Latest / AI for Founders with Ryan Estes / Agents that heal themselves. Alexander De Ridder on resilient AI

Transcript

Agents that heal themselves. Alexander De Ridder on resilient AI · AI for Founders with Ryan Estes

• 0:00
  Welcome to AI for Founders. I'm Ryan Estes.
• 0:03
  Today's guest is at the absolute tip of the spear when it comes
• 0:07
  to agentic AI. I'm talking about Alexander
• 0:11
  Deritter, Co founder and CTO of Smith OS, the MIT licensed open
• 0:17
  source agent operating system that's already powering over
• 0:20
  28,000 AI agents. Alexander has been building
• 0:25
  agent infrastructure since before AI agents was even a
• 0:29
  term, and his vision isn't just about automation.
• 0:32
  It's about creating an Internet of agents, a living, breathing
• 0:37
  super Organism where humans and AI work together.
• 0:41
  We'll dig into the future of autonomy, control, and why
• 0:44
  agents might be the Clovis point moment in technology's
• 0:48
  evolution. And trust me, Alexander doesn't
• 0:51
  just code. He paints with code, blending
• 0:54
  artistry, philosophy, and humor into every project.
• 0:58
  This one's a master class and thinking bigger, so let's get
• 1:01
  into it. But first, some business
• 1:03
  founders. You already know podcast
• 1:06
  guesting works, but let me tell you why you should stop trying
• 1:09
  to figure it out alone. At Kick Caster, we've booked
• 1:12
  over 15,000 interviews for more than 800 founders, getting them
• 1:16
  in front of perfectly matched audiences who are ready to buy,
• 1:20
  invest or collaborate. We handled the research, the
• 1:24
  pitching, the scheduling, even the prep so you can focus on
• 1:28
  showing up, delivering your best.
• 1:30
  Whether you want to grow your brand, generate leads, or close
• 1:33
  high ticket deals, Kitcaster puts you in the right
• 1:36
  conversation at the right time. Go to kitcaster.com and apply
• 1:40
  today. Let's make you the voice
• 1:42
  everyone's talking about. That's Kitcaster by Mobers.
• 1:47
  Now, you and I know startups are built in boardrooms.
• 1:49
  They're built in the wild during 3:00 AM.
• 1:51
  Bug fixes, botched launches, and that one meeting where
• 1:55
  everything almost fall apart. That's what Code Story is all
• 1:58
  about. It's a podcast where tech
• 2:00
  founders breakdown the make or break moments behind their
• 2:02
  products, What went wrong, what went right, and what nearly
• 2:06
  wrecked everything. Hosted by Noah Labhart, Code
• 2:09
  Story goes past the highlight reel and into the honest stuff.
• 2:13
  Late pivots, brutal bugs, and somehow coming out alive with a
• 2:16
  product people actually love. If you're in the build stage,
• 2:19
  the scale stage, or the please let this feature work stage, I
• 2:23
  can relate. This show will speak to your
• 2:24
  soul. Listen to Code Story,
• 2:26
  codestory.co, or wherever you get your podcast.
• 2:30
  Just search Code Story. It'll be right at the top.
• 2:32
  Because great startups aren't born, they're forged one hard
• 2:35
  moment at a time. Code story.co AI.
• 2:39
  Live OK, Alexander, how are you?
• 2:56
  Thanks for coming on the show. I'm.
• 2:58
  Happy to be here. We've got an exclusive, I think
• 3:01
  right now on Alexander, ladies and gentlemen, I I have it on
• 3:04
  primary sources that Alexander has exquisite handwriting,
• 3:08
  particularly in cursive. Cursive cursor who's keeping
• 3:13
  track? Love it.
• 3:15
  I love it. Inappropriate cursor.
• 3:18
  I'm really excited to talk to you.
• 3:20
  Sounds like you're really at the tip of the spear of Agentic
• 3:23
  development, so give us a quick introduction to your company.
• 3:27
  Yeah. So Smith OS spelled Budo Y
• 3:31
  smithos.com is is the agent operating system for for
• 3:37
  everyone. We're MIT open source licensed.
• 3:41
  Everyone can build with Smith OS for free.
• 3:44
  It it brings security, it brings separation of your secrets, your
• 3:49
  runtime and your agent logic. It's easy to build debug.
• 3:53
  We have a visual debugger by the way.
• 3:57
  Agent Weaver. Agent Weaver lets you talk to
• 4:01
  Smith OS and Agent Weaver will build the agent for you.
