Software Stocks Have Massively Crashed. Here's What Founders Need to Know.

Long overdue. I'm so happy that we were able to commiserate over the fantastic business that Davos is. And now we have our planning session. We just talked to Tyler Cowen about how you're supposed to keep everything a secret. But here we need to do some open brainstorming. We need to do some open brainstorming about how we, how we drive to half a billion in revenue, because apparently that's what Davos is doing.

Well, I'm sure you want to stay on point. If you want my advice, which you should ignore, starting next year, take some off the table.

Okay.

Listen, TBBN is on fire, right? It is great. You're— you've reached the point where you almost have too many sponsors. You have to turn them away. Sure. But listen, if both you guys could take out, say, $4 million next year, just put it away. Might not be the worst idea for a media business, right?

Yeah, yeah, yeah, yeah.

Small bit of advice.

When did you get into media? Like, what was the inciting element?

We were the opposite. I mean, there's good and bad for being in the early days or old school, right? You don't know what you're doing, just like the early days of SaaS. But we just did meetups for our community in 2013, 2014. Like 1,000 people would show up to a meetup. And now in SF, it's like— well, at least last year, it wasn't a big deal for AI. But imagine in 2013, 1,000 folks coming together to talk about enterprise software. It was nutty, right? And so it built a community and then accidentally did events, which you guys should not do. And then did media. But I did it backwards, right? You guys were very intentional, and it's pretty cool. I was utterly accidental and dragged into it. But it's convergent evolution.

Yeah. What are the best practices for events? I mean, people talk to us about it. We've said, no, we're staying focused. There's plenty of events. We get invited to a bunch of them. I have to say no to too many already. There's a good event in most spaces. And I'm wondering what you think makes for an actual good event.

Well, you mean from a business model perspective or from a quality of attendance perspective?

They're perfectly aligned, right?

Yeah, they're not aligned. They're poor. They're arguably poorly aligned.

Okay, let's start with the consumer.

No, and we've experienced this because we haven't done any events and, and we've done one that was not— yeah, but it was basically like a cocktail party. Exactly. Uh, and, and like, it— yeah, you, you I feel like too many people, like, you got to really know what you're in. You can be in the events business.

Yep.

Or you can do events. And if you're just doing events and you're doing events that you really want to be at and participate in and you're making events for yourself, it's very likely that it's not a great business, right? Because, like, it's very possible you don't want to charge for tickets, right? Or you charge for anything or all these things.

A lot of conflicts. Right. Well, look, I'll give you— I thought we might be talking about agents or the SaaS market crash. We can do what we want.

We want to get to that there for sure.

But if you want me to summarize learnings for events, which you guys have, there are a handful of public companies in the space which are boring but have very high operating margins. Yeah, they tend to get bought out by PE firms and go public and go private. Here's the basic thing. There is a nut for these large events like Con Lyons and Money20 and even Saastr. And it's going to cost you $10 to $15 million, maybe even $20 to turn on the lights. Whoa. Just to turn on the lights. So if you bring in $10 million or $15 million, like a lot of corporate events do, Dreamforce, they lose money, right? If you can build that—

Well, they're also taking on extreme risk because let's say you have some black swan or whatever, you have a bunch of—

We lost $10 million in COVID. I lost $10 million as the majority shareholder. I lost $10 million. It wasn't that fun.

Yeah, that's brutal.

But if you get over the nut, if you think about the nut, and they're high, this is why it's a terrible business. But if you get over it, it's almost pure profit. That extra person you charge a ripoff ticket to for $2,500, that costs you $80, right? But getting there is— it wrecks the ships. And if TVPN has money, it's the last thing you guys want to do is to struggle to make $200 grand or $500 grand off an event because it's totally distracting, right? But if you get massive, And you keep your heads down. These things, the best of anything makes a lot of money, doesn't it?

Of course. That's always the nature of these things. That's why we like Focus. That's why we like not doing events. I am interested in your reaction to my take that the DreamForces, the Meta Connects, the Google I/Os, this has put pressure on the cross-platform tech conference. Mark Zuckerberg doesn't need to go to CES to announce the Meta Ray-Ban displays. He hosts his own conference. Sony and Xbox, like, they can— they don't need to go to CES even if they have a presence. It takes away from, like, the aura and importance of those events. Is that something that you— that resonates with you, or is that sort of overstated?

Well, I don't know that events have worked for a product launch since, like, Steve Jobs. It's all— it's now performative.

It's very memetic with Steve.

