Navigating the SaaSpocalypse – The Inevitability Curve EP3

This is Textron tv. Hello everyone, and welcome to another episode of The Inevitability Curve. In this series, we talked to interesting folks about interesting topics that have evolved over time.

And with me today is an old friend, Roy Pereira, who is currently the CEO of Unified, API. How are you doing, Roy? Hey, I am doing well, Chris, As you should be.

As you should be. So, as we were talking about in the green room a, a minute ago, you know, you and I have had some shared experiences. We worked together quarter century ago in the, uh, security space at Cisco.

Since then, you spent a lot of time looking at AI and, uh, in, in, in the conversation leading up to this, I love the SaaS apocalypse, right? You might be coming to a crux in this software as a service for those of you who don't know what SaaS is, um, era. So, I'd like to, to take this conversation, let's start at the start.

Let's go back as far as, you know, either of us think we should. How did we get to where we are today? So I'll throw that to you.

How did we get to where we are today? Yeah, I mean, you mentioned 25 years ago. Uh, I, I try not to think that I'm that old, but 25 years ago, we, Hey, bonus rounds.

com, uh, was going crazy. People were, uh, getting massive valuations going public and so forth. I think you and I were both working at Cisco Systems, which was the darling of the time.

Um, but it wasn't about Cisco actually. It was really, I mean, Cisco was, was creating the conditions, uh, that we all needed for SaaS to come out. Salesforce was the company that really pushed it, uh, back in 1999, 2000.

They came out and said, Hey, just don't install software. Just go to the cloud. Just use ours.

Um, and it's a, it was a really interesting turning point, uh, because it changed the way that everyone today uses the internet, uses applications and is productive, basically. Um, it, we did not realize what that would allow us to do. You know, we were thinking about, oh my God, I don't have to ever upgrade software on someone's computer.

That was like the benefit, right? Obviously that ushered in, um, subscription pricing, which that wasn't a thing before, but it also ushered in connectivity between applications. All of a sudden now you are, uh, able to connect to other applications on the internet 'cause you have public.

Um, so this interconnectedness happened, um, and it, it massively grew the number of applications that we had, right? I if you think back to, you know, 1999, you know, you could count the number of applications in each category on a hand, right? That's not the case today.

Um, you know, you look at say, recruiting software, recruiting tech, I've lost count. It's like in the hundreds, multiple hundreds. Uh, and, you know, a lot of them do the same thing, but a lot of that is because of sas, Right?

In, uh, and in thinking about this topic, right? You know, so in, you know, I got in, you know, involved in this, in the early firewall market in about 92, more or less, right? And Bell Canada was one of the first VARs evaluated reseller for, uh, uh, the modern parlance.

And in selling some of firewall and teach him how to use it and getting it set up, you know, was, was trouble. So I think that was really the first conversations I had about the managed firewalls, firewalls as a service, right? And by the late nineties when you and I got together, it's starting to happen.

Um, now we've gone to this phase where, where, as you say, it's all out there and just, you know, all along the value to me, and I find myself in, in the security context, having these conversations with other security people, it's like most companies shouldn't have a security staff. They make shirts, you know, and, and it writ large. Like how many people, how many specialists do you need to have on board to make denim denim jackets?

At some point, this needs to be outsourced. So we get the cloud, we get SaaS, and I think that's led to issues that we sort of have to work through in the near term, right? For, for sure.

I, I, I think it's, uh, interesting that you bring up security. Um, both, both you and I obviously are, uh, were insecurity. I was in security.

You, I think there still are, but, um, I don't think people really think about security with new, um, scenarios. Like when SaaS came out, we didn't really think about security in the cloud. We started to think about it, obviously once we got into it.

Um, I think it's very similar to what we're in today, uh, which I know we'll talk about, which is ai, right? So you see a ton of companies moving into AI with, with little regard to security, whether it's privacy or leaking data. Um, just so much, um, issues that I don't think most companies have really thought about.

So security typically is still an afterthought. Well, and yeah, Don Alsup, my boss at Sea Change, you know, he, you here in, uh, Ontario in the early nineties, um, I talked him into, you know, going ahead with my crazy idea for a firewall. And then we ran into the problem that there weren't enough IP addresses, right?

8 or whatever you on didn't exist at the time. And this somebody, you know, put it in my head and I'm like, oh my God, you know, the internet's not gonna work. And John, John, John, you know this, you know this.

