Is The Software Margin Era Over?

Contrarian shouldn't be an identity or a strategy. I tell people you should focus on finding companies that are good. Sometimes companies that are good are misunderstood, so investing in them will feel quote unquote contrarian because other people don't see. But if you just say I'm just going to do the opposite of what everybody else thinks as a strategy, that seems silly to me.

The low stakes nature of this experiment is precisely what made it tolerable in this moment. So if it was applied to exactly what you mentioned, higher stakes industries like healthcare, fintech, similar failures could actually result in some serious consequences. It could range from financial fraud to loss of life.

I've met some people who are like, until AI is perfect, which is a higher standard than we hold humans to, it should never be used in these like, well, that's probably not what we're going to end up. And I've had other people who said, look, other people's pain and suffering is the cost of us getting this stuff right, and so we should use as much of it as we possibly can and live with the consequences also feels irresponsible.

Welcome to the Learning Corner by Precursor. I'm Mia Farnham and I'm joined by our GP and managing partner Charles Hudson, where we will be talking through the articles and Goodreads that caught our team's attention over the past week. Our first topic is a recent Wall Street Journal article about Taco Bell's bumpy rollout of voice AI across stores that was kicked off over a year ago. Customers have reported glitches, long wait times, and even trolled the system with prank orders like thousands of cups of water. Taco Bell's chief digital officer admits the tech is still inconsistent even in his own experience, and says humans are often still faster in busy stores. The company is now rethinking its strategy, leaning into a hybrid model where staff decide when to use AI versus step in themselves. They're not alone in this decision. In fact, McDonald's and Wendy's have both had to revise or pause similar AI pilots. Taco Bell's parent company, Yum Brands, is now working with Nvidia to improve their strategy. But across the industry, everyone's still figuring out where voice AI can thread the needle on value add. Similar to last week's episode where we talked about the high failure rates of AI pilots, stories like these suggest we're starting to make progress in figuring out what works and what doesn't when it comes to enterprise AI rollouts. Anything else that you wanted to note here? From your perspective.

I love a good anecdote. And the 18,000 cups of water cracked me up. I continue to think this is a positive development because my sense is what's happened in a lot of companies, talking to my friends who work in bigger public companies is the message from the board and the CEO was like, roll out AI, get on it asap. And vendors said we have an AI product that can do everything you needed to do, just like turn the whole operation over to AI. And many of those vendors had no real world experience, but I think either naively or in a salesy way had confidence in their product's ability to deliver. And what's happening is like with any new technology, once companies actually use it, they can figure out the limitations. So I think for those who poked fun at Taco Bell, I think what Taco Bell probably learned is that AI can do a handful of things super well today on its own, a handful of things pretty well with human supervision. And there's probably a set of tasks and things they hadn't thought about, like these prank orders that a human would be able to detect that it's not ready to handle. And so my hunch is that Taco Bell's conclusion will not be we should not use AI at all. It will be we should be more judicious and thoughtful about, as their Chief Digital Officer said, where humans put their hands on the wheel or where we let AI do its thing. And I think there are going to be more stories like this over time as companies. And look, I think not to belittle Taco Bell's business. Drive thru is a pretty safe domain in which to test voice AI. Like no one's going to lose their life or lose their livelihood if you get my chalupa order wrong.

Yeah.

And I think we're going to continue to learn that there are limitations for what unsupervised AI can do in actual production deployment environments.

That's a great segue into my next question because when I was looking through the comments across X and scrolling through, there was lots of different quippy things coming in and out from frustrated customers. But it was all done in kind of in like light hearted fun. There were some people who kind of pointed out that the low stakes nature of this experiment is precisely what made it tolerable in this moment. So if it was applied to exactly what you mentioned, higher stakes industries like healthcare, fintech, similar failures could actually result in some serious consequences. It could range from financial fraud to loss of life. So curious to kind of blow that out a little bit more like what kind of guardrails should exist in these critical industries, especially when others are able to experiment more freely. Like Taco Bell can play around with this and get it wrong a hundred times and it's okay, can you get it wrong 100 times when it is applied to health care and fintech or even broader industries, et cetera.