Yeah, yeah. Satya getting up on stage and talking for 4 hours about what's at Microsoft, it's already been leaked on TBN and everywhere else yesterday. There's no there's no energy. But so I think that, and to the extent events are still doing that in the age of AI, they feel very dated. They feel super dated. But the meta question is, in the age of AI, when— I mean, literally at Saastr, we have 20 agents that replaced 8 people, OK? We're running it all the time. We're at the bleeding edge. When and why do we want to meet people in personal and business lives? And meeting— I mean, you guys know— I mean, we're meeting over Zoom. It's not a tenth as valuable as if I was there in person with you. And we're friends forever, right? We're fake friends right now. And so what does that mean? What does that mean for building relationships, business relationships, pretend relationships?

Well, yeah.

So getting people together is super important.

Yeah. To maybe say it differently, AI could be incredibly bullish for events businesses because if you're just constantly online being just completely flooded with inbound and bots and bot phone calls and all this stuff, and you don't know really who's real and who's not, then in order to figure out like actually how you want to, who you want to partner with, who you want to make, how you want to make decisions, maybe you do need to get together in person.

You do. The thing that just like a lot of things in AI, the surface level misses some of the complexity. The flip side is as humans, as culture and tech, you know, since 2020, we just don't want to get out of the house. We don't want to go anywhere. We don't want to travel. We don't want to meet, like sales folks don't want to meet customers in person anymore. Right. And so even if we know we need to do these things to go to these— like we just, we all, we're either 996 or we want to work from home in our pajamas. There's nothing in between these days. And so, it actually creates a challenge for all the reasons we need to get together. 99% of us would rather be working from home. So, we'll see. I don't know.

Yeah. Yeah. Yeah. Sorry.

What— I'm sure they're all wildly different, but what are your conversations like across the portfolio today? You have a bunch of new AI-native companies that, and those Conversations look one way and then you have, you know, a massive portfolio that all is navigating AI in different ways, capitalizing on it.

Maybe AI companies are becoming SaaS. The SaaS companies are doing AI. Everything's bleeding together.

Sort of.

Sort of.

Sort of. My— so here's my simple rule, guys. If growth isn't accelerating, you're not an AI company. Oh, this is the flaw with the publics. All the performative stuff, all the public companies you guys talk to and follow has growth. Great that you built an agent. Great. But is growth reaccelerated?

And that's the case for Meta. I mean, that's the bull case for Meta. Like they did accelerate growth. They're spending a ton on CapEx, but there's clearly the AI that they're baking into the ad matching platform is helping accelerate. And so it's working.

And they are an advertiser that are generating the ads.

We talked to Eric Suford about this yesterday.

Well, sort of, but like, let's look. So that's for sure. That's why. But let's, let's look at why Microsoft crashed. And I never know why publics react the way they do. But on the one hand, the AI side of the business is still blowing up, right? Granted, a lot of that money is coming from OpenAI, but they did miss on the software side.

Yeah.

So it's not, you know, there's nervousness and you talk about advertising being up, the worst performing software stock of the last 12 months is The Trade Desk, just, just destroyed by all of this. It's complicated. So You want the portfolio? My, my, my advice is last year was deeply tough love. Now it's just tough.

Mm-hmm.

It's like you've had a chance to reaccelerate growth. Like, yeah, everything got better since Cloud 4.5. It got really good at 3.7. That's why Replit and Lovable blew up. That was a year ago, guys. Yeah, you had a year, whether you're Agentforce or my startups, you had a year to reaccelerate growth. And honestly, Salesforce is doing better than some startups.

Mm-hmm.

At least we're actually probably the only organization of our size using AgentForce for real. Like, we use it every day. And it works. I can't tell you how many startups their agentic product is like, it's a copilot.

Interesting. Yeah.

So did revenue grow? You had a year, you had 2 years. You're supposed to be agile.

Digging the tough love to just tough transition, I feel like there's a lot of SaaS founders who got to growth stage, they got to scale, that maybe they started a decade ago and AI did reinvigorate them. They got sort of back in the arena. Maybe they never stepped down or stepped back, but they just, it was a new reason to go into the office, get fired up, be tinkering with the tools, pushing the team harder. Is there any, is that just something that's innate to the tinkerer, a specific type of founder leans into that? Like what are you seeing across the portfolio?

Well, okay. Maybe two things. First of all, that's a great narrative. I don't think in the real world it's that common. I talk to behind the scenes, off the record. I talk to public company CEOs and B2B, my own portfolio, others a lot. Since our agents blew up, everybody thinks we're some sort of, at least for GTM, we're the agent gurus.

Okay.

