Chamath on why young people need more agency, risk, and adventure

23:25Jason Calacanis:And then you can imagine Imagine if instead of saying that, instead you said, I want to build an airplane company. Maybe it's not 10 agents, maybe it's 10 ,000. But the point is that there can be a conductor, an orchestrator at the absolute top level who can help you execute on all of those things. That's where I originally started in the spec that I wrote. The problem was that the market was not ready for something like that. And my best guess when I read it over and over again is that's a 30-year or 40-year journey. So then I had to wind it all the way back and say, what is a logical starting point today that over time can earn our way into being that omnipotent co-founder for every human on earth?

24:11Jason Calacanis:And maybe just to double click on that, why that? Because I think for me, this idea of enabling economic mobility, giving every single human on earth the ability to be economically independent, self-sufficient, take care of themselves, take care of their family, have money to spend on the things that they want. That to me is a very powerful idea. It's an extremely decentralized view of capitalism and democracy at scale. Yes. So just seem like, you know, I had a lot of passion. Big audacious mission, right? Yeah. And if you look at what's wrong with America today in our great journey as a capitalist society is self-reliance has been replaced in some ways with, you know, victimization and handouts have replaced pulling you up by your bootstraps and agency.

25:01And what you're talking about is, you know, basically reconnecting people with that. What the pioneer spirit of this country and as an immigrant to the country and somebody who had to fight for tooth and nail for everything you got, which, you know, people take for granted when they meet Chamath circa 2020s. You know, I know the story. You have to fight every inch. So if you have to fight for everything, you really want to see other people have that opportunity. It's the clearest path to fixing this K-shaped recovery. The issues we have that we talk about on All In every week is just to empower people to start their own companies.

25:33Yeah.

25:34Jason Calacanis:Then I had to kind of say, okay, if that's where I want to be, where do we start? And, you know, I learned two really incredible facts. The first is that if you look at the world's GDP, call it, for the purposes of this conversation, $130 trillion, I would say 90 % of it requires some sort of technology to make it come to life, some sort of software of some kind. There's a certain percentage of the GDP that does not require that. Yeah. Increasingly smaller and smaller. Yeah. Increasingly smaller and smaller. So it's kind of what Mark Andreessen said, software really is eating the world. So that was one interesting factoid.

26:16Jason Calacanis:And the second is when you look at software, what I learned was software can be bucketed into two categories. It costs around$5 trillion a year. So think of that as$5 trillion enables, call it$100 trillion. Okay, so there's a 5 % take rate that software extracts to enable most of the GDP of the world. And that 5 trillion can be further subcategorized into a trillion dollars of licensing revenue. So this is what you would pay Workday or ServiceNow or Jira or Linear, GitHub, whatever, right? A trillion dollars of those licenses, and then$4 trillion of maintenance, migration, and services. So meaning once you implement Workday, you typically bring in a suite of consultants to just help you make sure it doesn't break all the time.

27:15Yeah. And that you get your value from it because somebody has got to make sure the distance between this huge contract that was signed by management, some CTO somewhere actually gets implemented and it's worth the money you spent on it or else somebody loses their job eventually. It might take five years, but I remember when I was in IT, this was an acute thing. Some top-down person was like, hey, we're going with Lotus Notes. And then it was up to us to make sure Lotus Notes actually delivered the value, not Lotus Corporation.

27:43Jason Calacanis:But it occurred to me when I looked at that, because I had not really looked at that very carefully, I had this light bulb moment, which is I contrasted that to what we had done and a lot of what my team had built at Facebook. And I thought to myself, well, hold on a second. Why is it that Facebook, which is an exceptional business, only uses an extremely small amount of that$5 trillion? And most of it was built internally. And then I remember what I had learned from my friends at Google, and it was the same thing inside of Google. And then I remember what Elon had taught me about Tesla, same thing at Tesla.

28:23Jason Calacanis:You had all these examples where at the extreme end of success, all the best companies almost had an allergic reaction to that traditional software stack. They refused to use it. And so the light bulb moment for me was, oh, I understand. Like, imagine, Jason, you and I both want to start a restaurant and we're given ingredients. We both get sesame seed buns. We both get three burger patties. We both get tartar sauce. We both get the same pickles. At some point, you're going to start to wonder to yourself, how do I differentiate myself from Chamath's burger company? Yeah, source better meat. And you're going to say, I'm going to source my own ingredients.

29:07Jason Calacanis:I'm going to make these things myself. And I'm going to take the risk because I think there's better margins and better success in doing it my way. But for most companies, even though they say that on the surface, underneath, with respect to the software that they use to run themselves, it is exactly the same. That is why companies like SAP and Oracle can be as big as they are. They've structured and standardized this idea that at a certain level of scale, the only solution is their solution. But I had seen firsthand and I had enough friends and examples of other great companies that took a different route.

29:50Now, normally you would say, okay, but that means nothing.