And he didn't blink an eye. Of course, he was old and wise like our age now, right? And he just sat there and he said, look, anytime there's enough of a demand, a technical solution is found.

And within 15 minutes, you know, uh, Andrew Flint and Oma Lundy, and I invented network address translation, as did two other groups totally independently at the same time. And now we all have net and we're still using IPV four, and we still haven't run out of addresses. And so many issues since then in the security and non-security space see the same sort of thing.

It's like, well, well, let's, let's pick a point in time. Cisco Secure network, right? You know, the turn of this entry 2000, you and I and the rest of the Cisco team built up an architect, uh, put together this, what what, you know, is reasonably called a market architecture, right?

We weren't necessarily selling all that, but we said, this is what it should look like, and it still should. But if you really want to do that with the scales we have these days, how are you going to have people do this by hand faster? And you can't, therefore, unless we're wrong about this, something else is gonna take its place.

So SaaS and Cloud, that was obviously right. Not everybody can afford to have their own teams and as economies and scale, um, securing that and to the point I'm, I know you'll be, be bringing up actually using it, like bringing effective with this stuff, um, requires us to change our models. You know, we, we're not doing things by hand.

We're using, I'll say it, artificial intelligence by the definition we tend to use these days, and we're gonna be automating and scaling stuff up because, you know, we, the, the old ways just don't continue to work. Uh, but the architectures were right. Yeah.

So the way that I look at it is the last 20, 25 years have been a boom time, uh, for software, uh, for all types of software. Not just enterprise, but consumer as well because of the cloud, right? The, the infrastructure, the plumbing that companies like Cisco set up in in the late nineties.

Um, you know, maybe we didn't see the benefit in 2000, 2001. com, um, right? To, to really, uh, see the benefit of all that, uh, that fiber that, uh, we were laying down.

Um, but, uh, over the last 20, 25 years, I think we've just sort of gotten used to that. This is the norm, and it's been really a boom time. It's, it's been like a gold rush.

Um, 'cause when you look at VCs, venture capitalists and, and what they fund, they know that they can consistently get returns on the average, uh, by investing into a SaaS company, a, a new software for SaaS that does whatever, X, Y, Z just because the market itself was growing at a phenomenal rate year after year after year for 25 years. Um, and, you know, there was a lot of truth to what, what, uh, mark at, at Salesforce said early about how software's eating the world, what what he meant was that software's eating the old world, the on-prem world, where again, you were installing software, uh, on-prem, um, in your closet kind of thing. But the problem is, right now, at least, uh, we've hit a, a wall.

There is no on-prem and there hasn't been in a couple years. Uh, and so that growth in SaaS, I think, has started to, to slow down. I think we've seen, um, people concerned with Salesforce's ability to grow, um, and, and their stock price.

Uh, but I, I think the bigger issue, and I think you started to allude to that, is now we have ai, right? And we had ai, we've had AI for forever actually. Um, we now have useful ai, whatever adjective you want to use this stuff, even though it doesn't think for itself, it's not a human whatever, whatever.

It's actually pretty damn useful. And it can actually do stuff now as opposed before, it was a, it was like a parlet trick that was kind of cute. Um, and that's where I think we're about to see a massive, massive change.

We're seeing it starting, uh, but it's going to, uh, it's gonna be a massive change in, in the next five to 10 years. I, I, I, I frankly, uh, uh, agree. I think this is, you know, I, in the last episode, uh, uh, uh, Tim Roxy, good friend was on, and he is, uh, if he's been involved with the na, the power grid, you know, for, you know, 40, 50 years, a nuclear, uh, engineer.

And we were talking about artificial intelligence, and you know, I, I love, you know, common memes are great, you know, we understand ai, you know, we can talk about it and so forth, but to be clear, it's not artificial, nor is it intelligent, nor is it really new. But the differences are important. We have, so the deterministic AI in the past, and that's been, you know, very useful in certain ways.

And now we have this emergent, um, um, AI that's allowing us to, you know, to do things we didn't think we could do before. Right? You know, if, if I had a thousand interns, I could do X, Y, Z, but since I'm not going to 'em, won't, won't consider doing it.

And I think that plays into your point, right? You know, to scale this, this SaaS apocalypse, um, doesn't mean that everything's gonna stop working. It means we're driving past a point and the old way of doing things isn't gonna work very much longer.