It's pretty interesting. And everybody says 80% of their team wants to work like it's 2021. It's not simple. Everyone has to create a skunkworks team or something, and everybody's complaining because Here's one reason the AI-native companies are doing so well is they don't have to deal with the 20,000 customers pre-AI who still have feature gaps, who are still using clunky software, who are still have other competitors. And all of a sudden you have 10,000 new AI competitors that don't have to deal with that old stuff. And I would love to say I know so many folks that have that, you know, beginning of 2025 were growing 40% and now they're growing 80% or 110%.

Yeah.

I can only think of a handful.

Yeah.

How are you— how do you advise around competition in specific markets? Earlier this week, I forget who we were talking to, but I had been talking to a founder, and he had some idea. And I was just thinking to myself, like, sounds like a great idea, but I can guarantee that 4 different YC companies are going to end up on this problem, and you're going to be going head to head, and it's just going to be like— 996 versus 996 and like it feels like, you know, somewhat of a coin flip who will come out on top even though you're super talented, super experienced. What's your view on, you know, the current state of, you know, if a category is exciting, it'll have 10 companies kind of running at it aggressively.

It's worse. Like that was— I think that was the problem 6 months ago. The problem today, it— and literally a seed investor who's very successful but relatively new said to me yesterday, I'm giving up. Because everyone can vibe code something. I can't even tell the difference. I can't tell it. Now, I'm very— I vibe coded 20 apps that have been used over a million times. I'm in the top 0.1% of Replit. I know a little bit about this.

Yeah.

And we'll run out of time. You're not going to vibe code Salesforce.

We're going to hit the gong for your Replit skill set.

Yeah, hit that gong.

Numbers. Huge numbers.

Huge numbers.

You're not going to vibe code Salesforce for real. But you know what you can vibe code? Something for Demo Day that looks really good.

Yeah.

Real. And the stuff that like even 18 months ago you'd be like, I want to fund that. Oh my God, this is an agent for dental follow-ups. That's all. Like, my God, your jaw would drop 18 months ago. And, you know, we're talking about ClawBot and OpenClaw and all this stuff today. You were just talking about it, right?

Yeah.

I mean, I built my own version on Replit a week ago. I thought it was pretty cool and now it's obsolete today. For real. It's called Ren. Yeah, I built it a week ago and now it's worthless. Wow.

So I mean, Demo Day is coming up. What metrics— I mean, obviously you can have a polished product. So does this shift to being able to control narrative? Are we in the age of storytelling? Do we need to be focused just on ARR or cash flow at the earlier stage? What changes about— because there will be companies that are created today and wind up being successful. I'm sure you're not bearish on just startups generally right now. No, I'm not bearish. But the landscape is different, right? So what's different?

Look, here's the challenge. The challenge isn't even all the clones. I think we've accepted there's a lot of clones. I think we've accepted there's 1,000 competitors now. The challenge that— I mean, you guys know from the show, but not all founders have internalized. It's just investors are expecting insane levels of growth.

Mm-hmm.

Insane levels of growth, right? They want— like the idea that you could go from 1 to 100 in a year. Is now seen as what you want to invest in. That used to be almost unprecedented. It did happen in the old, old days, right? Now there's companies that like I invested in early, like Higgsfield for video. People never even heard of Higgsfield and it's $100-something million, right?

It's over $200 now.

Yeah, but that's not like Harvey, which we're talking about every week, right? Or Lovable or Replit. And so you— and then the problem, guys, is you look at Figma, and it's terrible, but Figma is almost soul-crushing for investing because you can't get much better than Figma. It's down from its IPO. It's trading less than 10 times revenue, and that's for owning and creating a category. Who's invested in something much better than Figma? I mean, not me.

Sure, sure.

So on the one hand, it's great, the best times of others. On the other hand, it's like, Figma isn't good enough? Like, let's call it a day, guys, and do media companies because it's just too hard.

Yeah, I mean, I've had at least a couple a couple of times over the last year where a portfolio company is like a company that is maybe a pre-seed company trying to go out and raise a seed or maybe an A. And they're like, hey, can you look at my deck? And I'm like, the deck is beautiful, but you're projecting to grow 3x this year, and that's going to be a nightmare for you because it just doesn't look like the business is best in class anymore, even though you're like, wait, I'm going to 3x. Revenue this year?

Yeah, it's— and unfortunately, there's no great answers to this, right? There's no great answers to the question that some great businesses will compound to epic rates, but they're unfundable now, and they would have been fundable 2 years ago. There's no— there's literally, I don't know what the answer is today, right?