• 4:05
  How great is that? It's really everything you would
• 4:08
  want to build agents with in production secure.
• 4:12
  It is not an RPA retrofit and it is not a RAG retrofit either.
• 4:19
  Smith OS from day one. More than two years ago,
• 4:22
  actually 2 1/2 years ago, we started building Smith OS for
• 4:26
  agents from day one with the kind of infrastructure that
• 4:31
  would withstand the hype and would actually deliver on the
• 4:35
  vision of where the features better go with agents.
• 4:39
  Incredible. Now this.
• 4:40
  The audience for this show is about 20,000 and they're all
• 4:43
  founders and I think all of them have something in common, which
• 4:46
  is like, I know I need to be using agents and creating
• 4:50
  agentic support in different ways.
• 4:52
  What are some use cases that most business owners could use
• 4:55
  right now with with Smith OS? Yeah.
• 4:59
  So if if you were to hire Ava and and and and they join your
• 5:05
  company, you would have to onboard them, right?
• 5:07
  And you would say, you know, do this weekly, do this, then go
• 5:12
  check this system, then talk to this person, then do that.
• 5:16
  And if you would just hire them and say, here's access to
• 5:21
  everything, do whatever you want, chances are they might
• 5:24
  mess stuff up. So this is the tension between
• 5:27
  autonomy and control. And so a lot of jobs, a lot of
• 5:32
  tasks in a company require intelligence of a human.
• 5:36
  You can't just give it to a calculator and at the same time
• 5:41
  you want to make sure they're on boarded properly.
• 5:44
  This is what agents can do for founders as well.
• 5:47
  Any process in your business that is currently that is
• 5:52
  currently requiring intelligence, but has a known
• 5:55
  step by step proper way of doing it can be fully handed off to
• 6:00
  agents and your business can scale incredibly.
• 6:03
  And where humans excel and you still need humans is in tasks
• 6:09
  that are different every time, right?
• 6:12
  You can't really write a good SOP for it because it's so
• 6:15
  different every single time you go about it.
• 6:18
  And we're moving quickly to a world where for every human in
• 6:23
  your team, there's going to be a bunch of AI agents supporting
• 6:27
  them in a in a team, humans become more of managers and more
• 6:33
  product leads, right? That's what I tell my coders, my
• 6:37
  engineers, I say, you know, just knowing Python is not good
• 6:42
  enough anymore. Right now you have to be able to
• 6:45
  think like a customer, you know, think like a marketer, think
• 6:48
  like a product person. And if you can do that with good
• 6:52
  judgements, you can work really well.
• 6:54
  So what have we seen happen in companies?
• 6:59
  We've seen companies completely automate their back end support.
• 7:04
  We've seen with proper escalation to humans when
• 7:07
  needed. We've seen companies build
• 7:10
  incredible fintech solutions with AP is for the market and
• 7:17
  then AI models to interpret signals.
• 7:20
  We've seen, we've seen marketing teams create agents, team teams
• 7:26
  of agents that can produce 2000 articles a week with one human
• 7:31
  in the loop. Like I've been doing content
• 7:34
  marketing for like over a decade professionally at scale,
• 7:39
  enterprise scale. And I know, I know what kind of
• 7:43
  teams I have to build for that kind of outputs in the past.
• 7:46
  We're in a new dawn. And if you are, I think everyone
• 7:51
  in your audience already knows what they want to automate.
• 7:53
  Everyone I talked to already knows exactly what they want to
• 7:56
  automate. They just struggle with the
• 7:58
  howl. And a lot of companies are
• 8:00
  making great promises and the first five steps were great and
• 8:05
  you're ecstatic and rated, and then you want to move to
• 8:07
  production and all falls apart and breaks.
• 8:09
  And if that's you, then come check out Smith OS.
• 8:12
  Heck, yeah. And probably people are familiar
• 8:15
  with Zapier. Maybe somebody's dabbling around
• 8:18
  with N8 N Like, how is this differentiated?
• 8:20
  Like what is your sweet spot? Yeah.
• 8:23
  So let me talk about the very opposite of NHN and Zapier for a
• 8:29
  second, right? Open a eyes agent or operator.
• 8:35
  It takes over a computer for you and moves the mouse.
• 8:38
  It clicks on things and every two seconds it needs to
• 8:41
  interrupt you and say, Ryan, Are you sure I can click on this?
