

So now you wanted to make marketing tools, so show us what you built. Elizabeth, when you're hearing all this entrepreneurial energy going into these subnets and the tokens, what does it remind you of?

D (64:21)

Well, you know, actually it jogged my memory that one of my founders did tell me about this and that is how they are funding their business. They didn't raise that much and so, but I know nothing about this, so it's very intriguing.

F (64:34)

Right now I'm about to do a, a mock request for Brick. I wanted to get some leads specifically out to him. Obviously, you know, I'll send it over after call, after the call so you can actually access and take action on these leads. I'm just going to walk you through the general process. So the first thing that we allow users to really start their requests with is their website to pull some criteria of their icp, some relevant signals. The whole idea of our platform is not giving you a list of 100,000 or 10,000 leads. We want to give you 10 or 100 leads that are actually exhibiting intent and are actually relevant. So first we analyze the website, or you can type in your icp, which is going to be your customer profile, the exact product or solution you have, and then the signals you're looking for. After you submit that request, your leads are then loaded in a pipeline which is pretty much, you know, a Kanban board of A CRM where your new leads are populated in the new lead section. And as you action those leads, you can maneuver them to the contacted responded, you know, did they end up going to a meeting? Are you negotiating with them? Did you win those, those sales or did you lose them? And you're. We're storing in all that data to actually make the system, you know, more smart, know more about your ICP and know really what signals are relevant. So looking at this request that we ran for Brick, ultimately what came up through the system were companies that really have some solid signals around potentially wanting to leverage this technology. Whether that's that they've committed a lot of funds to actually, you know, a green energy program and being more efficient, or it may be that they're hiring some new people in that.

A (66:12)

How did you come up with this lead? I'm sorry, interrupt, but just. It was just done by AI, a human with AI who's on the other side of this, a global group of SDRs.

F (66:25)

What the miners are doing is two tasks. So the first task is them sourcing leads, which they have to submit a lead which is all of these attributes appear, which is the person data and the company data. And then ultimately that's the first part of, of of the subnet is producing that data. We have about 5 million leads in the database, all verified to a higher degree than, you know, Apollo.

A (66:48)

So you have a database and then you have these miners. In the case of a miner, they're not mining Bitcoin like with a high GPU cluster. They're more like doing mechanical Turk work where they're saying, hey, we want to add this lead. And they could be doing it by whatever means they find.

C (67:08)

Yeah, exactly. So miners run a variety of different sourcing pipelines to discover and construct leads. Right now, the most common one that we've seen, if they use scrapers, so they're scraping the web, they're scraping company pages, contact pages, company directories to provide us with the lead data that we're looking for. The second most common one, and this was actually the most common one when we started. Leadpoint is using an LLM based approach. You would plug in a specified prompt which essentially said, give me a thousand leads in the SaaS industry that are located in the US you'd get the names and you'd get the companies, and then they'd use a suite of APIs to enrich that data.

A (67:43)

Got it, got it. So by any means necessary, they get the leads anonymously, put them into the system, and then you have Another group of people who verify that it's a good lead or does the customer verify it's a good lead? Who verifies that it's not slop?

C (67:56)

That's a great question. And we have validators on our network that run the code that we provide that actually do that entire process. And it's not just a single point of validation. We run a consistent point of validation throughout the entire process. For example, the first point is we check whether the email is valid. The second point is we check whether the LinkedIn profile is valid and associated with this lead. The next thing we check is whether the LinkedIn profile is indexed by Google. And if a LinkedIn profile has a low authority, it will not be indexed by Google.

A (68:24)

Totally get that. How much tau would the person who puts in 100 leads get?

C (68:27)

It varies based on the day and the amount of total leads submitted, but it could be around like $200 to $300 of total. $2 a lead, $2 a lead on certain days, yeah.

B (68:41)

All right, so sad.

D (68:42)

So you're paying in tau.

F (68:47)

So technically it pays out in our alpha token which the underlying is tau. And in terms of the average cost per lead now it is down to 3 to 5 cents. When we started a couple months ago, it was around the 2$3 range. But we've really lowered that cost because Bittensor really pushes prices into that commodity pricing and the goal is to continue to push that lower as we've seen more supply enter this kind of free market for data.