29:54Jason Calacanis:Nobody's going to, you know, how do you go to a defense contractor and all of a sudden convince them to make custom software? They can't do that. No. They've got other things on their plate. Yeah, I would have agreed with you until AI. Right. And the reason is you saw, I saw this cost curve where the unit cost of production was going to approach zero. So there was no task that was either too menial, too manual, too repetitive, or too difficult. So both ends of the spectrum that couldn't be accomplished at some point by these models based on how fast they were accelerating their learning. Right.

30:32That was what you saw just around the corner was, hey, this vibe coding stuff, which people are dismissing. hey, this is bullshit. It's just stupid. It actually was, if you tracked it and you used it every month, you could see every two or three months that it made some significant leap forward.

30:49Jason Calacanis:My thought was that vibe coding was going to be a great way to realize what the art of the possible was, but end up with a bunch of trash. Like prototypes, basically. At best, but probably not even that. Proof of concept, yeah. And so in the first instantiation of 80-90, I said, we have two goals. Goal number one is let's help companies get the margin benefits and the upside of the kinds of custom software that companies like Facebook and Tesla and Google benefit from, number one. And number two, along the way, how do you create a mechanism of learning the kind of data that can be collected by that process, by the unbundling of that process, so that incremental runs of software become safer and faster and simpler to build.

31:42Jason Calacanis:So first is do a job. Second is try to orient and create a network effect and a flywheel inside of doing the job. Okay, so you've identified a real problem and you put together a solid solution and a business model that you believe in. So you're all set to launch your new company, right? Not so fast. If you want investors and potential customers to take your new business seriously, you need to consider forming a Delaware C Corp. And that's where Northwest Registered Agent comes in. They're going to give your new company a real identity. That means an address for your public filings, a domain, a custom website, a business email, and of course, a phone number.

32:19And that's going to take just 10 minutes and 10 clicks. They don't charge hidden fees. Customer service is available around the clock. They're not overwhelming your inbox with spam, and they make it easy to cancel at any time. So get all the advantages of a Delaware C Corp independent, regardless of where in the US you're operating from. Visit NorthwestRegisteredAgent.com slash twist for more details. And the links are in the show notes. What we did was we created a product and we called it Software Factory because we felt

32:51Jason Calacanis:like that was a very honest representation of what this V1 goal is. It's just to create a factory that makes software. And we thought it could give value in both of those ways to people. So let me explain. So software factory is exactly what you think of if you went to a Tesla gigafactory. Raw material comes in the front, finished, Teslas come out the back. And in between, there are all these very logical stations and different tasks are done at these stations. In our case, what comes in at the front is just pure raw intention. And that can be for the kinds of business people and senior leadership of a company that deals in high-level business requirements, McKinsey decks.

33:45Whatever your business is, whatever your business goals are.

33:47Jason Calacanis:regulations, new laws get passed, maybe say it in a different way, a consent decree that you get from the US government, whatever. The point is that at the senior level, you're managing risk and you're allocating capital towards ideas. You want to give a shape to that. And then you want to be able to put that in the front part of the factory and have your colleagues and your teammates and the people that work for you extract from it a pretty detailed PRD, a product requirements doc. and what's great is it's not just the humans that collaborate on that. Now you can have agents working behind the scenes to make that PRD incredibly high fidelity.

34:28Jason Calacanis:And those PRDs at AOL was where you learned it. Yeah. Like they were big on these PRD docs. They were huge. By the way, you're bringing up a very good point. Like when I talk about the software factory, what it is, is an embodiment of what people call the product development life cycle or the software development lifecycle. They're interchangeable terms. The reality is that PDLC and SDLC lost its way. It gave way to a lot of what we created at Facebook in terms of an engineering-driven dev culture, which is move fast, break things, build first, poor documentation. And when we did that, we did that in a very specific moment in time, not because we cared about those words, but we were the 35th social network on the scene.

35:16Yeah, people forget. Yeah, there were like many before, MySpace, et cetera.

35:20Jason Calacanis:We had all these giants above us, including Friendster, including MySpace. We didn't have money. And so we had to move fast. Right, that was your advantage. But now engineering organizations start up with this idiotic idea that you should document, you should just wing it, you should just grip and rip this stuff. And don't write anything down. And yeah, that one happens. You get a lot of crap that gets into prod. You have an enormous amount of tribal knowledge. It's unfortunately trapped in many humans' heads. It's not well-documented. You have no SOPs. So when things break, things get really bad.

35:54Jason Calacanis:Why not have a system where intent comes in the front? You can shape it, humans, AI, agents. And then when you lock that down, Jason, the next most important step is to extract from it a very detailed, what we call an engineering plan or a blueprint. Like when you sketch out a house, the next step is you give it to the architect and say, give me the buildable blueprints, which give you all of the little technical details so that the house can stand on its own. The foundation is strong, all the wiring makes sense. And so our module, that step in our software factory, allows you to do everything from write, you can upload you know, mermaid diagrams.

