从 SaaS 到 AI-First：企业如何重塑创新

The anxiety that I see is that if you can generate an enormous amount of code and no one is reading it, you don't know the quality of the code. Nobody deeply understands the codebase, and there's more fragility, right? It's like the slop problem: vibe-coding slop in my actual production codebase. But I think the broader problem that a new company could go solve is that nobody knows how to manage that issue of human attention to engineering. I think it's open season around this really, really big problem.

Hi listeners. Welcome back to No Priors. The market is freaking out about the end of software. Today, Elad and I are hanging out and asking, “Is SaaS actually dying, or are people just projecting five-person startup behavior onto the Fortune 100?” We'll talk about what's real—incredible revenue growth, collapsing token costs, and faster turnover of incumbents—what's just hype, and how to size the opportunity. We also discuss the changing bottlenecks in building a software company and some parallels to the internet and cloud eras.

Let's get into it. It's good to hang. The market is freaking out around us, so amid all that noise, what are you thinking about?

Oh, you mean the SaaS—the SaaSpocalypse?

The SaaSpocalypse. The end of software.

Yeah, the end of software. It's kind of interesting. I feel like there are some meta-trends that people are getting right and then a lot of specific companies that people are getting wrong.

1. The SaaS Apocalypse Is Overstated

I think the basic premise is that SaaS software and per-seat software will no longer exist, everything's going to be replaced by AI, and everything's just going to get vibe-coded. So why would you pay X dollars for a Salesforce instance when you can just vibe-code it internally? All that stuff strikes me as incredibly shortsighted in the near term. Over the long run, who knows what happens in 20 years or whatever, but there are lots and lots of companies that are quite durable.

I think an interesting example of that, where I'm still a shareholder, is Samsara. Nobody's going to vibe-code a fleet-management app that will then get distributed through vibe sales or enterprise sales or something. You're going to build a vibe-coded, in-cab camera sensor that everybody will install in these fleets, and then you're going to support them using vibe agents. It's just very overstated.

I feel like it's one of those things where there's a massive market correction around something that, in the long run, has a lot of truth to it and, maybe in the short run, for certain types of companies, has a lot of truth as well. Ultimately, I think Decagon and Sierra are examples of companies where you're moving from per-seat software to basically utilization-based, customer-support-related agents. That is a real shift. That may impact some of the prior wave of per-seat software companies, but this isn't going to be every single SaaS company.

I view it as very overstated in the short term. In the long run, who knows? How about you? How do you think about it?

I think the idea of vibe enterprise sales is hilarious, because we have portfolio companies with hundreds of millions of dollars of revenue that are very committed to as much token usage as we can have and as few great people as we can have. Today, they have fewer than 50 engineers.

Mm-hmm.

They went from zero to close to 100 salespeople very quickly.

Mm-hmm.

Right? It's just a view from the growing AI natives that vibe sales is not happening, right? It's not happening.

Oh, yeah. Vibe sales is definitely not happening anytime soon.

No.

And so, again, all this just seems like a very strong market reaction and market correction. It seems like it's very overstated, especially relative to a handful of companies where you're just like, “Why?” How will you displace this company with coding? In the fleet example, you're not going to have the fleet managers writing their own apps to do all this giant surface area of stuff. It's just not going to happen in the short run.

I think a lot of it's actually driven by some assumptions that people close to my heart—engineers and builders—are making about the rest of the world, right?

Mm-hmm.

Because there's this implied belief that everyone will want to make their own software, and I think it's probably—

Software is eating the world. Is that what you're trying to say?

I am not—

Time to build, Sarah. Time to build.

I don't think that everybody wants to make their own software. I think some set of people will want to make it, and others will want other people to do it for them.

If you think about a good example of this, engineers sometimes have a personal, labor-focused picture of the world. Should you build Jira in most engineering organizations?

Yeah. It's not the best use of your time if you're focused on product. The other piece of it is the examples that people use: “Oh, my five-person startup built our own CRM, vibe-coded it,” blah, blah, blah. Yeah, of course. Before that, you just did it all on a spreadsheet, and that was fine too. You didn't have to vibe-code anything.