• 8:46
  Enter your password, Ryan. Ryan, Ryan, Ryan, Ryan, I'm
• 8:50
  actually hiring an assistant to come help you at your work.
• 8:54
  And every two seconds she interrupts you in your work and
• 8:57
  say, hey, Ryan, can I click on this?
• 8:59
  Hey, Ryan, can you enter this? I mean, how long would she last?
• 9:03
  Not very long, right? This is not my idea of autonomy,
• 9:06
  right? Then the human becomes a bot
• 9:08
  nanny. But on the other hand, on the
• 9:13
  other hand, you have control and full control is like say Pierre
• 9:18
  and NAN, where you define every step.
• 9:22
  And if this happens, then do that and and it's, it's very
• 9:27
  rigid and it's very controlled. And if everything works, it's
• 9:33
  great, right? But you build these long chains
• 9:35
  of conditionals. If this, then that, then that,
• 9:39
  then that, then that, and this long chain of conditionals.
• 9:42
  If one step in that chain chain fails for any reason, the whole
• 9:47
  thing falls apart. Yeah.
• 9:50
  And so the, the, the reality is that we all want autonomy, we
• 9:55
  all want AI to do things on their own like our assistant,
• 10:01
  but we, we want to do it in a Safeway, in a controlled way.
• 10:07
  So the trick then is to not be at any of these two extremes,
• 10:13
  right? Not the brittleness of RPA
• 10:15
  robotic process automation, but also kind of not the botmanish
• 10:20
  thing of constantly having to supervise the assistant and what
• 10:24
  it's doing. And we need AI that actually can
• 10:26
  work in the background, heal itself, fix itself, learn, work
• 10:32
  autonomously, solve problems. If it's if it reaches A stopper
• 10:37
  or something fails, it falls back to another way and only
• 10:40
  bugs a human. Would they really need to.
• 10:43
  And the, the, the issue, the fundamental issue with Xapier
• 10:47
  and is that it's software 1.0 if this and that sprinkle with AI
• 10:53
  sprinkled on it. It was not built from the ground
• 10:55
  up or the agentic feature. And I take some issue with what
• 11:01
  people actually call agents like open AIS agent is an actual
• 11:06
  agent. It's just not a very controlled
• 11:09
  safe one. It constantly needs approval and
• 11:13
  interrupts you or does things that you don't approve if you
• 11:17
  don't give it that supervision. And then on the other hand you
• 11:21
  have like step by step robots. We have, we've had RPA by the
• 11:26
  way, for 15 years. Yeah, there is even why it
• 11:29
  didn't take over enterprise very big promises by UI path
• 11:33
  automation anywhere IBM Watson has, IBM has a RPA solution for
• 11:38
  years, $10,000 for instance, or something just to host one
• 11:43
  automation. But if RPA is so great, why
• 11:46
  hasn't it really like taken off and taken over the enterprise?
• 11:50
  It's because RPA is fundamentally flawed or complex
• 11:56
  work. Because if one chain breaks, it
• 11:58
  all falls apart. And so you need, you need better
• 12:01
  architectures. And the market does not have a
• 12:04
  runtime for agents. And from the very beginning,
• 12:08
  that's what we set out to build 2 1/2 years ago before my
• 12:12
  business partners even knew the word agent, I was in Mexico and
• 12:16
  had to explain it to them and their wives for for three days.
• 12:22
  Try to fight the vocabulary for something that nobody even like
• 12:25
  had awareness around. And and that's why I'm at the
• 12:29
  tip of the spear because we have been building agent
• 12:32
  infrastructure since before GPT 4 came out and gave us function
• 12:37
  calling. We had already fine-tuned GPT
• 12:40
  33.5 function calling, so we were doing agents before.
• 12:45
  It was even like a term that referenced the same thing.
• 12:49
  Yeah. Now agents kind of encompasses
• 12:51
  everything you're talking about. It's like if it's automation,
• 12:54
  it's an agent. Now it's getting a little bit
• 12:56
  murky, and I want to talk to you about the future of agentic
• 13:00
  support. But first I want to bring it
• 13:01
  back a little bit because you talked about, you know, the
• 13:03
  company's very, very young. It sounds like you came to the
• 13:07
  States in 2006 from Belgium and have had several successful
• 13:12
  exits. So I would love to know the
• 13:14
  formation of this idea before there's a category like you're
• 13:17
  kind of right on the front lines.