36:39Jason Calacanis:You can do anything that you need to represent the system. And then when you lock that, only then can you do what most people go right to doing right now. Now, look, you can skip these steps, but we don't recommend it. The third step is you extract work orders. And this is where now, you know, vibe coding failed. if you could direct an agent to code with relatively strict guardrails, the agents are excellent. You've experienced this in some of the projects that you're doing. But the reason, I suspect, is because you've specced it cleanly and precisely. Yeah. And the more you give it, the more precise it becomes.

37:20It's a very counterintuitive thing. You would think that a stream of consciousness and rambling and throwing everything at it would create chaos these machines are very good at taking a large set of feedback and actually making sense of it and making a plan out of it it's kind of actually its strongest um its strongest ability right now and so this is why i have the foot pedal because i just when i do um a problem now i hit the foot pedal and i just keep talking and talking and talking and talking let it go and that it makes sense of everything I said. And then it asks follow-up questions because I asked at the end, ask me any follow-up questions, but what's not clear or things I'm missing and blind spots I might have.

38:05When you do that pedal to the metal strategy, boom, everything changes, which I think is your refinery thing, which is, hey, we're just going to pull in all of your stuff. What are your rags?

38:14Jason Calacanis:Again, exactly. Right at the front, our thing is you just dump everything, Zoom meetings, whatever. We'll sort through it for you and we'll give you a spec of a PRD. and now you can be there collaboratively in multiplayer mode in one doc, debating, questioning, asking, approving, denying. And then you go to an engineering plan and then you go to work orders. This is the same as a GitHub issue. And then you can send it to a coding agent and we support all of them via MCP. So you can use cognition, you could use codex, you can use clod. Cursor, whatever. You can use cursor, you can use whatever you want.

38:51Jason Calacanis:And by the way, the reason we support those is our view is all of those vertical tools are going to converge. That's not where the skill is. It's all going to reach parity is what you're saying. Yeah, they're all going to be the same. Plus or minus, whatever. Real skill is being able to keep this whole system in sync, right? Because then when you make the code, how do you generate evals? How do you generate unit tests? How do you wrap those together? And then how do you get it into prod and then maintain it as a system? So for example, you build a product. I'll give you an example of one of our customers, regulated healthcare company.

39:27Jason Calacanis:And with the BBB, a law changed. They just feed that into refinery. We understand the diff between that PRD and what needs to happen to respect the law. And once that's approved, it gets auto-propagated into the engineering plan, auto-propagated into a set of work orders. Humans review and approve them. and then put into the code because an agent goes and writes it. And now this system is completely in sync. The biggest problem that companies have is the drift that happens, right? Everything gets out of sync. By the way, it also works in reverse. So we built the factory so that it can work in forward or reverse.

40:10Jason Calacanis:So for example, an engineer gets a pager duty, wakes up at 3 a.m. in the morning, they all of a sudden have to fix a bug or patch something and they put production code in. We detect that. And now we auto propagate the work order. We auto make the change in the engineering plan and then we make the change in the PRD. So now everything is linked together. We call that binding. Everything is binded together so that everything is in sync. Now, Now, that may seem like overkill for a small product, but when you go to the big guys and you talk about what they're dealing with, they're not here to fuck around with some stupid vibe coding tool.

40:52Jason Calacanis:Like this is the level of governance and auditability that they need. Why? Because if you think about the kinds of products that many of our customers make, they're in highly regulated markets. They have huge upside for doing it, but they have huge risk as well, Jason. Yes, you have to mitigate, like, as a finance, health care, education, any of these verticals. Pharma. Pharma. You make a mistake. Defense. The blast radius can be large. Yeah. And can take years. Look at, I mean, your time at Facebook. I don't think you were there for all of it. But they've been under three or four consent degrees for mistakes made, speed errors, whether it's fair or not, whether it's political or not.

41:33put it all aside, it all speaks to, you know, having a tight plan, having this harness, people are calling it, having the software factory allows you to trust the AI more. And it allows you to feel less chaotic. We call it a control plane. Control plane, yeah.

41:49Jason Calacanis:The reason we call it a control plane is like, it's a plane that sits, you know, above all the fray, all the chaos, you know, models will come, models will go. We work with all of them. we allow ourselves to be able to take advantage of the best in class model for the best in task for the best task at that time. But it mostly is a product that's about multi-user collaboration, control, governance. And then what I told you is this emergent property where whenever you're building these products, we sit it on top of a knowledge graph. And what that is, is this is very similar to sort of the architecture that I had in mind at Facebook and the architecture when we led the Series A in Slack.