For very limited, niche applications where it's a technical team doing something really quickly because it's useful, custom, and bespoke, amazing. Of course that's going to happen. Does that mean that a Fortune 100 company is going to displace its CRM with some internal thing it got vibe-coded over the weekend? Probably not.

I think it's also extrapolating or projecting the behavior of very small technical startups onto the world's biggest enterprises. That's the second thing people are getting wrong: they're misunderstanding the moment. I think the internal software stuff that people are building is amazing. It isn't impressive that you can do that—it's incredibly impressive. It's just that extrapolating that behavior so aggressively and so early doesn't make that much sense right now.

I think, to your point about the five-person company versus the very large enterprise, if you ask that same engineer who's pissed about paying $10 a seat for Jira—

Mm-hmm.

—if you asked him or her, “Do you want to do the change management at Bank of America to get everybody to do this the way you think is right, and then deal with all the security considerations, manage other people's opinions about potential changes to the story-management workflow, and maintain the system?” the answer is probably not.

Mm-hmm.

I think it's focused on that. I actually think the idea that the actual production of code becomes not the bottleneck, if you know what the spec is, is incredibly interesting. But I do think it overstates how much of the overall software-vendor problem that is.

Yeah. I think people also misunderstand how much demand exists for software products. By software products, I mean everything. I mean AI, I mean different tooling.

Is software eating the world?

AI is eating the world.

Is AI eating the world?

AI is eating the world, so I think that is actually true. I think Mark's post on that was really thoughtful and forward-thinking.

I think that fundamentally there's so much demand for software, and there's so little supply of engineering in reality relative to that demand, that as you add this enormous boost of productivity to software engineers, it just gets soaked up, right? There's so much more stuff to build and to do.

I don't see teams, you know, startup teams continue to hire engineers for a reason, you know? I think the nature of the work is shifting, and I think some people are going to have real issues with that shift. Fundamentally, you're shifting from, in some cases, a few different types of mindsets around engineers.

One of the mindsets is really bespoke craftsmanship. “I'm going to do the aesthetics of the thing that I'm doing really well, and I care about the code quality and the artisanal version of what I'm doing.” Then there are people who write code because it's a utility that allows them to build products. There are some people who really like aspects of the math or algorithms. There are lots of different motivators for people to write code, and I think a subset of those people are going to be less happy in the new world.

It's kind of like indie game developers who make these handcrafted individual games for themselves and then for their friends, and then launch them on the App Store or whatever, versus the people who work at EA.

Mm-hmm.

They each have their own version of craftsmanship, but it was just a different type of thing. I think we're going to see a lot of these really great engineers who care about the bespoke craftsmanship of everything they do become unhappy working at larger companies as these coding tools get even more accelerated, because it goes against their approach to how they like working and what they enjoy about the work.

For other people who are really focused on the utility of just building product, it's going to be freeing in some ways. So I think there's also a variance in terms of the reactions to this stuff, depending on the type of utility function that you have relative to the work you're doing.

Yeah. I think related to that, one thing I've seen is that if you have an engineering identity that's based on a value-based ranking of difficulty or skill, the specific types of engineering that are considered impressive or high-status can actually be less hard for agents, right? So I think there's an enjoyability element and then an identity element.

Mm-hmm.

And actually, one of your founders from Applied Intuition wrote a good blog post—an essay—where he says, “Keep your identity small.” I think that's wonderful overall advice for this period of time.

Mm.

You're more adaptable if it's true.

Mm-hmm.

But I think your overall view—that there are a lot of unsolved problems, and making an abundance of software can better address that—I strongly agree with. And one thing that actually is near and dear to the audience that is really unsolved is that we've broadly been thinking about what happens if you have abundant code generation. In all of our teams, agent-first engineering management and thinking about code quality is an unsolved problem.

Mm-hmm.

Yeah, and we'll get there. It'll be your cohort, and we'll get there. What do you view as the major problems?

2. Human Attention Becomes Scarce

Well, the anxiety that I see is that if you can generate an enormous amount of code and no one is reading it, you don't know the quality of the code. Nobody deeply understands the codebase, and there's more fragility, right?

Mm-hmm.

It's like the slop problem, but instead of it being vibe-coding slop for random websites for nontechnical people, it's vibe-coding slop in my actual production codebase for every lazy engineer—which is every engineer. I actually do think ticketing systems are at risk, but I think the broader problem that Jira could go solve, or a new company could go solve, is that nobody knows—

Mm-hmm.