I think it's harder in healthcare in particular because there are some clear and obvious use cases where AI seems to be as good or better than humans and more consistent. But also it makes mistakes and hallucinates. I think the challenge is I don't think there's a universal belief in what the trade off between those two things should be. I've met some people who are like, until AI is perfect, which is a higher standard than we hold humans to, it should never be used in these. And I'm like, well that's probably not where we're going to end up. And I've had other people who've said look, other people's pain and suffering is the cost of us getting this stuff right. And so we should use as much of it as we possibly can and live with the consequences also feels irresponsible. So I think socially we're going to have to work out what are the acceptable error rates or if we're going to use a lot of AI. Ultimately somebody, we talked about this last week, somebody has to bear the risk and liability in the end. And with humans it's, I wouldn't say easy, it's straightforward. It's like the person who did the thing bears the responsibility. And what happens when the person who did the thing isn't a person but an AI agent or a bot or an algorithm. Like who's going to bear that liability? I think we'll, we'll sort this stuff out.

And that's why I think in health tech we're seeing a lot more applications of voice AI or AI more in back office operations to kind of streamline, note taking, transcribing, et cetera. I think it'll be a lot more of a discussion when it starts to bleed into, I don't know, like on a robotics nature of like taking part in surgeries, et cetera, all of which like already happen today. Like there's a level of machine tech that do really in depth surgeries that human hands can't specifically do. So that isn't necessarily new to everyone. The topic that we talked about last week around like mental health and therapy and chatbots, et cetera, OpenAI released an announcement this past week of trying to create some guardrails, especially with all the news that's kind of come out of difficult consequences of chatbots bleeding in a bit too freely with users in areas that maybe they're not quite ready to do so. Lots of things to be ironed out there for sure. Our next topic is a recent post from Martin Casado on X where he shared the brief but quite impactful tweet. The idea that non consensus investing is where the alpha is is actually quite dangerous in the early stage. Follow on Capital tends to be more and more consensus aligned. Obviously this calls into question the legitimacy of the long held belief that non consensus investments are the strongest ones and certainly brought people to the comments section to make their opinions heard. Martin later posted quote I'm much less interested in whether consensus based investing works and much more interested in the dangers of being non consensus, he says. I actually believe the following two statements are roughly true. A VC consensus is often hype driven and totally disconnected from markets and b early markets are efficient and therefore good deals get identified early and that's reflected in marks. In other words, deal hunting is the wrong approach in vc. Charles I sense this is a topic that was widely discussed prior to Martin sharing his post. Wanting to get your perspectives on this long standing discussion or debate.

I think this is one of those evergreen topics. This and like should you send monthly updates seem to like bubble back up into the discourse. I think people seem to forget the other half of being contrarian. It's the and right part. I've never understood this. Like we're a contrarian investor. I'm like you can't short private companies. Yeah, you can only go long. You can go long or you can not participate. Those are actually your two options. So until you can short private companies, you can't really be contrarian. Shouldn't be an identity or a strategy. I tell people you should focus on finding companies that are good. Sometimes companies that are good are misunderstood and so investing in them will feel quote unquote contrarian because other people don't see. But if you just say I'm just going to do the opposite of what everybody else thinks as a strategy, that seems silly to me. Like very silly. At the same time, I think VCs are very good at identifying the kind of people that they think will be successful, which is different from the kind of people who are in fact in the end successful. And I think right now I think consensus good businesses. There was a period of time where it was very clear that Stripe was a consensus good business investing at Stripe at the moment it became a consensus good investment was a good deal because the company has grown in value probably 150 to 250x since the point it was obviously going to work. I'll give you another example. OpenAI at a billion was probably a consensus good company. It's up 500x as of today from then. So Martin's right. The funny thing now is that we have a relatively small set of venture firms who define taste like they have all of the money for follow on capital. So consensus is whatever they like. So if Martin's statement is you should invest in the kind of companies that I think are good if you want my follow on dollars, I think that's probably true, yeah. It's unlikely that Martin's going to take a 16Z's follow on dollars and put them into a company that he doesn't think is good. But I think by the time it gets to Martin, that's different than what it looked like or Martin rather it's different than what it looked like when the company started. So I think if you're going to do things that are weird, a couple of things have to be true. One, you have to believe eventually the company will be well understood. Two, that it will in fact be a good company. And three, that you and the other people around the table have the resources to finance that company through that dip where it's maybe hard to finance. So I don't know, I think whenever we do things that are not well understood or kind of hard to get your arms around, I just assume there might be one or two financing rounds for that company that are really difficult. And either we or the other folks in the syndicate have to have a plan for what we're going to do if we hit that speed bump.