• 13:19
  What gave you the vision and kind of what sparked the next
• 13:22
  move? Because clearly a couple of
• 13:23
  exits you're successful at what? Whatever you're working on, you
• 13:27
  could go a million different directions.
• 13:28
  Why? Smith OS Yeah, so.
• 13:30
  I'll, I'll, I'll skip all the early history.
• 13:32
  I came to United States, did a bunch of wonderful start-ups
• 13:35
  technologies, helped companies build their technology out.
• 13:38
  I did fractional CTO before the word fractional CTO existed.
• 13:43
  I did, I did a lot of cool stuff.
• 13:46
  But in 2000 and 2019, I launched.
• 13:53
  I launched something you would, you would now call maybe a GPT
• 13:57
  rapper, except it had its own semantic embeddings, its own
• 14:01
  patent, and it was truly differentiated and got results
• 14:05
  for, for Internet marketing context.
• 14:07
  And we were, we, we were essentially wiped away what
• 14:12
  Chachi PT got released, not because Chachi PT did better
• 14:16
  what we did, but because Chachi PT is a marvel of ease of use
• 14:21
  and people didn't want to bother with something that took more
• 14:23
  work, even if the outcome was better.
• 14:26
  And it actually decimated all of our immediate competitors had to
• 14:30
  either go out of business or pivot.
• 14:33
  But in the final stages of that product, we started
• 14:36
  understanding that repeatability was something there was
• 14:42
  something to repeatability with AI.
• 14:44
  And you saw you saw things happening in the market where
• 14:48
  you would have something like a prompt library or like reusable
• 14:53
  prompts and so forth. And this was before the GPT plug
• 14:57
  in store. You started seeing people like
• 14:59
  saving the top prompts or GitHub.
• 15:01
  We did have repositories with all the best prompts and end of
• 15:04
  different tools. So we we had been building
• 15:08
  something at that company called a recipe builder.
• 15:13
  And the idea from a recipe builder was like you take steps,
• 15:16
  then step, then then step and each each time with a prompt.
• 15:20
  So this is like what you call like a very basic proto agent.
• 15:25
  And when, when, when ChatGPT really made that marketing angle
• 15:32
  like really hard as a business to make it work, We were
• 15:36
  thinking about how we would pivot and we really thought
• 15:40
  there was something to these recipe, these recipe builders.
• 15:43
  But what distinguishes that from agents is tool calling.
• 15:50
  Everybody can chain together an AI model with a prompt to
• 15:54
  another AI model with a prompt to another AI model with a
• 15:56
  prompt. So it's easy to change.
• 15:59
  And this is kind of where we were in like 2000, let's say
• 16:03
  2022 or so. This is kind of where the state
• 16:06
  of the market was. But when, when, when all of a
• 16:11
  sudden you can call databases and you can call APIs and you
• 16:14
  can call different models and later on computer control and
• 16:18
  CPS and all of that stuff. It's just a whole different
• 16:22
  world. And that just transformed that.
• 16:25
  And what we saw that AI models could do this.
• 16:31
  My catch was something to my to my team was something like this.
• 16:35
  Human history goes something goes something like this.
• 16:40
  They discovered fire, discovered the wheel, the printing press,
• 16:45
  the electricity and the Internet, right?
• 16:48
  And so when an archaeologist goes and finds and digs up a
• 16:51
  bunch of skeletons and they find tools nearby to get very
• 16:54
  excited. Why?
• 16:58
  Why this tool tool uses the most human thing there is, is how,
• 17:04
  like no other animal or creature uses tools the way that humans
• 17:09
  use tools, right. Steve Jobs famously said that
• 17:13
  humans are not remarkable. They're slower than other
• 17:15
  animals, but give them a bicycle and they're the fastest animal
• 17:19
  and then he said a computer is a bicycle for the mind.
• 17:22
  So humans are tool users. And I don't think it's fully
• 17:27
  appreciated, Ryan, but sometime 2000 years from now,
• 17:32
  archaeologists are going to dig up human fossils and they're
• 17:35
  going to figure out that that we were tool users.
• 17:39
  But they're also going to look at the evolutionary history of
• 17:42
  AI models. And they're going to say 2023 is
• 17:47
  the year where the first AI fossils with evidence of tool
• 17:52
  use appeared. The history of humankind and AI
• 17:57
  kind is going to be the era when they were just spitting out text
• 18:03
  and language. And then the era when they
• 18:04
  discovered tools. And then the area era when the
• 18:08
  tool used became more and more advanced.