What about how— what do you think is going on in—

so staying there, doesn't that just mean that maybe there's the same amount of entrepreneurs, same amount of building, but just more founders opting out of the traditional venture capital track? Because if they don't actually have a solid use for capital, they should get profitable earlier, monetize earlier. I mean, integrating payments is easier with one-line prompt. So just do it and actually pull your roadmap forward and get and stay lean and be higher leverage. Like, I don't know, is that not an unreasonable conclusion that you'll just see more bootstrap successes?

You know, there's— of course, that's the dream, right? Don't raise too much money for TVPN. Keep it lean. Yeah. And I think for a little while when everything was easy, but when things were growing well, but the markets were down late 2022, '23, and '24, that was the dream. Guys, listen, if the markets have fallen out of love with us, we'll just do the Mailchimp. We'll just get to $1 billion in revenue on our own, pay us all out a couple hundred million in dividends and see what the Lord brings, right?

Yeah.

The problem today is that the pace of software development is so fast that you better, you better have one way or another, you better have 4 or 5 amazing folks on your team or you're going to get crushed by the kids at Y Combinator, bootstrapped or not. It does. There's no time. And I remember back in the day, Michael Cannon-Brookes, who was the co-founder of Atlassian, came, you know, to SaaStr Annual way back in like 2019 or 2018. And he said, well, I was lucky I had 5 extra years. Like, if I had any competition in the first 5 years of Atlassian, we wouldn't have made it. There were 3 companies.

That's crazy.

So, that's hopeless today.

How are you going to compete when people pushing out agentic, crazy agentic products weekly? How are you going to compete with them?

So, I want to get your take on this. So, we were on a podcast that got released earlier this week. John was talking about something that has informed our strategy with TBPN from the beginning, which is that Media is now a barbell. You're gonna do great if you're a platform, a Spotify, a YouTube, Netflix, et cetera. And you're gonna do great if you're a personality, individual, low, low opex. You're just like Joe Rogan. You're creating content. But if you're in the middle ground where you have like, you know, a 200-person team and you're trying to roll up talent, they're constantly leaving. You're competing with personalities and the platforms. You're gonna be in a rough spot. A buddy of ours, John Palmer, was saying yesterday, he thinks that software could go in that direction where if you're one or two, lean, lean, lean operation, you're competing with big companies, you could do well. If you're a huge platform with distribution and scale, like a Salesforce or a Google, you're going to do well. But there's kind of this messy middle that could get kind of churned through.

Well, it's just— well, first of all, I think all of this vibe coding, I think people miss the point because they're not really doing it, right? And you had the great guy, The Economist, on before me, who was—

Tyler Cowen.

Yeah, he was great. And you asked him, have you actually used any of the Agenta products? He was like, no, he was honest, right? Kudos to him, right? Most folks talking about it haven't. What Vibe Coding is already unleashing is if you want to, to your point, if you want to build a super niche app for real, I want to build software for web broadcasters with scale, okay? There's 11 customers, there's TDPN and 10, that's all I want to do for my dream. If you are willing to, to do it for real, not for an hour, not one shot it, do it for real. You can now build that software without an engineer. So that is amazing. And so we're seeing whatever, 1,000 flowers bloom. I was the first investor in a company called RevenueCat that does— powers mobile subscriptions for 50%— for 50% of mobile apps. And Andreessen published basically the same data this week. At the end of last year, just all of a sudden, the number of mobile apps like quintupled. Wow. At the end of last year because of vibe coding and it's just starting. And so the very bottom of the market, if the three of us want to get together and build an app for real, we can do it now. It's very exciting. And Salesforce isn't going away, but that middle is going to be harder. It is. It's a good point. And there's just— what I really want to build it today, you can. And it's super exciting.

What's going on in PE land. I would hate to be a private equity firm with a bunch of kind of legacy SaaS, and you're having to explain to your LPs that, no, everything's fine. We took this company private. We'll take them back. It's fine. The Thoma Bravos of the world, I'm sure they're very savvy. How do you think they're kind of thinking right now? How intense is the fear? Where is it sort of unwarranted?

Well, there's one thing I know for sure, like for 1,000% sure, and this is slightly inconsistent with the data Carta and others put out and others on the internets. These B2B companies, pre-AI ones at $50 million, $100 million, $200 million, $800 million, no one wants to buy them. It doesn't matter that you— hooray, you got profitable. Thank you. That means you're not going bankrupt. You have not solved your existential problem in the AI. I mean, literally, I'm an advisor, a friend to a company about $140 million in revenue. Pretty good growth, but not AI growth, but pretty good, right? And we were doing a review the other day on M&A, and I was shocked with the folks on the block that would sell— folks much bigger than them, much bigger than $140 million—