—how to manage that issue of human attention to engineering, and there are a bunch of ideas—

Mm-hmm.

—like testing, smart review—just let agents do it—formal verification. But I think it's open season around this really, really big problem.

Mm-hmm. I think the one other thing people are bringing up that I don't quite buy is that agents are already making big decisions for vendor purchases and things like that. I think somebody near and dear to your heart posted about that, and I think the statement was, “Agents are increasingly making decisions about what software people are using.”

And really what that is is, well, you have a partnership with cognition or your cloud or whoever, and as part of that partnership, you spin up a Superbase instance, and you use very specific tools because you have a partnership to do that. And that's always happened, right? If you're using Airtable and they're on AWS, you're spinning up an AWS instance without knowing about it, right, in the background.

So I also think that whole notion that in the short run agents were making these choices is overstated. I think in the long run it's true, but then you get into all sorts of agentic commerce decisions: Do they understand your persona and what you actually want and need, and all this stuff?

So I just feel like we're in a little bit of a noisy moment. I'm somebody who's very pro-AI progress and a believer in all the changes that have happened and are coming, but I think we're having a lot of overstatement now of what's actually happening in the world. Part of that is this AI apocalypse and this giant reconfiguration. And part of it is, um, you know, extrapolating that the future is here already, when in many cases it's just, hey, we did a big deal or whatever. So I just think people kinda need to... Or, or you know, the Malt book stuff where you're like, “Yeah, a lot of that seems human-generated.” You know, in terms of the emergent behavior. So I don't know. We're, we're, we're in this odd moment where I feel like this was the month of hype in a way that we haven't seen in a while, where a, a bunch of stuff got overstated in all sorts of ways and people believed it. And by people, I mean, like mainstream media and others are like, “Oh my gosh, look at this behavior of, you know, these agents trying to cut out humans from their forum where it's Reddit-like,” and blah, blah. And you're like, “Okay, like maybe you should see where the posts are coming from in some cases.” And it's exciting, by the way. Don't get me wrong. I think there's very exciting behavior that's happening. I just think, you know, a subset of it was planted for marketing purposes.

Yes, certainly. I think people are also figuring out that there are things that tap into deep emotional reactions that people have to their view of things that feel very human, right?

Mm-hmm.

From a marketing perspective.

Yeah, yeah.

Mm-hmm.

And that's clearly one of the things that's happened around the Malt book stuff. I also think that one of the things that actually happened was that the idea that demos are different from the reality of the full software that you need has not quite arrived on many equity researchers' desks, right? And so I'm like, “Guys, your whole job was to think about the structural advantages of your businesses and what is going to compound.” The theory of competitive advantage didn't just poof, disappear, right?

Mm-hmm.

Software markets have been a fight about how to do things and how to distribute to customers, as well as a battle over how to produce code, for a long time. So I feel like that has been missed a little bit. But I do think in the long run, the fundamental thing—that the bottleneck on production of expensive-to-produce software being loosened—is really cool, right? It just means that if you think of it, there's a lot of embedded points of view in software on how to solve a problem, right?

Mm-hmm.

You know, if it's engineering or enterprise sales—not a very software-y problem—or general productivity, right? Notion is a way to do things. It's a building-block system, but it's definitely got a point of view. And so if you reduce the cost to express that point of view in software, I think it's cool that we're going to see a lot more ideas.

That's amazing.

Yeah.

And again, I think it's a revolution. So don't get me wrong. I've been involved with coding companies really early on, and I'm very excited about everything that's happening. I think it's transformational, and I think it's revolutionary, and I think it's really important. I just think we had a month of kind of bullshit hype.

3. AI Growth Defies The Hype

Okay. So if we ignore the noise of the last month, where people got a little frantic, what do you think is a signal that people are not paying attention to enough in such a noisy landscape? You were telling me that the growth pace of the biggest companies is still underpriced.

Yeah. One thing that Jared on my team put together that I thought was super interesting was that he pulled data from Capital IQ, where they just predicted some projections on OpenAI and Anthropic. And then he graphed out—and maybe we can share these graphs as part of this episode—how long it took different companies, in years, to go from $1 billion in revenue to $10 billion in revenue.