And as, as, as luck would have it, we do have the tools, I think, to do the things we need to do. Yeah, I think you brought up a really other interesting point. Uh, you know, as technologists, we always talk about technology, software, hardware, um, but it's the human side.

That's actually interesting because I actually, I'm concerned, and I was concerned 10 years ago too, so my, my, my last company started off as an AI assistant 10 years ago. Nobody wanted to pay us anything. Um, and which is kind of ironic, right?

'cause then now every other company is an AI assistance company, uh, and there's a ton of, uh, funding available for that. But, um, what we learned was, um, that, uh, you know, AI has the ability when done properly, when it's useful, uh, to not just affect software and technology, but really affect humans and the way that we could live really. Um, and I was concerned 10 years ago that didn't actually take shape.

Might take shape this time, because again, the ai, um, that we have today, the, the technologies is much, much better, uh, on magnitudes, right? So what you're getting at is, do I have a job anymore as a, or if, if I'm an intern, can I get an intern job? It's kind of hard.

Uh, if I'm a junior lawyer, can I get a job in a law firm? Hmm. You know, maybe today, uh, if you're really good, uh, in two years, maybe not, right?

And so, like, if you're a junior and you're, you're just stepping out into the workforce, can you get a junior level job? Because the AI basically is just as good, uh, and you know what doesn't take breaks, doesn't talk back, doesn't want, you know, um, doesn't want all sorts of stuff. Um, so I am a little concerned about that because that has a waterfall that's quite negative, uh, in terms of the way that we, we structure our civilizations.

In fact, like it's that fatalistic. But you know, you can see why some of the tech leaders in Silicon Valley are pushing for guaranteed basic income, right? Because they see the pitchforks and the riots happening.

If, you know, the vast majority of people don't have jobs because they're being taken away by, by useful ai. And, you know, which is an interesting twist on it. You know, I, I'm, I am, uh, you know, I have, I've said it on this in the show before, but you know, I belong to all three major American par uh, parties.

I get it, right? And UBI, universal basic income first crossed my bow maybe 10 years ago. And it sounds a little silly the more you think about it.

I like thinking about that as a national dividend. I assume, you know, you're Canadian, I'm a yank, you know, so, you know, take a country, everybody gets one share. You know, if the com country's profitable, you'll, you'll get more in your UBI or less, right?

I, I like the capitalist drivers. Um, but I also have three 20 something kids, right? And what happens over the long term, we can talk about in a minute, you know, might be deterministic.

I think there are certain things about the future getting past the, the details of the, of the immediate future, the next couple years. And these sort of cases will work out fine. However, that may not help you a whole lot.

Yeah. You know, during the Industrial Revolution, it did work out. Um, right.

So this could be one of those instances where as humans, we adapt, we, we survive and we're, we're gonna be forced to. Uh, and so we'll come up with different things. Uh, maybe it's gonna be like a Star Trek universe where, you know, nobody gets paid, uh, and everything is free.

Uh, but maybe not, uh, probably not. But, uh, I, anyways, we don't have to get into that, but I, I, I, because it's, you know, who knows? Uh, but I think it is an interesting after effect of this technology.

Well, I think, you know, I'd love to get into the AI assistant thing with you longer. We probably won't do it here, but Right. It is funny, you know, talking to you this morning, I have been sort of, that's been stuck in my head.

And when I say AI assistant, I mean that everybody gets one. Everybody, everybody. Right?

You know, what does it even mean? You know, not when it's expensive, but when it's basically free. You know, what would it mean if, if everyone you knew had a personal assistant that was paying attention to them, you know, think next level, past, you know, current ai, give it a generation or so, um, how much more time does that free up Now if we couple that with the same AI taking all their jobs, we have other problems, but the opportunity as, as, as individuals and as societies to leverage more of, uh, the population, you know, and call it workforce or not, you know, how, how we're measuring the productivity.

I see, you know, mid to long term, uh, future opportunities for societies to do phenomenally well. Right. I, I agree.

I I I think you said something that I, I think we need to be open to, which is like define work, right? Um, if you go back a hundred years, work was defined differently, right? Than what we are working today.

Like, I'm looking at you on a screen, right? Um, you know, is this work, uh, you know, I'm working at home right now on Friday. Does that work?