to sell to them. So to sell—

okay. Any exit, any— no P for— they've already gone to Thoma Bravo and Vista and Insight. They ain't going to $140 million ARR for a weird crammed-down combo exit unless everybody else in the PE chain said no, right? It's rough out there. And I, you know, until even late 2023, you got to $20 million in revenue, you were still growing and you were efficient, someone was going to buy you. And the question was, was it 5x, 6x, or 10x? And that playbook died for PE and it died for exits. And I do think that Tomo Bravo will be they will figure out how to accelerate AI in their portfolio and go deep and it will work. I really believe if you can own the agents on your platform, that's how you reaccelerate. You have to own the agents on your platform or the agents will take away all the value. But man, the PE has just said goodbye to B2B and it's terrible. I don't mean to be so draconian in 2026, but I talked to so many founders that are in la-la land and I think that was okay last year, but at least at some point you just be honest that PE is not coming to the rescue, unfortunately, for 95. I mean, everyone's on the block, guys.

Does PE ever come to the rescue? It feels like it's more of a grim reaper scenario typically. Yeah.

On the other hand, it's just so exciting to be building things today.

Sure.

Right. And there's either a malaise in these companies because they're building nothing, or you turn around and you can't believe what they built last week.

Yeah.

I mean, Ramp's your big sponsor, right? Even though Ramp is old, look at the pace at which they put out software now. It's crazy, right? And anything below that is going to die.

It's just atrophy. Do you ever go around to your portfolio and say, like, hey guys, why am I using AgentForce? Like, why is this— like, nothing against Salesforce, but I was surprised to hear you're trying everything, using everything, building your own agents, and yet you're sitting here saying, like, yeah, we're running everything on AgentForce and it's great.

Well, we run— we will actually— and this is a risk for everyone. For the moment, you can be promiscuous with agents. So we are running 4 different sales agents. We're running AgentForce. We're running a hot YC company called Artisan. We're running a company called Qualified that Salesforce just bought for almost $1 billion to get more reach. And then we're running— and then we're running a bit of Clay, which just raised at $5 billion.

Sure.

So we're actually running them all. We're— it's work. We're running them for different use cases. And that's actually good for startups because in the short term they get more customers. But overall, it's hard at the moment to dominate. So it's not just AgentForce. And I think that's the risk. That is, that is the risk that I think AgentForce is going to work. I can tell you why as someone who's actually using it, not Sony baloney, but we're using 4 vendors instead of 1.

Yeah. Can you give more—

instead of 1, there's risk as well as opportunity, right?

Give more concrete examples of how you're using AgentForce, what it's doing for you? Is it sending emails? Is it drafting emails? What?

Yeah, that's all it's doing is sending and drafting emails. We gave it the use case, which is a good one and which is the one Mark talked about in the early days. We gave it reactivations.

Okay.

So, and this will resonate a little bit with you guys, folks that used to be sponsors or used to come to our events. This is very simple and atomic that a human being was too lazy or unwilling to follow up with. And, but they were good ones. So we scored them and sent them out and we got 70% open rate.

Wow. That's great.

It's pretty powerful. This is something when we had 8 people on our sales team. Now we have 1 in AIs. We went from 8 to 1 in AIs where it just wasn't worth the human's time. It just wasn't worth John and Jordy's time because they're hunting the big deals at Vanta and Ramp. They don't want to hunt the little deals or, or there was turnover at Brex and they don't know how to meet Pedro. Pedro's gone and they don't know. So the humans give up, but the agent doesn't quit.

Sure.

Yeah. Agent just looks up who replaced the person at Ramp or Brex. They follow up with them. They have no shame. And we had an agent that closed a $100K deal on Saturday night. I mean, let's be honest. Oh, you want to close a deal? Yeah, for real. How many humans want to do that? What, wait for the e-sign document to come back? But how many are going to do the work on Saturday night? They're streaming, right?

Yeah. What about— do you think we'll see any— are you excited about any turnarounds, public market companies that have been that have been completely beaten up. I mean, we just had Jim on from Yahoo, and that's a business that people aren't talking about. Certainly the tech media isn't spending a ton of time talking about Yahoo, but he's sitting there being like, growth is great. We've got hundreds of millions of users. We have all these opportunities to put AI in across the ecosystem. And you have to imagine there's a bunch of companies that have been beaten up in the public markets that maybe founders, people just wanna start something new. They don't wanna try to deal with, legacy platform, but I'm sure you've been pitched random ideas like that.