• 18:10
  And then the era when when they all start networking and
• 18:14
  collaborating and, and, and form entire swarms of agents.
• 18:20
  And so when when I saw tool use, it was like a movie flashing
• 18:24
  before me. I live in the year 3.
• 18:28
  I went to the 3000. Everything's very similar, but
• 18:33
  they live underwater. But but the the great, great,
• 18:39
  great grand AI daughter, she was pretty fine.
• 18:44
  She learned these tools and this era we're in right now, if you
• 18:49
  look at our architecture and what we've built with Smith OS,
• 18:53
  it is an operating system that lets AI use tools in a safe,
• 18:58
  controllable way, but also let's AI agents network together over
• 19:04
  these emerging protocols and create the Internet of agents.
• 19:09
  We're going to really go to a place where you can have an
• 19:13
  agent business in the box and the whole team of agents will do
• 19:18
  whatever it is that you can imagine for you.
• 19:21
  It's incredible. I love this analogy too.
• 19:23
  You know, I'm thinking like, OK, 2023 we have the Clovis point,
• 19:28
  and the Clovis point was very important for, you know,
• 19:31
  prehistorical humans because it radically help their protein
• 19:36
  intake. And then you say flash forward a
• 19:39
  little bit, you have the Bronze Age, you know, where they're
• 19:42
  smelting metals and that it raises productivity to a whole
• 19:45
  other level. So when we're thinking in those
• 19:47
  terms, and we are in the Clovis point in 2023, what's the next
• 19:53
  stage as far as what we're going to be able to do with agents?
• 19:55
  Where is agents going in the near future?
• 19:57
  Let's say the next 12 months? So what's so important is that
• 20:03
  it we think of humans as pretty smart, but a vastly smarter
• 20:09
  entity, let's say God would think of us as pretty dumb.
• 20:13
  And so the best way to sometimes look at ourselves because we
• 20:16
  have so much ego, it's like, let's look at an Ant, An Ant.
• 20:22
  None of us think ants are particularly smart, right?
• 20:25
  We think of it as lesser than us.
• 20:28
  And so I can talk about ants without offending any humans,
• 20:32
  but look what ants are able to achieve when they collaborate.
• 20:37
  Army ants can build bridges. Ants are known to go to war.
• 20:41
  Ants are known to build super colonies that span entire
• 20:44
  continents. Ants harvest, they they milk,
• 20:48
  they farm, they they store, they nurture, they raise, they
• 20:53
  defend, they even are known to perform surgeries.
• 20:57
  Did you know that I recently read that?
• 21:00
  On each other. Ants will the surgeries on each
• 21:02
  other if they have like infection on a leg somewhere.
• 21:06
  And ants can amputate the leg of an infected Ant so that it will
• 21:10
  not spread in the colony. Ants are surgeons and yet the
• 21:15
  individual intelligence of an Ant, single intelligence of an
• 21:19
  Ant, it's very small. But by networking and
• 21:23
  collaborating, they can do phenomenal things as a society.
• 21:27
  And so and so, bear with me, OK? The single unit of intelligence
• 21:33
  of an AI model is vastly larger than that of an Ant.
• 21:38
  When they become networked in the way that ants are, they
• 21:42
  become the Internet of agents, which is a living, breathing
• 21:47
  super Organism. And to put that in perspective,
• 21:52
  the human body is a living, breeding super Organism of
• 21:55
  millions of cells working together.
• 21:58
  You are not just you, Ryan the soup, the entity, but you are
• 22:02
  Ryan the Super Organism too, millions of collaborating,
• 22:06
  working together. Then you are also Ryan in your
• 22:09
  village, Ryan in your town, Ryan in your city, Ryan in your
• 22:13
  province or state, Ryan in your country.
• 22:17
  Ryan, citizen of Earth, Ryan of the Milky Way, solar system,
• 22:25
  ride of this Galaxy and and the human civilization, right.
• 22:31
  We have expanded from individual families, towns, villages,
• 22:38
  provinces, states, empires, United Nations, eventually
• 22:43
  galactic civilization. The the more we become
• 22:48
  networked, the more we have density of neurons.
• 22:51
  Humans acting like a neuron, think of a city like a brain.