42:38Jason Calacanis:You know, I wrote this memo for Stuart, which was called Intercompany Edge Effects. You know, like, I don't think I'm like head and shoulders better than anybody at most things, which I'll be told. But there is one thing where I've just been in it for a lot longer than most people. And I see these patterns. And that is around network effects. And so this third time around, I wrote the PRD for this network effects and how to build it and how to architect it. At Slack specifically? No, no, here at 8090, yeah. But I'm saying it's built on the experience of Slack and it's built on the experience of Facebook.

43:14Jason Calacanis:And what I see is this ability to use the nature of these models to help improve the N plus first piece of software based on its experience building the end pieces before it. I don't want to give too much of that away, but that's working too. So basically what we have, Jason, I guess in a nutshell is we've created a factory. It gives companies a lot of control. It allows them to document things. It allows them to take out of tribal knowledge what is in humans' heads and document it for the first time in a long time. It allows these systems to stay complete and in sync. It allows them to unbundle a lot of that$5 trillion they're spending and slowly but surely kind of, you know, how do you eat the elephant one bite at a time?

44:04Jason Calacanis:It's like, OK, I'm going to take a shot at this specific workload. Oh, great. So, you know, for example, like today, there was a guy that tweeted, you know, that this is a third party. And he said, we've, you know, used Software Factory now to unbundle$5 billion of ISV licenses. $5 billion. And I saw that and I was like, man, this product works. This is great. Yeah. I mean, you've got the signs of early product market fit, which means you'll get to pull eventually where people will just be like, I need this. And what people don't realize is in those large companies, there's a long list of projects they would like to work on.

44:44They have two choices generally. They can hire IBM. They can hire a software company. Then they have the implementers they can hire alongside them, McKinsey, Ernst & Young, whoever it is. And now you're giving them a third way, which is, hey, are you smart enough to use these new tools to control your destiny and then take those middlemen? We're talking about the middlemen in wine. And those folks, like, are they actually adding value or are they damaging the core brand? And you could argue, and, you know, for every time one of these implementations happens and we saw it off close and personal, it was a waste of money and didn't get used and get It's thrown away or it's just years of frustration and you lose anyway.

45:23Here, at least you're getting closer to the craft. Yeah.

45:26Jason Calacanis:Well, look, you know, here's what I'll say about that. I mean, I think that there are some phenomenal organizations that I think have the potential to really thrive. I think the big consulting firms. So I'll take two specifically that I'm relatively close to, Ernst & Young and then Deloitte as an example. I think that these guys can thrive in the world of AI. Then you look at companies like, you know, PwC or Accenture, and it's like, it's like very critical to me, but it's like, they picked a model provider. And it's like, how can you pick one of Anthropic and OpenAI? Not because they're not good.

46:00Jason Calacanis:These models are incredible, but two reasons. One is you're then tying yourself to only one technological roadmap and you see the leapfrogging that's happening. You have no optionality. Yeah. And then two is, I think Anthropic and OpenAI have realized what we've realized, which is the end boss of tokens. So if you want to generate ROI, you must be able to sell ROI, meaning return on investment, on token spend to corporate Fortune 2000, Fortune 5000 CEOs, CEOs who can see a line of sight to making more money than they're spending. The thing that we all can see is that they are the most important customer.

46:47Jason Calacanis:Why do you think Anthropic and OpenAI have started consulting JVs? So for Ernst and Young and Deloitte, what I've told them is like, wow, you guys have made the best decision. Now, obviously I'm biased because they picked us, but when you have a control plane, you can now be model agnostic. You can solve your client's problems. You can actually work together. We work together really closely with them. They're incredible partners. It's incredible because now we get the benefit of decades of relationships, right? And trust that they've built and independence and auditability. And now we can work together to actually like solve people's problems.

47:23Jason Calacanis:So step one for like the next few years, we're going to be doing a lot of this, which is selling these big transformations into large corporate enterprises, and then doing the methodical work of implementing it and making it successful. I think if we can do that, Jason, in three or four years from now, we move into phase two, which is how do you take software factory and then slowly submerge it under the waterline? And that's the setup for phase three, which is how then do you have this just little, maybe it's a voice pedal called co-founder that exists. But what will co-founder sit on top of?

48:01Co-founder will sit on top of an extremely robust

48:07Jason Calacanis:software factory, right? Ultimately. And the understanding of how to get work done and how do you tie it to actual GDP and outcomes. And if you think about it in terms of infrastructure, we turn a faucet on, water comes out. We don't even see all of that underpinning and we take it for granted, right? And we're kind of moving towards that with the solutions you're building and intelligence as a service in just in general. But there's a lot of things that happen before that water comes out of your spigot. And what we saw with Uber and some other folks were you give people unlimited water, they're just going to let it run.