Um, you know, five years ago probably that wouldn't be considered work. And, and so I, I do think that we're in, uh, yeah, we're about to go through some bumpy rides where we're, uh, contemplating what work does, what do I do, uh, how do I benefit society, maybe, um, you know, in the tech of work. So I think that's interesting.

But do we actually execute the work? I think is also a really good question. 'cause more and more in, in my team, in my company, we don't execute, you know, the AI executes, the AI is, there's multiple, right?

And so we come up with stuff and in fact, we actually get AI to help us come up with some of that stuff. But the execution a lot of times is automated and it's done by an ai. And that's where I, I really do see us moving towards.

And that means that I need less people, right? Again, back to my issue about, uh, you know, we can probably be a hundred million dollars revenue company with 10 people, and we're not going to be an anomaly. I think that's gonna be the norm.

Um, so what does everyone else do? Uh, that was my initial question. And, um, yeah, but again, not to be too fatalistic.

'cause I, I'd rather be, uh, on, on the more positive side of things. Well, you know, I'm, I'm born in 1965, you know, so I'm basically the oldest, the Gen Xers, but I was, you know, the youngest of all my family in all directions skip years of school. So I was the youngest of my peer group.

And yeah, I, I think I, I think a lot of the, the, the, and I, you know, the current, uh, uh, cultural gestalt, whatever you wanna call it, has a lot of pessimism built into it. But after 59 years, you know, we have seemed to have, have survived all the apocalypses, you know, that have come towards me. The, the one of my first, my first memory I have of a talking head, you know, people like us, you know, on a screen somewhere, it is 1969, I'm four years old.

A guy named Paul lic had published the population bomb, you know, was on the every Boomer's bookshelf. 54 billion people. 52 billion people.

So a hundred million Americans will starve to death in the seventies. And you gotta remember, I'm four years old, right? And I'm just thinking, crap, that's gonna suck.

And we make, you know, to be clear, 30% more food with 88 million billion people than we did with three, right? So I think we, we find solutions to this. I agree.

I agree. I think the Gen XO will definitely find a solution. Well, I, I have a lot of, I have a lot of faith in the, the generations coming behind us.

I look at, let's talk about the future, you know, for a minute. 'cause I look at, you know, we, the Gen Xers and the Boomers and so forth, you know, the me generation of the sixties and seventies have, as we always do, you know, as humans do, gotten older and we're, we're like kids these days, uh, much like was said about us in our teens and twenties. Um, I think kids these days will work it out.

I think they will develop the tools and come up with the methodologies to keep making the role to place worth being, Right? Yeah. No, de definitely.

I, I, I, um, I think it's gonna be interesting. Like I said, I, I, I think we're, uh, just entering it now in terms of this massive upheaval. You know, if we go back to software and SaaS, um, you know, we, like I said, we've had pretty easy game plan for the last 20, 25 years.

And I think that's changing right now. 5 came out and everyone realized that it was actually useful. And now, you know, we're into four.

They just launched, uh, what was it, oh, one, uh, yesterday. Um, anyway, so it is just gonna keep getting better and better and better. And it's growing exponentially in terms of benefits.

So, uh, we definitely do see in our own customers, so we, we provide unified APIs to B2B SaaS companies, right? Infrastructure so that they can talk to other SaaS companies and get customer data. And so, uh, SaaS apps come to us, uh, for this infrastructure.

And we're seeing these trends, and we're at just the forefront. And it is just amazing because 12 months ago, or actually 15 months ago or so, we weren't seeing this ai, uh, shift. And now we are, uh, and in fact, it's almost a hundred percent coming in.

It's all ai, ai, ai, um, and it's no longer traditional SaaS applications that would fit neatly into these categories that we've always known for 20 plus years. So yes, there's a new category, uh, and it's not well defined. It's just like AI first, whatever that means.

Um, traditional categories, SaaS categories are adding AI to compete. Um, and then the fact that AI now can build, that's really interesting because I theorize that we are going to get inundated with everyone and their grandmother building a SaaS app, oh, and a B2B uh, B2C app. And you know, we have, I think, what, 35,000 SaaS apps globally, uh, according to our count.

I think that's gonna go to like 35 million maybe, right? Um, 'cause it's just so damn simple. Like anybody, you know, your 6-year-old can go on to Claude or maybe the new chat g PD four, and ask it to build an app.