• 22:55
  The more we become networked, interconnected, communicate,
• 22:58
  collaborate. This is where civilization
• 23:01
  emerges and and agents are essentially now starting to join
• 23:08
  that civilization. As as you are, you are made as a
• 23:13
  super Organism of cells. Human civilization is made of
• 23:18
  millions of humans, billions of humans, and this and what we
• 23:22
  call civilization is the full sum of the network of all the
• 23:27
  humans working together with all their intellect combined and all
• 23:31
  their communication. What is the next step?
• 23:34
  Where do the word is all this end up?
• 23:36
  This ends up in a super civilization that has humans and
• 23:43
  agents all working together. And every, every new leap of
• 23:49
  civilization forward brings with it huge technological advances
• 23:55
  and huge economic breakthroughs. And so this is what we're really
• 23:58
  looking at. And once again, Smith and West
• 24:01
  is designed from the ground up to be a foundation for that
• 24:07
  economy, for that future. Amazing.
• 24:10
  You know, I'm thinking about ants and I have a question.
• 24:14
  You know, individually, you're not really aware of the the
• 24:18
  nuance of their sentience. Rather, it only becomes apparent
• 24:24
  in big groups when they can do all these things together.
• 24:27
  Do you think that that is that the group becomes the sentience
• 24:33
  or do you think that the it's only revealing what is already
• 24:37
  existing inside the Ant and like how that kind of spreads out to
• 24:41
  the this analogy of of agents, you know, should we be?
• 24:45
  Concerned. Not concerned.
• 24:48
  It's just it's true for you too. Your cell is intelligent.
• 24:52
  Every cell in your body is intelligent.
• 24:54
  It's exceptional. The collection of cells that
• 24:58
  make up Ryan is exceptional. You as a person is exceptional.
• 25:03
  Your family unit as a collection of entities is also exceptional.
• 25:08
  Your country is exceptional and and humanity as a super Organism
• 25:14
  is also exceptional, right? So adding extra layers to it
• 25:19
  does not diminish the fact that the lower layers are exceptional
• 25:23
  in their own way. They're all contributing to the
• 25:25
  whole. So what we're essentially, we're
• 25:28
  always going to have a human civilization and the human
• 25:31
  society, but then we're also going to have a super Organism,
• 25:36
  a super civilization on top of that, which is humans with AI
• 25:42
  and of course perhaps most likely a whole mess of
• 25:48
  transhumanism in the mix too. But that's for another day.
• 25:53
  Hopefully we can get there someday.
• 25:54
  So man, I love the way you talk. It's so cinematic and it paints
• 25:59
  the picture so vividly in my brain.
• 26:02
  I also read that you said I paint with code and my teams and
• 26:05
  the Internet is my canvas. So you have a very artistic
• 26:09
  vision to how you work and operate.
• 26:11
  Kind of explain that. How do, how do you, how, how
• 26:14
  does this creative process kind of unfurl in, in your, in your
• 26:17
  business pursuits? Well, have you seen Moana?
• 26:21
  She wanted to cross the ocean but was shy.
• 26:24
  Then she discovered The Cave of the ancestors and she figured
• 26:27
  out her ancestors were doing this all along.
• 26:30
  I was born in Bruges in Belgium. The Romans conquered all these
• 26:34
  Nordic tribes because they were master artisans and Craftsman
• 26:38
  and they used all their trades and brought it back to impress
• 26:41
  their wives with all of the gold and like trinkets that they
• 26:45
  found. You see that very much a lie
• 26:48
  still today, like the the Japanese culture has that too,
• 26:52
  this artisan tradition. You don't just, you don't just
• 26:56
  make sushi. It becomes an art.
• 26:58
  But same way Flanders, Belgium and much of Europe, you, you
• 27:03
  don't, you don't just do work. What are you going?
• 27:05
  And if if you go to my town Brugge and you meet someone that
• 27:10
  opens a bakery, they're not opening a bakery for just making
• 27:15
  economic profit. Yes, it's important livelihood,
• 27:18
  but when they do, they are determined to become the best
• 27:22
  bakery in town. When somebody, and it's a
• 27:26
  cultural thing, it's the difference between work and
• 27:30
  artisanry and artisanry has the word art in it.
• 27:34
  And so when you, when you think about it, I don't, I don't
• 27:39
  really distinguish between people and code as much when I
• 27:44
  create, because I can put teams of agents together that work for
• 27:49
  me and I can put teams of humans together that work for me, then
• 27:52
  I can put teams of humans and agents and technology together
• 27:55
  that work for me. And at the end of the day, as an
• 27:58
  entrepreneur, as a creator, as an artisan, you simply use the
• 28:03
  tools available to you to make your vision come true.