48:43Man, they're just going to be like a golf course. It'd be like making almonds. In the bathtub, because there's a valve

48:50Jason Calacanis:where the water just runs out. Yeah, let it run. Who cares? Like, I forgot to put the bath on. And this is what happened with token maxing, I think. And also, it's got a slot machine type addiction to it, which is you give it a job, it comes back. Oh, the job's almost correct. And you're like, oh, I almost hit my flush. You know, if you're playing cards, let me see another flop. Let me see another flop. Give me two flops. And you get this incredible rush. But more often than not, it's 60 % or 70 % of the way there. So you've got to pull the slot machine again, which means you've got to put another$100 bill in for tokens, right?

49:25It's pretty funny. Big announcement today. you've done your, I believe, is it the series A and the other one's considered seed or series B?

49:33Jason Calacanis:Series A. I mean, you technically are the seed, although we're calling you the preferred A1. Okay, I'll take it. But yeah, you guys, so yeah, we raised$20 million when we started two years ago, literally almost two years ago to the day. And it was great. It was all the besties, you and fun uh freeberg and sax uh and then some incredible angels um nikesh aurora is the ceo of palo chairman and ceo of palo alto networks adam d 'angelo who i worked with he was a cc and founder of quora on the board he is podcast shy i tried to get on this podcast for a decade it's he's just so podcast shy he just loves his work huh what's adam like he's great um he seems very smart and then you know like a bunch of our friends from the from the group chat like andrew David Lee, Skye Dayton, Diego Burdick.

50:23Jason Calacanis:And it was, so it's, it's wonderful. Isn't it nice to pass the hat with your friends for that first round? It just, it's a really great feeling. And we've seen it now with, um, uh, our friend, Sonny Mandra passed the hat twice, uh, for us. We had, you know, two nice outcomes there. It's just a nice feeling if you can, uh, get there with your friends. So we raised, we raised 20 million in the seed two years ago. And then, um, yeah, I'm very excited. Uh, we, uh, raised about 100 in the A, which is great. And we had an incredible lead, Mark Benioff in Salesforce Ventures. Amazing. And then we had an incredible new group of folks also join, Thomas LaFonti-Cotu, Yuri Milner, Xander Lurie.

51:15Jason Calacanis:Oh, wow. Just great people. Great humans. Now, this is the largest you've raised for a startup, obviously. What's it like? What are you learning as a CEO versus a capital allocator? Because I went to see the office and sometimes I call you like, hey, let's get lunch or can you come to this event or do something? You're like, I have a sales call, which by the way, I will tell everybody that's I do it's I sell the enterprise software now. That's what I do. You're an entrepreneur. You're doing the founder led sales. No. Which is always the best way to do it because then there's nothing between you and the truth.

51:50And you can then have this clarity of vision. That's why founder-led sales for the first couple of years is so important. Because your sales team will massage you, massage the client. The truth gets kind of obscured 20 % on either side. And now you're living in a house of mirrors. Like your sales team is making you feel 12 feet tall. They're doing the same thing to the client. And like reality is just distorted. Um, but, but talk a little bit about what you're learning being a CEO versus a capital allocator or, or, you know, being somebody's top lieutenant.

52:21Jason Calacanis:I think it's the same job. Meaning the difference between being an investor is I was only investing one of those five units, but being a founder CEO, you are allocating all five of those units. Um, now, you know, I, we benefit, I benefit definitely from having, um, you know, distribution that most founders don't start with, right? You know, large Twitter following, obviously being able to do the podcast with you, that definitely has helped me and it's given me an advantage. It gets you to, what I tell people is, it gets you the meeting. It doesn't close the deal, right? It doesn't close the deal.

52:57It will get you the meeting. People ask you for a selfie in the lobby, but you're not closing the LP unless your returns are there. You're not closing the client unless the value's there. So you can't overestimate it, yeah. You know, where do I spend my time now?

53:10Jason Calacanis:it's like, you know, it's, it's this, it's this constant state of worry, to be honest, like it's, um, existential dread. I don't want to let my friends down. I don't want to let my investors down. Um, I don't want to let my employees down. Um, in terms of the actual, you know, so that, that creates this kind of like omnipresent anxiety that I think I'm actually pretty good at absorbing. In many ways, Jason, I grew up in a pretty complicated household. I've been pretty open about that. That was incredible training for this moment specifically. My dad was an alcoholic. My dad didn't really think twice at times just to beat the shit out of me, which it was what it was.

53:58Jason Calacanis:I have no grudge, but it's really made me able to absorb chaos, I guess. I mean, if you look at all the great founders, you're going to find this like dysfunctional childhood on the margins or at the core of it. And, you know, it's the pressure that makes the diamond. Right. And then you get into a business situation and you're like, this is nothing compared to getting the shit beat out of me or living in terror or, you know, whatever. I just want to do right by the people that have taken, that put their faith in me. And so where do I spend my time maybe? So, you know, I spend a bunch of time obviously kind of like in the machinery of getting more business because I think that that's very crucial.