For example, ADP took 20-something years to grow from $1 billion to $10 billion in revenue. The next wave of companies, like Adobe, took about 20 years to go from $1 billion to $10 billion. Then you fast-forward in time and you have things like Salesforce or SAP, sort of an even more modern cohort, and they took 8 or 9 years. Microsoft took 7-ish or 8 years. Google, Meta, and AWS took a couple of years—3, 4, 5 years. But the AI labs did it in roughly a year, right?

It's a wild chart.

It's a wild chart, and so we should add it, right? But you just see it go from 20-something years with Adobe to a year for the AI labs. And then if you look at the projections that are sort of the public projections, they aren't necessarily the company-driven data, but the public projections on where the labs will end up, or how long it'll take them to go from $10 billion to $100 billion in revenue—

For Microsoft, that was something like 27 years. For Google, it was over a decade, same with AWS, and roughly the same for Meta. And then for the AI labs, it's like 3, 4, 5 years. It's very fast.

And so we're seeing the fastest time to real, massive revenue that we've ever seen in the history of software. It's just these insane curves, and again, we should post them. Part of that, I think, is that the internet has created this global pool of liquidity, and suddenly every customer is online. It's much easier to distribute than it's ever been. There are more people with access, higher GDP, and lots of drivers for that. But then, simultaneously, you're creating enormous business and user value at massive scale, and these capabilities are so rich that you're seeing this take off in terms of revenue. It's unprecedented. It's really impressive, and I think people are ignoring the revenue and usage side of the equation.

The other thing that we actually put together was the collapse in token pricing for equivalent models. I think this was done initially by David, who worked for me, and then Shrin. For example, we looked at the cost of a GPT-4-level or equivalent model a year or two ago, and basically, in 21 months, it went from 37 bucks for a million tokens to 25 cents. Pricing dropped by 150× in 21 months.

Then we tried to extrapolate that curve, but obviously people aren't really using GPT-4-level models anymore, even though they're 2 or 3 years old. We looked at o1-equivalent models, and the cost of a million tokens on an o1-equivalent model in December 2024 was about 26 bucks. In November 2025, it was 30 cents. So we saw another 88× drop—not 88%, but 88 times cheaper—in 11 months for that next generation of models. We're having pricing collapse on the token side while we're having revenue ramping insanely on the usage side. That's insane if you think about it—the pace of the shift in cost, revenue, utilization, and everything.

This gets back to the fact that I'm incredibly bullish on everything that's happening. It's more about modulating that against this odd over-extrapolation of what's actually happening, the actual capabilities, or what these things are really doing.

Yeah. I think one thing that people miss in the bear case in all this stuff is, as you said, revenue numbers, which are hard to miss. But then there's actual token-inference count, right?

Mm-hmm.

If you look at where the inference is happening, it's either happening in inference clouds—Baseten, Model, or Fireworks—or it's happening at the very large model providers.

It's insane.

It's happening in Elad's brain, which is still much more—

It's all happening up here as well.

—two orders of magnitude more efficient.

And humans.

And humanity in general.

Right.

And humanity in general, yeah. That's true. In terms of power utilization, the human brain is really impressive. What is it, tens of watts? 20 watts? What's the power utilization of a human brain?

I don't want to look it up right now. It is—

Something like that.

Two magnitudes.

It's like 10 or 20 watts, I thought.

I think, to the point of real data, the inference clouds are growing 1,000× in terms of consumption, right? And then they're getting more efficient, so revenue grows at some lower rate than that, but it's wild.

It's 12 to 20 watts of power, which is comparable to a dim light bulb or a computer monitor in sleep mode.

Yeah.

It's not even—

Like a computer monitor—

It's when your monitor is sleeping. That's the amount of energy that your brain is consuming as it does all these crazy calculations.

It's one blade of one GPU fan in one of these data centers.

Yeah.

That's how I think of it.

It's nuts. I feel like Noam Shazeer's brain, though, is probably consuming 1,000 watts.

Well, I think that's great. I think we have a lot of efficiency work to go.

I meant it the opposite. He's so smart, he's probably consuming more energy. But to your point, maybe he's more energy efficient.

Oh.