Well, look, there's a bull and a bear case here, right? And the bull case is AI is so early, and especially in real enterprises, not all the tech folks buying these products on a circular basis. Everything's circular, right? Including GPU sales and everything's circular. If you go out and talk to the real world and bring folks from Archer Daniels Midland and whomever on the show, they're going to tell you it's early, they're experimenting. And I think that is going to benefit Salesforce and ServiceNow. It's not too late for them at all. Right. There's a lot of stress at those companies today, but it's not too late. The flip side, though, I got to tell you, now that we're in 2026, you've had time. Where's the turnaround? Like, hooray that I can do all these things at Yahoo! But it's already like it's already just utterly changed video and audio. And everything. Why? Why? Show me the money. Show me the lift in revenue. And I have lost patience with founders at $1 million, $10 million public companies that have not seen the lift because great. But ElevenLabs just crossed $350 million. Yeah, well, show me the money. Like enough talk and enough arm waving. And, you know, you— AI doesn't count if the revenue doesn't grow.

It's—

the problem is, here's the tough part for everyone, including Yahoo and most startups. It's hard enough just to keep up. It's hard enough just to ship an agent. It's hard enough just to achieve parity, figure out the guardrails, figure out how to make this product work. And then the other guys have pulled ahead. Like, it's just— you're almost bailing out a leaky bucket. And so I know we want to be optimistic that there's so much agentic opportunity, but I'm at the point you got to show me the money.

Mm-hmm.

Yeah.

I don't need to see you go from— and look what happened. Like, Mongo, like, dramatically reaccelerated. Yeah.

Yeah.

In the age of AI. I'm not looking for you to go from Mongo from like down to like the teens to the 30s or whatever they did. I'm just looking for a little bit of gain.

Yeah.

Yeah.

Even truly got a little bit.

What about, how are you thinking about AI discoverability? Are you paying a lot of attention to this with your business?

Are you guys—

To find vendors? We need to get business?

Well, yeah, yeah. Just like how you're showing up in LLM queries, all that stuff. Do you think it's— Are you advising portfolio companies that they need to pay a lot of attention there?

GEO broadly?

Yeah.

I think people are underestimating this. I think GEO is— it's almost a disservice. You don't just want to show up randomly in what's a next-generation media company focused on tech and it says leaders include folks like TVPN and the others. That's not like you got— that's nice.

Yeah.

What you really want to show up for Like, I'll give you a very specific example. I did this as a use case yesterday. I went into Replit and I said, what's the best CRM for me to use? And it said HubSpot. That's where the money is today.

Yeah, totally.

Us asking all the agents we work with, whether they're Claude or ChatGPT, or if we're building something in Lovable or Replit, we're not, you know, we're not going to go to Google. We're just going to ask the agent, what should I use?

Or you won't even ask the agent. You'll just say, I need a website. And it will be like, well, I need to pick a database and I'm picking this database and you don't even know what database it picked.

And that's, that's an issue. But I think like software isn't dead. I mean, the public, the leading public B2B companies are doing $2 trillion.

Totally.

But when people do discovery, they're going to ask their agent. And I even built a digital Jason. It's been used 175,000 times. People just ask, what should I use?

That's great.

Why would you go to Google and pay? Even, even you might use the AI summaries if they answer the question, but we're not discovering with a sales rep that doesn't know the product. Right. And endless webinars like that stuff. It just, just ask, figure out what is the most trusted agent for what you do and ask them what to buy. And, you know, you should just buy that product.

Yep.

Where should I— I've got $500K to sponsor this year. I want to reach tech leaders. I want to do it on media. I want it to be persistent. What's the best place to do a $500K sponsorship? And if it says TVPN, I should just like, let me not waste my time. I'll do one more call. But if you guys will take my money, I got to move on.

Yeah, yeah, yeah.

The other, the other dynamic is just agents selecting different vendors themselves. That's a whole, that's another level of it, which I don't think enough people are kind of even caught up to yet, which is like if I just have agents running parallel and they're— and you, you can assume if your agents are closing deals now on a Saturday, maybe you start to give them a little budget not too long in the near future and they're starting to actually, you know, spend money on behalf of the company too and just make certain vendor decisions. And you trust them because, hey, you're doing great work.

Yeah, I don't know the big examples, but there are startups that have blown up in the last 12 months like Resend instead of SendGrid for email, WorkOS, and others where the agent just said to use them. Yeah, like I use Resend instead of SendGrid. I couldn't get SendGrid to work in Replit. I just couldn't get it to work. I could tell you, and you know the reason why? The founders are long gone and they decided to throttle the free accounts so much that they don't work. Like you just can't get a free account working. And so I asked the agent, well, what can I do? I'm banging my head against the desk. They said use I'm like, yeah, that guy seemed cool. I've never gone back. Wow.

Right? That's remarkable.