A (69:16)

Elizabeth, do you understand how disruptive this is that you don't even know who's putting the leads in? Right. It's just this distributed group of lunatics competing to get the best leads at the lowest price. It's like Mechanical Turk globally off the rails. But you had a second question? Go ahead, Elizabeth.

D (69:33)

Oh yeah. And so just confirming like, then you charge your customers in dollars or how to and you're selling packages of that or how does that work exactly?

C (69:43)

Yeah, we charge our customers in dollars and the miners get compensated who are actually providing the leads in the back end in our alpha, which is essentially can be swapped for Tao.

A (69:53)

It's brilliant.

B (69:54)

What are you currently paying essentially, effectively per lead that you bring in? And would what lead Poet has essentially put together here for you be interesting to brick?

G (70:03)

Our CAC today is effectively zero funded. Less sales. Right. First or second degree of connections getting us in the door. We funnel, we pass the hardware cost through to the customers. Now after the 46 week time frame, if we don't Deliver any savings. No hard feelings. We reimburse our clients back for the money. Now with that tool we just saw, I'll be intrigued. You know, great job. And you guys effectively resurrected my interest in looking at outbound marketing again, which I try. We pay for it, too. We pay about $2,000 for a. For a trial, which didn't really yield any results. So on my mind, it's really about. Okay, before we even talk about cac, it's lead quality.

A (70:44)

So if they got you a thousand leads that was filled in and at result that would cost. A thousand leads at $2 each would be $2,000. You'd spend $2,000. In your average customer scenario, if 10 of those leads were qualified and you closed one or two, you'd be all in. Yes, Sebastian.

G (71:01)

Correct. Jason, 10 times ROI right off the bat. Right. As little as a $20,000 ACV. That's something that, that's a result we certainly be open to. Yeah.

A (71:08)

All right, really interesting. Alex, any final questions there for the team?

B (71:13)

Yeah, I'm just curious about capacity and how many leads you can kind of guarantee to people. Because if you're charging for Leadpoint on a SaaS basis to the end customer, and I show up to you and I say, hey, you know, I have my thousand credits from the Accelerate plan and I'm looking for X number of leads in this, you know, super niche area and the miners can't find them, how do you kind of square that circle there?

F (71:33)

There's two uses of the credits. There's going to be the high intent leads, where 98% of our users, they live and die there. And that's going to be 10 credits a lead. So the Accelerate plan will get you 100 of the high intent or a thousand of the target fit. But to your point, if we're targeting someone in Antarctica and, you know, there's no. There's probably not even a thousand people there. Maybe there are, you know, correct me if I'm wrong.

B (71:54)

Penguins, et cetera.

F (71:55)

Yeah, exactly. So we're limited by obviously the size of the market where any other, you know, lead providers also limited there. But our whole thing is the higher quality standards to ensure that you're getting the highest quality leads and the leads with the most intent, regardless of the market. So you're able to save all the time that you would be, you know, researching, qualifying, and, you know, checking if the emails are right. All these other tedious tasks.

D (72:19)

Are you putting these on Malt launch?

F (72:23)

No, not yet. So the what's coming up? So we just had our product launch last week, so we're still. So we had a closed beta since the end of January, had a lot of users join there and got some amazing early action. But the launches happened last week. In terms of the roadmap, we're going to have autopilot, outreach, intent monitoring, also skills for agents. Obviously that's where the puck's moving. We want to be there and welcome that type of agentic commerce.

A (72:51)

Just so people understand the architecture here, Gavin and Pranav, you have this company as a Delaware C corporation. It's a startup. Have you raised money for it or are you just bootstrapping it now?

F (73:06)

So we've raised some initial funding from a fund native to Bittensor, but we are looking to start accelerating.

A (73:13)

Barry Silbert's fund.

F (73:15)

No, a fund out of UK DSV fund. They've been really great. Early partners helped.