Maybe he's at 1 watt, and I'm at 1,000 watts or something.

I meant for the computers.

And then you get the algorithm just going better.

We're all stuck until the brain-computer-interface work improves. But—

Really good interface.

I'm just interested in how much efficiency we can get out of the models.

Yeah. Obviously, based on the human brain, there's a lot of room.

4. Market Cap Becomes A Weapon

You know, one thing I do think about: I was talking to a friend who leads a bunch of purchasing at a traditional large enterprise this morning, and he was like, “This whole thing is overstated. We're so committed to all these big enterprise vendors.” His other view was that the incumbents have the money to buy and go fight back on these dimensions.

One thing I immediately thought of was that reflexivity in markets is such a good concept. And here it's like, well, they do unless they don't have the market cap to do it, right?

Mm-hmm.

To your point, first the labs, but then a series of the very best application companies: if they're growing to a billion-dollar run rate rapidly and valuations grow in concert with that, then I do think there's a question about whether or not you have the currency to compete, too.

Yeah, I'm already seeing that in the San Francisco housing market, right? San Francisco housing is starting to rise again, in part due to, I'm assuming, outcomes from the lab tenders and things like that. Suddenly you have these companies that are worth hundreds of billions of dollars out of nowhere in a few years, and as employees are selling into tenders, there's this new influx of cash in the ecosystem.

There's also NVIDIA going from tens of billions or 100 billion to trillions in market cap. There's just this shift happening right now in terms of scale.

And there's an interesting question, actually. This is one other thing that we looked at as a team, and maybe I should just publish all these slides. We basically asked: What proportion of GDP is tech, at least in the U.S. economy? How has that grown over time? And what has that meant in terms of market caps?

If you look back to 2005, Google was worth $100 billion, and Exxon was the world's most valuable company at $400 billion in market cap. It took until 2018 for Apple to be the first company with a $1 trillion market cap ever, right? Everybody was shocked that anything could get to a trillion. At the time, tech represented about 30% of the S&P. Before that, it was around 10% back in 2005.

Now the top 8 tech companies have about $23 trillion of market cap, and they make up well over 50% of the S&P in terms of value. At the same time, they went from basically 4% of GDP in 2005 to about 12% of GDP today. So the question is: What proportion of GDP eventually just becomes tech?

AI is a driver of this, right? You're taking services and certain types of jobs, augmenting them with AI, and converting them into effectively software spend or tech spend. You can make different assumptions about growth rates, and based on that, you can end up with anywhere from 15% or 20% of GDP to 30% of GDP in 2035. But that means that the market caps of these tech companies get even bigger.

It's a metric for how big these things can actually get as they aggregate portions of GDP. I think that's the other lens that people aren't really thinking enough about in terms of what some of these terminal values might be 10 years from now. How much more can things grow, and what are your assumptions around that basis for growth?

This gets back to that ramp-up into revenue. It's a very interesting set of questions that we've been asking on my side, just in terms of these meta things—what are the bigger trends that people may not be paying attention to that may be super interesting?

Okay. Well, then, I have a set of structural questions about how to invest based on this for you. Asking for a friend: my funds are small. I think there are good implications and bad implications based on what you said.

One might be: If everything's going to get a lot bigger, $1 billion is no longer late stage, right? That's just a marker on valuation that it's the beginning—

Well, even now it's not late stage, because people are raising at a $1 billion valuation with $2 million of revenue, right?

Right. Well, you can decide that's a—

I know of at least one company like that.

You can decide whether that's a smart idea or not, right? But the point we would absolutely agree on, I think, is just that the runway for some of these foundational companies is much larger than the conventional wisdom.

I think we've already believed that, though. I think everybody shifted. I remember I wrote a blog post 15 years ago or something, 10 years ago, that basically talked about how hard it is to get to a sustainable $5 billion market cap.

Because at the time, basically, once every couple of years, a company would actually get to that and stick with it. This is back to 10 or 15 years ago, when the biggest market caps were in the hundreds of billions at most, and low hundreds of billions, right? Then we saw everything grow 10X over the last 15 years, right? You suddenly have trillion-dollar market caps, and that means there are a lot more companies also worth $100 billion than there used to be in tech.