And WorkOS was around for years trying to do OAuth, like authentication for apps. A good solid CEO, good solid technology, but went through layoffs, was kind of going nowhere. And then everyone just started to use it. Every agent recommends it. Everyone just blows up. So, you know, that's not yet— that part's not yet in mainstream America. That's pretty nerdy. But some version of this is how we're going to pick agents. And so, you know, I think geo is nice, but some of it's scammy and smarmy. And I'll tell you how I know. In my opinion, when an agent is scammy or smarmy is when I have to put in a credit card before I can use it. Oh, I don't like that today. So all these geo tools are trying to get you to pay $5.99, $8.99 before— just let me try the thing for a week.

Sure, sure, sure.

Yeah, yeah, sounds free trial.

If it works, I'll pay you. Yeah, yeah, yeah, yeah. I mean, obviously you've talked a lot about how competitive these vibe-coded products are, the early tech markets. Are you optimistic or getting more interested in the more niche, far-field businesses where maybe there aren't really software solutions? Maybe you put Harvey in this bucket and some of the medical stuff, but also like hard tech. There's a whole bunch of pieces of the economy that have not really been touched by software yet, and maybe AI is the thing that makes it now we can go do something on the farm that's actually impactful, or do something in oil and gas, or do something in fishing, or some really far-out piece of the economy where there isn't an established conference.

Yeah, we had the founder on the other day that's doing super intelligence for dairy farms.

Dairy farms, yeah.

It's like, he has a product, you just walk up, if you're working on a farm, you walk up to the computer, you say, what should I do today? And it just gives you a bunch of tasks.

It's a good market. I almost did a dairy farm investment years ago that was using the— what do you call it, visual AI to optimize farm yields, right? And there's a couple of larger companies in it. I'll tell you, I can give you a couple of answers. Here's what I'm thinking of more in terms of the one, will a lot of flowers blossom here? Yes. Like we will see more and more of this niche software. It is great. To make money from it investing, here's what I'm looking at is, is the agent, however the heck you built it, is the agent so powerful? So ROI positive that you can charge 4 to 5 to 10 times more in this category than you could before. That's where the math compounds to something interesting. So it's one thing if you can just track the yield for the cattle a little bit better than the prior, like that's great. But if it's the same unit economics as the deals I looked at 4 or 5 years ago, I'm out because even getting to $100 million is tough, right? But let me give you a contrasting example. And these AI SDRs that people— well, we use a lot of them, people used to make fun of them, right? But these companies, startups like Artisan and Qualified and Clay, they're expensive, man. They're like $100 grand to start.

Wow.

Okay. And let's go look at older ver— when you, back in the day, I was one of the first investors in Salesloft, which was the last exit of the last generation. Okay. December 2021, $2.5 billion to Vista. And then—

Last chopper. Last chopper.

Wow. Yeah.

Funny story. The CEO was like, we gotta sell. And the VCs didn't want to, like, we're gonna be worth much more than $2.5 billion. But the CEO Kyle's like, no, the world's changing, man.

Wow.

But getting folks to pay $100,000 for those apps was hard, right? But today, it's just table stakes for the next generation version of these apps. So when you start to see an app that used to struggle to get $10 grand a year for get $100 grand, I'm all in on dairy or niche, you know, hydroponics or, you know, pool cleaner apps that are all really cool. It's just that the numbers didn't— I'm an investor in a company called Mango Mint, which is in a pretty crappy category. It's software for spas and salons and doctors' offices. Why is it crappy? It's mid-sized, and there's like 10 really good companies. If there's just one—

Yeah, it'd be fine. But it'd be fine. It's hypercompetitive still.

You know, after $30 million, they just blew up with everything they're doing that is agentic and automation-related because it's much more val— not because it makes the product much more valuable.

Yeah. So they charge more.

Yeah. So I want— show— I'm sorry, but show me the money. If you had a software that was $8,000 a year before and now you can charge $80,000 because honestly you got rid of 10 people in the back office, that's what's happening. You're getting rid of a lot of people.

Yeah. The other thing, when people, when people are like, oh, this agent's like super expensive to run, and like, could I— people are like, I could just get a normal person at this point. It's like, well, the point is that you have potentially an expert that can work around the clock that you can turn on and off in real time in a way that you just can't. It's like, why is it— why does a consultant end up charging, you know, an obscene hourly rate? It's because you're just tapping them in for, you know, a quick sprint here or a project or a couple months, etc. So you get a premium.