54:40Jason Calacanis:I spend a bunch of my time as a product manager of our network effects business. So, you know, there's the principal engineer and I, we sit side by side and we're like little buddies pairing on this thing. I spend a bunch of time there and I've been spending a lot of time right now in the organizational design of 80-90. So meaning we don't have an org chart and we don't have a hierarchy. And so you may say, well, how does a company, and look, like last year, we did 17 and a half million bookings this year if we execute maybe we get you know we should have a decent chance getting to 100 next year i've already told them what the plan is which is 500 i have no idea whether they're going to get there or not but my point is i wanted to take all the all the drama off the table um and we're making a lot of progress but the default jason was to create an org chart people would ask to hire people then it's like who should report to whom yeah then you You get Game of Thrones starting, fiefdoms.

55:46Jason Calacanis:It's not the way to succeed. This is not how it should work in a world of AI. And so instead, I came with a framework and I call it system on a chip. And so what does that mean? When you look at the iPhone, the iPhone's an incredibly beautiful product. It clearly works. It's incredibly successful. Some would say the most successful product of all time in terms of profits and distribution. Yeah. Let's take two examples inside the iPhone. Let's take the camera chip and the power management chip. What's so interesting to me is if you break open the iPhone, what you see is a circuit board. And what you see are a bunch of chips with a bunch of interconnects.

56:24Jason Calacanis:So in the case of power management and the camera, the camera has no idea what the power management function does, except that it gets a signal. That signal is very specific. You know exactly what it is, and it can act on it. And I tried to take that analogy and I said, you know, why aren't we organized like a circuit board? Why aren't we a set of chips and a set of interconnects? And what that does is it reduces each function into understanding what are the inputs and outputs of that function. I'll give you one example. So I went to marketing and I said, okay, don't ask me who reports to who, none of this stuff.

57:00Jason Calacanis:You have two inputs. One is money. And the second is all the content that you may need. And the only output that you can generate are leads. Now, those leads can be enterprise leads. They could be leads for people who want to buy subscriptions to Software Factory. The leads could be job applicants. But your chip takes in these two inputs, generates one output. Now, draw your internals for me, and let's go debate the internals. That's where you start to get into tactics and strategies and how the, and loops, yes. Processes. And so what you start to see are like people designing cores, right? To use chip architecture language, cores and chiplets.

57:40Jason Calacanis:And we debate the interconnects. And then, you know, the leads now feed into sales, right? And then I go to sales and I'm like, all right, let's design your chip. Sales to demo. And then I go to finance. You know, they get leads and they get, you know. And a customer client comes out the end. No, just TCV is only allowed to come out the end. That's all they're allowed to exude. Their exhaust is TCV. And so then TCV goes into the 8090 Enterprise team to then build the stuff. It goes into finance, to account. So my point is, we've been playing this game, essentially, of reimagining the company as a circuit board and reimagining everything as a chip.

58:23Jason Calacanis:and the chip is comprised of these cores and chiplets and everything has interconnects. And now you get to like a place, Jason, in a world of AI where agents can sit at these boundaries, measure everything. And I'm not saying that we've eliminated politics at all, but I'm saying it could be a path to it. So I've spent a lot of time trying to structure the mechanism of our company so that we can scale without the typical failure modes. Yeah, I mean, I think your customers are also going to go through this kind of change. We're seeing it when we debate, oh, these job losses, these layoffs, is it AI window dressing or is it actual reality?

58:59And then we debate it in a very granular way. It's obviously both, which makes the debate great. But the interesting thing about this moment in time, those middle managers who were responsible for massaging this chip and getting things in and out and then what happens between the chips, there were humans there. And those humans would bring their biases. This is a great example.

59:21Jason Calacanis:This is a great example. So let's just say that marketing generates leads. Well, what happens at scale is they'll generate a lead and sales will say the lead sucks. Yeah, the lead is weak. And instead, what should happen is at the boundary of those chips, there should be a DSP, right? Something that qualifies a signal as high or low, or it scores it on a spectrum. And now we know what the leads are. We know how they convert. And to your point, there's nobody debating and arguing. It's not about personalities and emotions. It's about what does the numbers say. A much simpler process. And, you know, it's so great to hear you doing this because I'm also learning, learning with you.

1:00:01And did a similar process where I took all the applications we have and we started scoring the applications for year zero, year one startups. And then we just have this great Socratic debate. The humans are in the loop. Hey, what should we weight the scores? And somebody had put into the weighting that the founders were coachable. And I was like, wait a second, how did that get in here? And they're like, well, you know, like we have these problems with these people who are hard to deal with as founders. And then we have these ones who are coached by the way, all of our money's made with the uncoachable ones.

1:00:31So you just put into the system something that takes the diamonds and throws them in the refuse pile because they're a little sharp. They're supposed to be sharp. They're fucking diamonds, folks. Go back into the garbage and find me the diamonds you threw away. I want the uncoachable, difficult people. Those are the ones we make the money off of. It is one of the great aspects of systems thinking. And I kind of feel like the future is people who can think in systems, whether it's first principles, second order effects, and just understand the, I guess some people call it the idea maze in business thinking.