So I think in general, we've seen these shifts happening already. The reason that we were asking the question internally about how much bigger these things can get is because that has further implications. How many more trillion-dollar companies can be supported? Is it 2? Is it 3? Is it a dozen? Is it 50?

Relatedly, if everything gets pulled up, how do you think about how you invest over the lifetime of a company in general? Or how do you think about that as a founder in terms of the end state? Then there's also a related question of what's the actual fail rate of startups. Should the fail rate go up or down in that world? You could argue it either way.

You could argue that the fail rate should go up because more and more value is getting aggregated into platforms, as has traditionally happened. Every single platform shift has seen a commensurate forward integration of that platform into the most important vertical application.

As an example, Microsoft very famously, on its OS, forward integrated into the Office suite—Excel, PowerPoint, and Word. They killed or bought companies in those market segments, and that became Office. Then they redistributed it alongside the OS.

Or Google forward integrated into vertical searches. They had a platform, and then they built out travel, local, and all these things. It's not surprising that the labs will forward integrate into the most interesting applications on top of them. You're already seeing that partially with code, but what else is coming there?

What implication does that have for people running startups? Which of those verticals are durable and defensible, and which of those are going to get eaten by the labs? You could make arguments in both directions in terms of whether more of overall GDP will aggregate into a smaller number of companies, which is already what's happening, right? Just ignoring the labs, that's kind of what happened with Amazon, Google, and all these things.

Or do you end up with this broader tail effect as well, where things happen simultaneously? We also have a lot more startups that are worth more because there's just so much more market cap to go around, but the internet continues to provide this global liquidity.

To me, I think the tail dominates because the surface area of what you can address with technology is just increasing more rapidly.

But is that true? If you actually look at market cap, it's very much a power law, right? The head and torso aggregate almost all the value. That's actually true of customers too, although people tend to misunderstand that.

I remember that book, The Long Tail, or whatever, about the internet. The claim was that the long tail really matters, and then you'd add up Google's ad revenue and you're like, "Actually, it's all the head and torso," right?

I feel like there are these head-and-torso effects that keep getting ignored. It's like Paul Graham's power law on startups, right? Most of the value of YC is probably 5 companies—like, 80% of it. I'm making it up, right? But it's really concentrated, so why would that change in this era?

Yeah.

I don't think it changes in this era. I think that it depends on what your measure was. If your measure is how many $100 billion businesses there are, I think there's a lot more, right? It doesn't mean there are fewer $100 billion businesses. Actually, there are more because the surface area is growing. At the same time, the distribution of how much is in the head is probably the same, and those are even bigger.

Yeah, it's possible. Yeah. It's an interesting question.

Do you think for investing, there's a thing that's good for me and then perhaps bad for me, or just a question for growth-stage investors? The time to market leadership and to revenue scale, I think, is compressing.

I mean, it's not "I think." This is happening. We have a large handful of companies that have gone from zero to a $100 million-plus run rate faster than SaaS companies that we'd seen 10 years ago.

Some set of companies that look like this are durable, and for some, leadership can still flip, right? A question might be—

Okay.

Is it you or is it Ant or is it OpenAI over time, to your point, that actually you could grow to a billion dollars of revenue and still face that question?

Mm-hmm.

And that is, I think, a risk that maybe some of the growth ecosystem would find to be a new thing, versus category leadership at a certain scale, which felt unassailable 10 years ago.

5. Founders Need Exit Discipline

I think there are 2 interesting historical precedents to this. One is the internet wave where, in 1999, 450 companies went public, and in 2000, another 450 went public. There was, say, one to 2,000 companies that went public during the internet age, and maybe a dozen to 2 dozen of them are still relevant, right? Everything else roughly died or got bought.

Then you fast-forward 10 years, and you saw the subsumption of things that people thought were unassailable. In social networking, people thought Friendster and then MySpace were unassailable, and then Facebook won.

In payments, I remember when I invested in Stripe, everybody said, "Why are you doing this? Braintree exists, PayPal exists, and all these things exist. Why would you ever invest in another payments company?" Of course, that ended up being the winner, or one of the winners, right? Payments is so big, it's a fragmented oligopoly.

I feel we've kind of seen this story before. As a founder, it's really useful to be asking about 2 things. One is, what is the durability of your business? Number 2 is, how should you think about when to exit if you're going to exit?