When someone says that, that's dumb. Like, because honestly, if you can— if you could— maybe at TBN you're pretty hot, it's easy to hire. I've never found hiring easy, no matter how hot I've been at any startup or anything. It's always hard to find good people. If you can magically wave your wand and get 10 great people for low wages to do the job you want, yeah, don't hire an AI, right? But they're going to quit. There's a tax to hiring them. There's a tax to onboarding them. They don't all perform. You got to train them. You got to manage them. I mean, good luck. Good luck. We are— it's harder to hire anybody.

Yeah, it's also more is, I mean, from our point of view, we had to— we went from, you know, basically 3 to 10-ish last year, so not crazy growth, right? We're a small business.

Not crazy.

But we actually like the size of the team right now. I like that we're all hanging out here in the studio all day long. I don't want to add a lot of people. I don't want to be a manager. I don't like managing people. I like working with people that are great. I don't like being a manager. And so if we can limit— if we can keep headcount low, stay small, that's amazing.

Well, that's why we shrunk to 3. I couldn't do it anymore. You criticize me, like, take your shot at me, right? But I have been doing it a little longer than you guys. I just couldn't take one more person paying them 6 figures quitting and for a worse job. I couldn't take one more person getting to work from home with high autonomy, setting their own goals, no drama, paying them hundreds of thousands of dollars and just saying, I want more. I just couldn't. Couldn't take it. And so last June, Amelia and I run SaaS. We're like, we're going to get, we're going to go so far on agents and we're going to break every agent. We're going to push it to the limit. And so now we have, you know, 2, 3 people doing the work of 15 and replaced 2 agencies with apps we built ourselves. We just couldn't take it. Wow.

Real quick, real quick. IPO, IPO market. You think, looks like OpenAI will beat Anthropic out the door. Looks like SpaceX, could do some type of deal, beat OpenAI out the door, a nice spite IPO. How much do you think—

I love the spite ones.

How much do you think kind of the ordering of these IPOs really matters? Like, is it— could Elon successfully suck some real oxygen out of the room and make things more difficult for his— for the other labs?

Well, look, I, I think I, in my limited experience, and I think some of this is a media creation for something to talk about. These, all these companies are so exciting at a retail level and at an institutional level. There is infinite demand in the private markets and there will be sufficient demand in the public markets to go public. So some of this is a media creation. Certainly, though, the IPO markets are wide open, but They're discriminatory. I mean, Wealthfront bombed. EquipmentShare crushed it. Wealthfront bombed. You got— why did— I mean, EquipmentShare is at what, $4.5 billion growing almost 50%? That's a pretty high bar. OK, so it's open-ish. So there's something to be said when the markets are good but not perfect to being the first out when demand has not been satiated, walking around. And maybe not everybody wants to do all three, right? And maybe it's tough if you're worse but still great than the other ones. There is some, you know, probably whatever xAI, SpaceX, especially if it's not SpaceX standalone, 'cause SpaceX standalone with Starlink's a great IPO, right? But if you start mashing loss-leading stuff in it, there's something to be said for maybe Anthropic will look better, so going first. But we're entering an era of just utter wealth creation from IPOs like we have never seen. Like, it's just— we're underestimating. You know, it used to be a billion, you know, a billion. A unicorn used to be great.

Yeah, that's a trillion.

And $10 billion used to be great. Now, now you go on TV and we're all talking about trillion-dollar exits like, like it's, you know, like it's snacks and popcorn and a show. And it is. But think about how many more. You know, there's 20,000 folks at NVIDIA that have made $20 million or more in the Bay Area, and it's just going to explode. And we're going to enter this weird world where the best VCs are going to make money like we've never seen before. Right. And the best engineers. And then, you know, the middle is going to have no jobs. Right.

Well, that is—

Well, Tyler Cowen disagrees with you. He thinks that the economy will—

Well, no, he did. He did say you might have to move to Houston and take a job in the energy sector. So.

That's good advice, or don't quit. But I will tell you one last thing. This week, I was at an event with a lot of C-level B2B executives. It was just a favor, but I've gone every year for a while to just— and this year, they finally made it. We can't find any jobs.

Hmm.

They're just— people just don't need these. I won't say exactly which level they were, but like, finally set in that like, I just— they just know no one needs folks with these 2021 through 2024 tech toolkits. Skill sets. They just don't need them. And it's, um, yeah, go to Houston, like, do it. But whatever you do, don't quit your job. Like, if you like it at TVPN or Yahoo or Cisco, my advice is stay.

Yeah, yeah, that makes a lot of sense.

Uh, well, so, so great to finally have you on the show. Let's do it again. Yeah, uh, very soon.

All right, you guys are the best. We'll talk to you. It's a pleasure.

Thanks.

Goodbye.