1:01:07But the more you can understand the entire playing field, each of the chips, the more you can make that incredible brand product, right?

1:01:15Jason Calacanis:I want to give credit to two people. Yeah. This system on a chip org model for me was the output of reading Jack Dorsey's memo. And to be honest, I couldn't understand it. And I thought I was dumb. And I was like, you know, Jack was talking about building a world model and a sales model. And I really couldn't get to the level of detail to understand what he meant and what he was doing. And so I was just kind of churning in my head, churning in my head, churning in my head. And I had an important meeting internally where I was talking to my co-founder, Sina, and it kind of out popped this idea.

1:01:50Jason Calacanis:I was like, Sina, I just want a bunch of socks. I want to just be able to see IO. And then I want to debate the pinout and the interconnects. And then he and I kind of sketched it together. And the second was Elon. I don't know if you remember, Jason, when Elon opened Tesla's first gigafactory, when we went to that party. In Nevada, we went there, yeah. Nevada. Yeah. He showed me the layout of the factory and I said, oh, what is this? Is this a chip? And he was like, no, dummy. This is the machine that makes the machine. This is the factory. And that has stuck with me to this day. And the reason is because to your point about abstracting one level up, it's not about making the machine.

1:02:29Jason Calacanis:It's about designing the system that then enables the machine to be made. So, you know, when I'm sitting there designing the org, I'm trying to think through, like, what is the repeatable system that makes it like an iPhone and less like a brittle org chart, right? Or when we're designing software factory, what is the network effect that allows this thing to just spin faster and faster every single piece of software that we see? And I think if we get these things right, we have the chance to build something quite interesting. Yeah. And it's great. You're in the arena trying things, running a company.

1:03:02All of those things. All of those things.

1:03:04Jason Calacanis:The haters are going to be so mad, Jason. I mean, you're going to pull it off, too. And they're going to be like, oh, he can actually build a company, too, in addition to a podcast. I took a couple pictures when I was there because I got to visit. Yeah, yeah. Here's the first picture. This is you working in the factory. Oh, no, this is you on the sales. There's Chamath. And there you are. Look at that. Wow, you did our day. There I am in the factory. Yes. Getting dirty. You're rolling up your sleeves, getting dirty, sweating. So you really can do both. You can do both. We have photographic evidence.

1:03:37You're able to hit the road in that. Let me see that suit again you're wearing. Oof, this is the JCPenney suit. Yeah. Rugged. Look at that with the blue. Burn that suit. I mean, listen, you got to go knock on some doors. You know, the young folks love to hear from you. Love to hear from other. And there you are in the drink with me. And that one looks actually, the way these Chinese, these are Chinese models, by the way, that we've been playing with internally. oh there you didn't like that one oh that one that that's a that bottle went bad this is what the idiots who work for me and i do we're like producing a podcast that they just make images to try to get me to break up on hair so they're showing me these as we don't we're we're image maxing we're image maxing we're meme maxing it's it's pretty funny but um let's um look back um and young people coming out of school now, whether they've, let's put aside like how in debt they are, we all agree you shouldn't go in debt.

1:04:33It's kind of a settled issue. We understand the future is AI. So for our kids, we have kids similar ages, we have teenagers and some younger kids, you know, over the next decade or two, they're going to be going through college, they're going to be into the workforce. What are you raising them? How are you raising them in terms of skills, hard skills, soft skills, whatever in between, but how are you thinking about getting them ready to take on the world and be productive in it? And then what's your advice to people who are in college? You got a lot of you, Waterloo alumni. What are you thinking now?

1:05:08And just dream of consciousness because this is a dynamic field here.

1:05:11Jason Calacanis:I don't have a good answer. I have not tried to answer this for my kids because I feel like I asked this question more than I answer it. And I don't really know. And the only thing that I've concluded is that I think it's very important for my kids to have their own adventure. I want them to, you know, do something interesting and have some wins, have some losses. Um, but beyond that, I don't really have much of a, I don't, I don't have a plan. Um, I just hope that they find an adventure. And if I can be a part of helping them find that, that I think is going to be the whole challenge. Cause like, and why do I say adventure?

1:05:52Jason Calacanis:Because I think it's kind of like all encapsulating. Like if you think that the food, the shelter, you know, your basic necessities are essentially covered for you in a, in this next iteration where most things are abundant. Yeah. Modern society, uh, society, society evolves. Abundance. Yeah. It's still not going to change millions of years of hardwired physiology. So how your brain has evolved, that's not going to change in one generation. No. And so I think if you understand that humans desire for agency and for risk. Problem solving, socialization. There's some core things that are just inherent in who we are.