Often for companies, there's about a 12-month window where your company is the most valuable it will ever be, and then it crashes out. For a very small handful of companies, the answer is you should never, ever, ever sell. For most companies, the answer is you should sell when the timing is right.

The question is, how do you know when the timing is right? Ultimately, you're going to hit a point of maximal value, and then it has a real potential to die, even if it got enormous traction. That was the internet wave of the '90s.

I think too few people are thinking about this. One tip for founders, from a hygiene perspective, but also just a way to make it a non-emotional discussion, is to preschedule once or twice a year the board meeting where you talk about exits.

That way it becomes non-emotional. It's not about "We're going to exit," and it's not like, "We should exit." This has actually been Horace's advice, I think, from when he was running Opsware. You just set up a non-emotional meeting once or twice a year. You're like, "Nope, still not time to do it."

Or you say, "Oh, you know what? Actually, the competitive dynamic has shifted dramatically. Somebody's come to us with an offer that's higher than anything we'll achieve over the next 5 years. Now's the time to do it," right? I think it's useful for you to be thoughtful about that.

Again, the default for a small number of companies is never, ever do it. For almost everybody else, it's worth considering at one point or another because you may otherwise get stuck with something that isn't working for a long time, or you may get crushed by a competitor. Many, many years of very hard work can just go down the drain.

I think this is an interesting point about the comparison, especially between the internet age and SaaS—or, I don't know what you call it, the cloud age—from the last decade, which are more similar. I wasn't around for this era, but from my research and from working with a bunch of people in that period, you're not old enough for this era either.

AOL was the internet for a moment, right? Yahoo was the—

Mm-hmm.

web's front page. Netscape was the browser. Internet Explorer was the web runtime.

eBay was the market. I think there are a number of these—

Yeah, and AOL exited at the exact right moment to Time Warner, right?

Right.

At their peak valuation, yeah.

Right, and I do think that founders and investors may over-rotate on the SaaS era. It did feel like, at a certain scale, in the internet era, there was a period of time when growth was the default—growth at a wild speed. That was not true in SaaS land, so it was more incremental, and beyond a certain scale, it felt very protected.

But I think this probably does look more like the internet era, where the question is: Does that growth compound to a control point where you're a very special company? Or do you actually think about exits in a different way?

Yeah, and if you even go back to the 80s, you had Lotus. I don't know if you remember this company, Lotus.

I have implemented Lotus 1-2-3 at an enterprise business as an intern.

Okay.

Wow.

So Lotus built 1 of the first spreadsheet products, and it grew explosively. It got into the hundreds of millions in revenue really, really fast, and this was the 80s, right? Then, a couple years later, it basically collapses into the arms of IBM, and Microsoft launches Excel and takes the whole market, roughly, right?

Again, it looked like a very durable business. It was the killer app on computers for its era, and then it just died. It didn't die—it ended up with a great exit to IBM—but it no longer exists in reality, right?

I think the same thing is going to happen for a number of companies of this era, and the question is which companies. That's a really hard question, right? Who knows? But for some companies, you're starting to see cracks, right?

For companies with these cracks, as the market structure shifts, as you see shifts in what the labs are doing, as you see shifts in usage, and as you see shifts in differentiation and defensibility and all the rest, it's a good time to ask, "Hey, is this my moment? Are these next 6 months when I'm going to be the most valuable I'll ever be, and then I'm at real risk?" If so, you should think seriously about what to do with that.

I view this not just as a right-now thing. Every 6 months, there are going to be shifts worth considering, and that's why you should pre-schedule the board meeting so it's not emotional. You're not putting something on the agenda and everybody's like, "Oh my God, do you want to exit? What's going on? Are you upset? Are you worried?" It's more like, "Oh yeah, we booked this 6 months ago, and we booked it a year ago, and we booked it 2 years ago," whatever it is, "and this is just when we talk about this stuff." So we can have a very logical, emotion-drained conversation around this stuff.

And maybe, again, in comparison to the internet era, as to why to think about it more now—

Well, people in the internet era should have thought about it too.

Sure, sure.