I mean, we just had this conversation, you and I, before we started recording this podcast of evaluating a deal and wondering. I felt like it was non consensus. But I did question how long will it take for folks to realize that it is good? And that's exactly to your point of like how many runs of funding is it going to be a difficult journey? And then also brought me to this question, like when you're a later stage investor, I assume consensus is the only thing that you want to invest invest in at that point. Because early stage, that's the whole way you can make your money in alpha. But at series B and beyond, you shouldn't be investing in truly non consensus things. Or do I have that upside down?

From your perspective, my sense is there's like two different jobs in venture. There's the early stage job which is identifying at a very early stage the companies that are likely to be good important companies five to ten years from now. And I think in later stage investing the job is how do you find these special companies separate from the good ones? Because basically like most companies that can get to 20 or 30 million in ARR, they're good companies but they're not all great companies.

Yeah.

And the question's like how do you identify the truly great ones? Because you only get a few shots on goal and the job of that late stage investor is really to find the special companies. And so I don't know, I think it's, I think this is the hard thing about venture today is what's consensus is really the decision of a handful of very senior partners at a very small number of firms and like the tastes of those people have a really disproportionate impact on what gets funded in terms of follow ons because they have all of the money. I like looked in the comments section, I was like guys, what Martin is really telling you is he knows what he's looking for.

Yeah.

If you want his money it should look like things that he thinks are good and if you show him stuff that you think is good that he doesn't think is good, probably not going to get funded.

The last piece on this, I know Yoni from Slow Ventures. He shared a post on his blog where he dove a little bit deeper into the post in more depth because I think there was a lot of comments in the comment section of. Oh, Martin has it all wrong. This is so like anti venture what he's saying. And Yoni was kind of like you're misunderstanding the purpose of what he's trying to deliver here. He said, quote, I think it's better to think about non consensus narratives that are overlooked rather than betting against the crowd outright or finding undervalued companies. That's why I distinguish between contrarians and non consensus. I feel like what you just mentioned was pretty aligned with where Yoni lands.

Yeah, I, I like Yoni's writing. I often find that he puts the word things I've been thinking about. I, I think our job at Precursor is to find founders that we think are good working on problems that we think are interesting that are maybe not well understood. And it's usually because those problems are a year or two away from being obvious to everybody else. But again the founders have to turn out to be good. And so it's funny sometimes I read those comments as like, oh, you guys want venture? It's like one of my favorite quotes from the Wire when Marlo's in the convenience store and he's talking to the security guard and he's like, come on man, I'm trying to do my job. And Marlo says, you want it to be one way, but it's the other way. And I think there's just a lot of things that have venture where people are like, I wish it worked this way. I'm like, me too. But it doesn't, right? And you have to recognize that, especially if you want to do different things, you have to actually understand how the business works today and what it would take for a founder with a different background to be interesting and fundable by a large venture fund. It's a lot to get those people to a place where those businesses are consensus Interesting.

Well, our last topic is a substack written by Sam Schillis titled Has the Economics of Software run Off a Cliff with AI? He explains that as AI tools make software faster to build inference costs, so the cost every time a user interacts with an AI are driving margins down. The old build once, sell infinitely model is completely eroding and software is starting to feel more like a cost heavy operational business. Throughout the piece, he even compares the future of engineering to manufacturing, where a small group of top engineer designs high leverage systems and the rest become lower paid interchangeable line workers. Sam shares a word of caution, which is product and engineering leaders who don't embrace AI risk falling behind the edge will go to creative hybrid builders, those who can understand both code and experimentation, not those who cling to what was historically the norm. This was the first time I had heard the manufacturing comparison when talking about a future for the engineering workforce powered by AI. I wanted to get any perspectives from.