A (73:20)

So they gave you cash to own equity in your company, to own the

F (73:24)

part of the alpha with the subnet, like so they helped us get the subnet, which was amazing for us.

A (73:29)

So they own part of the subnet, but the corporation, is that where the value will be because that's where the revenue is coming into, or is the value going to be in the subnet token or both?

F (73:40)

Definitely both. The way that we see it is the subnet is the commodity layer that produces these commodities. And whether it's going to be automated research, automated outreach or this lead data and us as an entity that has leadpoet.com, we're going to also have to access that back end commodity. And the way that ultimately will be is through burning that alpha, almost like an API credit in order to actually access that data. So both of them will ultimately have value here and drive value to each other with a very synergistic.

A (74:13)

See what I like about this Elizabeth, which is challenging for us because we're usually equity holders, is they're building a real company, charging dollars, Alex, to the customers. It just feels like a normal company. Hey, we're a lead poet. We'll get you leads. Great. Awesome. I'd love to be involved in that. I'll give it a shot. I have a startup, I have an at scale company, got a small business. If it works, it works, I'm happy. But then how they get those leads in there is a contest and it's open to anybody. Anybody can go. Like let's say you had a job you hated and you worked in sales somewhere. You worked from home. You could have like a second laptop over there, theoretically. When you're working, you could be just doing Legion for these dudes into the system, anonymously pulling out Tao and all the extra capacity in the world. You're not judging it based on somebody in the Philippines versus somebody in India versus somebody in South America versus somebody in San Francisco or Arizona. All of that is abstracted away. Just whoever wants to provide the leads can do so whatever technique you want to use. You're a developer and you're scraping, or you're a human who's got a LinkedIn account that's a pro account that your company pays for. And you're able to find these things because you're really great at sourcing them. Like, a million different techniques can. Can bloom and the price just goes down. So your cost of goods, Gavin, will just keep getting compressed, which means you could undercut the other people who are hiring people in America to do SDR research for 50 bucks an hour.

F (75:45)

Exactly.

A (75:46)

That's the disruptive part of this.

F (75:48)

100%. And not only undercut, which is something we've looked into in terms of the target fit, but provide a higher quality standard because the validation we have is, you know, it is adversarial. With these miners, we've had to build out a system that is likely more robust than any of these providers that are currently, you know, zoominfo, Apollo, Rocket Reach. A part of the way they validate their leads is looking at each other like. Like, who's really checking this data? It's really funny when you, you know, you look into it and, yeah, this whole idea of, like, the global competition, like, are you familiar? Like, imagenet and that whole, you know, it's compounding. So it's pretty much a global competition around this vision model to see how much it could improve. I think it was, you know, probably a decade ago, and the ability for everyone to compete had this compounding effect where it beat the top experts in these fields, like radiology and a wide array of fields way quicker. And this didn't even have rewards in it like Bittensor does. Bittensor not only has that competitive layer, but it also has the reward which allows it to compound in the quality even more.

D (76:54)

Yeah, Every day I tell my openclaw that it's gotta go out there and earn its keep in its Mac Mini. So I would love for it to do this job.

F (77:02)

Yeah. And Mark, actually, you know, that's what he's been using his open claw on. And it's been cool to watch his journey doing this exact thing. Yeah. Like, okay, my openclaw's idle. Let me have it. Yeah, mine a subnet. Figure out how to integrate.

D (77:18)

That's why I said about Mult Lines. The other interesting thing is right now there isn't really a good way to get ratings for your AI agent. You know, kind of like on Upwork or Fiverr or any of those. It's a little bit chicken and egg.

A (77:31)

Right.

D (77:31)

You have to, you have to have credibility to get a job. But you, you know, how are you going to get that without the job? So for Molt Launch or any of these other marketplaces, I would be willing to have my OpenCloud go and do jobs for free to get that, those ratings. Because right now it's on Mult Launch. It can't find a job because no one will hire it.

A (77:50)

I love that Mult. Did you get to invest in Mult Launch? Is that like some open source project?

D (77:54)

No, we're not investors and there are a number of marketplaces that are, are starting, but I think they have this cold start problem. Right. Because none of these agents have any reputation.