1:06:32We're going to want to play games. We're going to solve problems. We're going to want to socialize. And if you don't do those things, by the way, you're going to all of a sudden have people telling you you've got to put your kids on SSRIs or speed to get them to have those things.

1:06:45Jason Calacanis:and it's like yeah the best that's why that's why i think the best word that i have come up with is adventure i like it um and uh i would like my kids to have a chance to have an adventure i mean i've had a i've had a great adventure you know and it's still it's still only half done um yeah and you know you've had a great adventure like it's like this is what you mean the kid from toronto i mean you were toronto originally ottawa ottawa kid from ottawa kid from brooklyn and now look like and we still have 20, 30 great years. Who knows if they're going to solve some life extension. God forbid we'll be doing episode 2000 of All In somewhere.

1:07:23But yeah, I've been thinking about it a lot too. I don't have the easy answers, but the one I came up with is very similar to your adventure, which is exposure. Exposing them to possibilities, exposing them to high agency people. And I think if you can expose them to these things, then it's enough because young people have that energy they'll find something that will pull them to it but i wish somebody had exposed me earlier to entrepreneurship a little earlier to risk taking i had to find that myself right you had to find it yourself eventually um and that's like i'm taking them to uh tokyo with me for founder university and i'm gonna let them sit in on it And I was like, I just thought to myself, man, if a 10 or 16-year-old J-Cal had experienced Tokyo, you know, my first time on an airplane was when I was 15, 16 years old.

1:08:16I went to Florida. We didn't have the money for airline tickets. That was out of the question. Our vacation was, how far can we drive the van before it reasonably breaks down?

1:08:24Jason Calacanis:Were you upset when your dad had only bought you a one-way ticket and left you on? Yeah, he was just like, good luck getting back, son. It's always great to catch up. Congratulations, 80-90. I think I got to talk to you about branding. You're getting very good at branding. I love the with me branding. Great on that job. On 8090 and Software Factory, both of them are interesting names. 8090. Hey, 80 % of the cost for 90 or 80 % of the features for 90 % less. 90 % less is how it started. I'm not sure if we'll deliver that, but that's how it started. Still great as an idea. But Software Factory is iconic.

1:09:04so i think we could retire 80 90 at somewhere and just call it t sf like there's tfl the french laundry we went to this is like t s f and people start referring to it hey did you talk to t sf hey what did t sf say or have you tried t sf to solve that problem just you know t sf it or whatever because when we were up at the french laundry which how great of an experience was two days that French Laundry. Yeah, that's amazing. Give people a little sample of how crazy our lives are now and just getting to go hang out with TK up there.

1:09:40Jason Calacanis:I mean, I think being around these kinds of people are very inspiring for me. It's like, you know, you can be as successful as you want, but then there's just a lot of people that stop. And then there's a certain core group of maniacs that never stop. And I admire that, you know, Buffett, Thomas Keller. These guys could have stopped a long time ago. They keep going. And I just find it deeply impressive. And the intentionality. Why is that? I think it's because they're playing their own internal game. You know, the problem with like defining your life in terms of external measure is at some point you'll hit it.

1:10:17Jason Calacanis:Yep. Then what? And then you'll think, oh, I guess that's it. And then you kind of stop. And instead, I think you have to be deeply selfish about this, which is you got one trip around the sun. Just never stop. Yeah. Just keep, why not keep at it? All the things you want to do, just never stop. Yeah. Be relentless. That's another thing for these kids. I'm big on the, you know, some things take like four days or five days to get good at. Some take, you know, five hours. So if you look at skiing, right? Like, gosh, when you started skiing, man, it was like watching Big Bird coming down the hill blindfolded.

1:10:52And now you're like, whoa, cutting S turns. You could do a diamond, but you stay on the blues. Like, you could ski. You could ski. five years later you can actually ski ish you're like the sri lankan skier it's like unbelievable you could be on the sri lankan olympic team i don't know if there's the only the only person

1:11:10Jason Calacanis:more awkward than me is freeberg which gives me some solace i mean freeberg come a diamond every time he comes down it's a knee injury he's i'm like are you coming out tomorrow he's like no my knee my hip my this my that poor guy he's not billford but what i did with my daughter my 16 year old specifically was i was like what are things we can do together daddy daughter date that take like that third or fourth time to have the unlock where you go, ooh, I like this. So pickleball, skiing, anything like that that takes like three or four times to get good at, it just builds that resiliency. And once they've done the third or fourth one, they're like, what's next?

1:11:44What's next? What skill can I learn that's kind of hard? There's your hour plus with my bestie, Jamal Paliha. Love you, bro. I really appreciate you, Jason. Thanks for supporting me. I appreciate you too, bro. Yeah. Oh, thanks for coming back on the show. as one of the Twist All-Stars for summer of 2026. Nine more to go, folks. It's going to be a great summer. Keep tuned in to the podcast feed. Bye-bye.