You on this piece of I like Sam's writing too. Anyway. He's seen a lot from, you know, Google Sheets to now, and I think most talent markets that we're in end up with this highly skewed star system. Sports has gone that way, music has gone that way, art has largely gone that way, where the people at the top make an extraordinary amount of money relative to the average person in their field. And I think it's interesting to think about, well, what would software development look like with a similar star theory, you would end up with a handful of really highly regarded, highly compensated, hopefully highly talented software engineers who make a lot of money for some special level of Expertise. Maybe it's system design, maybe it's understanding how to get the most out of AI coding. And then you have a lot of people who do a similar job and make a lot less money. You know, I think we talk about the extremes are that like, everyone's going to become a software developer or there'll be no coders, there'll be only AI, or there'll be one person in the world who writes all the software. All of these weird extremes. But I do think the software job market is going to look different, but I don't actually know that I know the answer. The one thing he said that I do think is definitely going to happen is that you're going to have developers who will end up spending more in tokens than they're paid in salary. Like their compute consumption will exceed their. Because I think their compute consumption will grow faster than their wages. So at some point those two lines will. Those two lines will cross. And I think that will be interesting when you have a person who's paid, let's say, $250,000 a year who's generating a 500,000 DOL compute bill, but that person might be doing the work of what would have previously taken 10 people.

And like, how much of revenue is attributed to that token spend? I guess this $250,000 that I'm spending on them is like a drop in the bucket for what they're outputting cost.

I continue to believe the hardest thing VCs who've grown up in the SaaS era, including myself and I try to wrestle with this a lot, is we grew up in a world where the marginal cost of delivering software to the marginal user was effectively zero, which is how you got these sky high. And the products were hard to. And we've kind of inverted that now. The marginal cost of supporting a user is not zero. There's compute that goes with it. And everyone's like, oh, compute will go down. I'm like, but what if people just use more of it? The whole Jeevan's paradox kind of thing. Like there's nothing that says that the compute cost per user will go down just because the cost of compute goes down. Usage could go up faster than prices go down. So we don't, we don't know what that is and we don't know what's going to happen to customer willingness to pay when they know that these products are like, less hard to build. So we're like tinkering with cost and we're tinkering with value. We're tinkering with all the assumptions of software, but we're still funding companies as if the old rules are true. I think that's dangerous.

I wonder how many investors are just assuming that eventually the margins will end up looking like SaaS. Like eventually they'll flip to 70, 80%.

90% of people believe that.

Yeah.

I don't.

How many years will that take, if it happens at all? Say margins will never flip to be as fat as SaaS. Are there other KPIs and metrics that you should be leaning on more heavily to kind of show you?

I think this is the hardest thing for software investors. I deeply believe this and I think I'm right, but I'm willing to accept them wrong. I think software for most of my life has been a business where the producer of software captures all of the value. Yeah, 90% margins really means the customer has no leverage. That's what that tells me. Yeah. And of course they don't because software products took an army of people to build. They were really complicated, they were expensive to maintain. And if you built an ERP system, your customer wasn't going to do that on their own. And there weren't cheap alternatives unless you were doing some weird open source thing and support it yourself. So you could charge people 90% margins and get away with it because they didn't really have good alternatives. And it's not that the customers didn't get value. The value capture by the provider was aggressive.

Yeah.

If you think about the total opposite end of the spectrum is open source, where the economic value mostly goes to the person implementing the software, not the person who wrote it. The person who wrote it usually writes it for other reasons. I think the software business is going to look more like open Source, where a $10,000 piece of software in three years is going to probably be the equivalent of a million dollar piece of software five years ago. It's just deflationary. You're going to pay a lot less. Which to me means margins will go down because the consumer of software is going to capture much more of the value and that's good for society. If you. Everyone's like, software eats the world. I'm like, well, not at 90% margins. It's hard. Software's cheap. Software can go a lot of places that it doesn't go today and solve a lot of problems. So there will be more money made in software in the aggregate. But it might be a less attractive business from a margin standpoint. But I think we have to evidence like, but you're going to need fewer people, I'm like, but not none. And some of this is trading human workers for computer. So I actually think software is going to become a less attractive business but not an unattractive business, if that makes sense.

Well, we are all out of time today. Thank you, Charles, for stopping by. Thank you, our listeners, for jumping in. Please like and subscribe and we will talk with you all next week. Thank you.

Bye.