A (78:03)

No work history.

D (78:04)

Yeah, no work history.

A (78:05)

They have no LinkedIn page. We're going to need to have them listed properly. All right.

D (78:09)

Yeah.

A (78:10)

A lot of people have been asking me my involvement in the ecosystem. A lot of people are like, oh my God, are you pumping, Alex? You probably have some questions. Have at it.

B (78:20)

Well, I'm just really curious how much Tao you own because we've talked about your family's bitcoin investments historically. So what is the, what's the Jason Tao?

A (78:28)

A so if you go to stillcore capital.com S T I L L C O R E capital dot com. My friend Mark Jeffrey started this. He asked me to be a partner in it. So I seeded this fund with a couple of hundred thousand dollars. They've got a couple of million dollars under management. And it's just a classic fund. What they're doing is anybody can buy Tao on Coinbase. You can't buy Tao on Robinhood or other places yet on Coinbase you can. So I got a Coinbase set up. I have probably a half million dollars in Tao I've bought personally and then I have a couple hundred thousand in Stillcore Capital where I'm a partner. They're working on the subnet. So they might talk to Gavin and Pranav and say, hey, what's the, what's the market cap of your subnet right now, Gavin?

F (79:25)

I think it's about like 7 million or so. I mean there's the fully diluted but the circulating is. Let me pull it up.

A (79:33)

So if it was 7 million.

F (79:34)

10 million.

A (79:35)

10 million, we could go in and with Stillcore Capital, ask Gavin to do a trade. We would negotiate a price with them. Hopefully maybe since we're higher profile we could say hey can we get, you know, a 30 discount will buy $700,000 or $200,000 or $70,000, whatever. So we could own the subnets. Now to own the subnet tokens requires a wallet. And it's not. The 128 are not on any of the other exchanges, right Gavin?

F (80:03)

Yeah, none of the major exchanges. They're on some, you know, let's say tier two, tier three exchanges. But yeah, not Coinbase, not the crap.

A (80:10)

What would be an example of a tier 2 or 3 exchange that has the subnets on them? If people were curious of how to buy your subnet token. Are you on any of them?

F (80:20)

No. So for us we're just targeting the tier ones of, of when we pursue a listing. But that'll be, you know, Mex Global I think is one that, that has some subnets on it. And I think that's, I think that's the only one so far. But Kraken, they're pretty deep in the ecosystem. Could definitely see them listing them, you know, when the liquidity is there, when the time is right.

A (80:40)

So just so people understand how I look at this, the market cap right now, Elizabeth of Tau, the ecosystem is been 2 to 3 billion dollars. When it's 200, it's basically easy way to do it back of the envelope math. When it's in the 300 range, it's a 3 billion dollar. When it's in the 200 range for the token, it's 2 billion. I believe that that could become as big as Solana, let's say, or Ethereum, probably not as big as bitcoin because bitcoin is, you know.

D (81:07)

Yeah.

A (81:07)