Um, that's gonna be interesting because, you know, we already have too much, uh, as, as like an end user, we have too many options. Uh, and so it creates, um, it creates no winners. How about that, right?

And so there's too many options. And so every option gets, let's say, 10 customers and they can't grow anymore. Um, and I think that's what we're seeing right now.

Um, but I think that's gonna change. I I think this is short-lived. I think in the next, uh, 12 to 24 months, we're gonna see a massive crash of the number of SaaS apps as AI and AI agents take over, and they just kill off all of these single feature SaaS apps.

Um, and that's gonna be a pretty massive drop. Um, and we we're starting to see some of that. Now we're starting to see some of these larger enterprises who have maybe more resources in-house, uh, around ai, um, starting to like do an audit of all of the software that they pay for on a monthly basis and, and ask themselves why, why are we paying X dollars for these apps?

We could just build them, right? So either we build them with the use of AI or we get AI agents to do that for us. Um, and I think that's going to be the norm.

Um, and that's going to be the next, uh, thing that is going to reshape all of software on, I believe. Yeah, I'm just trying to think of the, the, the scale issues of this. And we'll take this as, you know, to, to end on.

'cause I think that, I think there is a, a, a, there's a similar pattern in this, right? You know, if we go forward and keep increasing x, you know, at what point is X larger than, you know, the mass of the planet, you know, you know, completely unmanageable. And if it's not going to decrease, um, we need to do something else, right?

You can't do it by hand. And the examples, uh, you put out there, we've gone from this, you know, look, I'm gonna buy software and install it to, I'm gonna buy, okay, I will rent software that's on, on the cloud somewhere, but I'll, you know, rent it monolithically, you know, somebody else has built this big thing. I'll sign into that.

Or I got my API, and you know, to be clear out there, this is not a sponsored piece. So, you know, not, not, uh, pushing your, your products, but as I understand it, you're simplifying that because there's, you know, unified API, you know, sounds to me like instead of having 7,000 ways to deal with all these different things, you know, how about one, right? And in a, in a similar, similarly, if AI can build the apps for us, um, then, then we're not gonna be, you know, choosing 'em all by hand.

We're gonna automate all that. Yeah. And, and, uh, you know, we, we can end on this one.

So we started talking about Salesforce and how they ushered in SaaS, right? 25 years ago. Um, we are now seeing larger companies remove Salesforce, remove workday, uh, stop paying these, these kind of large fees.

Um, and they're building these in inside. Now they, they may not need all the functionality. Um, but I think what's happening, and it will definitely be happening more and more, is the fact that AI is taking over the, the tasks that humans used to do in these large platforms, sources of truth like Salesforce, workday, and so forth.

Humans are no longer doing those roles, those, those little tasks. AI is doing it. And so the value of those large platforms is diminishing, and it's diminishing so much that they basically become a database and a very expensive database.

And so you don't need to pay that amount for a database. And so we're starting to see companies now replace them with a simple database wrapped around with ai. And I think this is where the AI infrastructure companies, um, you know, obviously the open AI is the anthropics, the coheres.

They're, this is where they're next going to. They're not like just building better models for ai. They're now going to explain, uh, expand into infrastructure that wraps around everything that an enterprise, uh, company needs, and that then we'll eat into the original, uh, SaaS, um, large companies.

So I, I, I think it's full circle that's about to happen. Well, the circle of life, as they say. That's true.

But, uh, yeah. Thanks, thanks for all that. You know, a lot to think about.

I, I, again, I, I worry about the future. There's lots of things to be concerned with. I am hopeful that we'll keep finding things to do.

And, uh, it seems clear that economies, you know, uh, if you will do just fine. It's how we, it, it's how we, we weave through all the details is kind of tell the text for the future. So, at any rate, Yeah, I, I just hope that it gives us humans more time to do what humans like to do instead of some, some of the work that we've had to do just because it wasn't automatable.

And now that it is, maybe we can actually be human, be more human, Maybe. Well, I, I'm a Walt Disney kid. That's a good way to end on it.

Yeah. I, I grew up with the world of the future and I still believe in it. So let's, uh, let's take that as our hopeful note.

Well, Roy, thank you very much for your time today, for everything you've done for the industry and the world, and for being yet another good human. Thanks, Chris. Thanks folks.

See you next time.