I mean, Mark Cuban did this. Mark Cuban's claim to fame is that he sold a company. Let's put it this way: It was early in terms of product, and he sold it to Yahoo for a few billion dollars. Then he collared Yahoo stock so that, as the stock dropped, he didn't lose any money.

It was 1 of the best all-time financial engineering moments in tech history, right? That's what made Mark Cuban a billionaire: He sold at Yahoo's high-water mark, and then he kept all the value as it collapsed in price. He was 1 of the few people who did that during that era, but people were thinking about it.

I think most people missed it, right? In retrospect, thinking about the flips that made it happen, where the ground was moving a lot, is useful, right? Because you have to answer the question: Am I that company or not? Or is my acquirer that company or not?

In the internet cycle, you had new distribution, new performance, new interfaces, and changing user behavior. It was just everything happening all at once in new exploration. That was not true in cloud land.

Yeah.

Right? It was more of a replacement market, and then niches that you could cheaply distribute to. It was a new business model. SaaS is amazing.

Yes.

But in AI, it's like, okay, is the next major capability jump from the labs going to screw me and reset the leaderboard? That is an important question to ask yourself.

Then there are also surface-area questions, right? Agents versus IDEs, voice as a default—there are things that change in product experience that also could reallocate power.

6. Bundles Create Durable Defenses

The best way to defend against this is to build a bundle—to build a multi-product surface area for your company so that you cross-sell multiple things into the same organization and become a default part of the workflow. That's the best way to defend against this, because then you're being used for 5 or 10 different aspects of the vertical or application that you're in, versus having a singular thing that's easy to clone or copy or for people to displace.

So I think the defensive advice on that is to do that.

Yeah.

Bundles are often seen as offensive, but I actually think they're amazing for defense, you know? I think that's the other thing that people are underdoing a little bit for some of these vertical applications, and that's going to be the way to win long term or to defend long term.

Well, I actually still think—I sound like I just hate the SaaS era. I think it is a mistake that people took as conventional wisdom from the SaaS era and applied it now without thinking about it, whereas the advice was, "Do 1 thing well."

Oh, the point-product thing?

Yeah. It was, "Do 1 thing well," and then people buy you, and then don't go and compete with a million things.

Yeah.

But we think—

That was bad advice. That was always bad advice, though. Substantially, in SaaS companies, it was bad advice, because before that, powerful companies were very acquisitive and very multi-product, and it was just the SaaS era where it became this singular thing.

I think the other piece of it is that the rate of change and the velocity of the technology during the SaaS era was just slow.

Yeah.

It was just like, "Let's just keep building out the internet." That was kind of the SaaS era, right? The difference with AI is that the velocity of change is so high that what normally would have taken a decade, with a normal decade-long displacement cycle, is now happening in a year or 2.

That's really the reason that these things are so turbulent. It's because the technology is shifting so dramatically, so quickly, and that's just part of scaling laws, part of reasoning, and all the post-training stuff that's been rolled out.

There's just been so much innovation in such a compressed period of time that that's the reason things are turning over, and things that normally would have taken a decade are happening in a year or 2. That's why we're seeing these displacement, or potential-for-displacement, cycles.

But that also means that, as a founder, your mindset should shift into this new-world framework. You should say, "Okay, if every 2 years is 10 years, I need to think really quickly about changes that are happening. I need to react to them in all sorts of ways."

Yeah.

It's just back to—you know, it's a fun, interesting, and exciting time, and I think it's going to be an amazing decade of transformation.

Yeah. I do think maybe one way to think about a lot of the defenses that people did not use in the software era, or the last software era, is: What does not depend on my little feature set just incrementally growing?

Platforms, ecosystems, networks, bundles, even hardware, like you described with Samsara—that feels like nontrivial control points. So maybe the takeaway for me and Elad hanging out today is: Hey, don't over-rotate on the last month.

But also, be intellectually honest about the position you have in the market and, in this speed-of-change era, actually think about what the control points are.

Yeah, well, that's coming. That's shifting. It's going to be fun.

Okay. Have fun.

Yeah. See you later.

Find us on Twitter at nopriorspod. Subscribe to our YouTube channel if you wanna see our faces. Follow the show on Apple Podcasts, Spotify, or wherever you listen. That way, you get a new episode every week. And sign up for emails or find transcripts for every episode at no-priors.com.