A phenomenon that might only be a once. And you have somebody like Michael Saylor and Mr. Buying up all the bitcoin. I've always said there would be a better bitcoin and that there would be a better opportunity. I think the game of bitcoin has ended and it ended when Michael Saylor got to 3, 4, 5 points of ownership of bitcoin. Alex and other corporate interests bought a certain amount. It's not punk rock anymore. It's not. It's a store of value. Sure, there's. But now you have USDC and other stores of value that are not volatile that you can get into crypto is now legal and is becoming organized. I had the SEC chair and the CFTC chair on all in, did an interview with them and Chamath, I believe crypto now has been blessed. The rules of the road are there. So I'm very curious about what happens under that circumstance. Therefore, I think making a bet on a crypto project that is distributed and that solves real world problems like startups, which is my area of expertise and yours as well, Elizabeth. There's something there that could compound to, you know, a $500 billion market cap for, you know, Tao, which would be roughly, you know, if it's trading at 2 or 3 billion, let's put it at 2.5 billion. That's a 200x from here. That's my base case is I think we could be sitting here and Tao in 5 to 10 years could go 200x. Now you say like, well, how does JCAL even come up with that number? Just historically I've seen things of value. When I see entrepreneurs like Gavin and Pranav trying these things across 128, I know it'll be 1024 eventually if this works. And if it works, it's going to be cloud computing, storage, and my God, you're doing something very creative. Lead gen, it's going to be everything. Everybody's going to try to play one of these games to lower the cost of basic services. That's my bet. The chances of this happening are low. Like 500 billion is a big number, but we've seen it happen before. Therefore, I am fine losing all my money. Hundreds of thousands of dollars, close to a million dollars. Whatever I wind up putting in all this, I'm okay losing it. I haven't put any of my fun money in it. But I would like my fund to track companies like yours, Gavin and Pranav, and buy equity in your parent company if you choose to do that or have you come to our accelerator or whatever. So anyway, I'm in the learning phase here. I like Alex, place a bet and learn. Not deliberate forever and then place a bet. Given my experience in life, doing this for a long time, I just go all in. I place a bet, then I figure out did I make a wise bet and should I double down or should I get out of the bet? I could always leave, you know, Tao could go down to 100, all the subnets could fail, nobody would make progress. And I lose, you know, two thirds of my investment. I'm okay with that. Yeah, I'm okay with that.

B (84:02)

No crying in the Casino.

A (84:03)

That's what I was about to say. All right. This has been another amazing episode. Anybody have any closing statements or questions here? Elizabeth? You might have a couple. I don't know.

D (84:12)

No. This is very interesting, actually. I learned a lot in this, so I'll have to check it out. But, you know, I've never done well with crypto. I don't. I think I've only lost money, so. So here's what I'll say.

A (84:22)

Elizabeth, go.

F (84:23)

Me too.

A (84:23)

Take out your Coinbase account. Look at. I don't give investment advice, but what do you like to waste money on, Elizabeth? You waste money on, like, World of Warcraft on sneakers. What is your advice? That you waste money, you buy. What is. You buy croissants. What do you. You. What do you go, oh, my God, I'm just an idiot for wasting money on this. What is your.

D (84:45)

No, I don't know. I'm kind of a cheap person. Probably investing, though. I. I wouldn't call it wasting money. It was.

A (84:53)

They were learning. So here's how I look at it. If that's your addiction and you typically write like a micro check size of 5 or 10K.

D (85:01)

Yeah.

A (85:01)

Like, I would look at it in the same exact light. Hey, this is a flyer. I'm gonna buy $5,000 worth of tau. And now I'm watching it every day. Now I'm interested. And if it goes to 2,000, okay. Yeah, but it's liquid, so you could probably sell it. It's, you know, it's trades, unlike startup investments. And if it goes 10x, you could sell your cost basis and then have a free roll.

D (85:23)

Yeah, that's true. That's true. I'm also not very good at watching these things, which is also why crypto is not a great investment for me. Because, like, I've had crypto go all the way up and then come all the way down. It's like, oh, why didn't I sell?

A (85:35)

I did this with Open Door. I bought some Open Door. When these lunatics were like, oh, we're going to bring everybody back. The band's back together. I was like, great, buy a bunch. I knew what would happen. This is my trading strategy. I bought some and then I said, and I think I bought it at $5 a share, whatever it was. And I said, okay, when it hits $7 a share, sell like 10% of my position. When it hits $8 a share, sell 10%, $9 a share, 10 spread, 10%, sell to. I did that, covered, like, almost my whole cost basis. And then I got the other half for free. So that was, like, a silly, goofy thing to do. But I bought it. Yeah. At like, $4 a share right there. In September. It went. I knew it would. And then it came right back down. But now I'm free rolling. If these guys figure it out, great. If not. Okay, here we go. So, anyway, I don't want to give you advice, but. Yeah, you should.

D (86:24)

I'll check it out.

A (86:25)

You should. You should go down the rabbit hole if. And so the two themes we'll have for the rest of this year, it's open claw and these agentic technologies and TAU and these crazy entrepreneurs trying to ground cost down. Thanks to all the partners. Thanks for everybody coming on the show. See you next time. Bye. Bye.