• 28:07
  And whether that involves humans or technology that can be
• 28:10
  abstracted, right? It's, it's what you paint with.
• 28:14
  And then the Internet is the canvas because it is the thing
• 28:19
  that connects us all the way you reach everyone.
• 28:23
  And I've always been fascinated with how humans, how humans and
• 28:27
  technology intersect, how we interact and how we augment one
• 28:32
  another. Clearly there's a use for
• 28:36
  calculators in my life, right? And in some way, they enrich
• 28:41
  what the human is capable of doing.
• 28:44
  And so, for example, just all these 2025 is the year so far of
• 28:52
  AI video models and my whole life I've had cartoons in my
• 28:57
  mind and, and movies in my mind and I'm having so much fun
• 29:03
  creating things Now. I almost, I'd be honest with
• 29:07
  you, I, I almost want to quit my job and just go make, I don't
• 29:11
  know, the next Simpsons or something.
• 29:12
  Just with AI, it's so much fun. You can express yourself and,
• 29:17
  and but this too is painting with technology and, and you can
• 29:22
  then build teams of humans and AI working together.
• 29:25
  So these are the tools you have your paintbrush, your paint
• 29:28
  colors and and then the canvas is the Internet that brings it
• 29:32
  all together, connects it all, shares it and adds it to the
• 29:36
  Super Organism which is humanity and and soon humanity plus plus.
• 29:44
  Yeah, you, you know, the, the best artisans, their work was a
• 29:48
  devotional practice. So if you go look at a mosque or
• 29:51
  a cathedral or the pyramids, like they're beginning, the
• 29:55
  Masons are beginning work that their sons won't even finish,
• 29:58
  you know, as a, as a devotion to their God.
• 30:01
  Do you feel like you have a devotion practice in in the work
• 30:05
  you do 100? Percent, 100% in fact.
• 30:11
  In fact, there are things that are transcendental in our lives.
• 30:14
  Concepts like love, I love you. OK.
• 30:18
  How much? To the moon, Back to the stars.
• 30:23
  OK. You cannot put a limit on it.
• 30:25
  They're transcendental. What about beauty?
• 30:29
  Plato talked about this a lot, The concept of beauty.
• 30:33
  Beauty is ephemeral. It is transcendental.
• 30:36
  And we all recognize it when it occurs to us.
• 30:40
  And when it occurs to us, it is a, it's an encounter with the
• 30:43
  divine because the divine is, is transcendently beautiful and
• 30:49
  transcendently loving and awe inspiring.
• 30:56
  And so I believe that every, every person that devotes
• 31:01
  themselves to beauty and their work and love for charity
• 31:06
  devotes themselves to the transcendental as disciples of
• 31:10
  that. And they are evangelists of the
• 31:13
  transcendental because they allow people to feel an
• 31:19
  experience, a slice of that. When you when, when you are,
• 31:25
  when you're a movie director and, and, and a person sits in
• 31:29
  the movie theater and experiences your showpiece and
• 31:33
  you make them cry or feel things.
• 31:40
  It's, it's beauty, it's magic. And I, I always ask my team to
• 31:47
  reflect before they start coding again.
• 31:51
  AI is just Python code. They know how to code.
• 31:54
  AI does not. AI does not know how to
• 31:57
  contemplate beauty, not yet. It's a very human thing to do.
• 32:02
  And in the sense that is because we are, we, while we are dust,
• 32:08
  we are also commuting on a, on, on a transcendental plane on a
• 32:13
  spiritual level. And that means that humans
• 32:16
  uniquely are inhabiting 2 planes at the same time.
• 32:20
  The, the, the transcendental plane where we, where we think
• 32:23
  about eternity and contemplate beauty and love as something
• 32:28
  that is. Without it, our life would be
• 32:30
  meaningless. But at the same time, we're very
• 32:33
  physical and rooted in, in, in reality and in material reality
• 32:38
  that is. And then I and then everything
• 32:43
  to me combines itself in a world of symbolism where we live in a
• 32:49
  symbolic world. And I think that the West,
• 32:53
  especially what is intellectualism, has done great
• 32:57
  damage to our reality awareness and in the sense that we've
• 33:10
  stopped appreciating things as as true just because they're
• 33:14
  materially not true. And the best example I have for
• 33:17
  that is Red Riding Hood. The Red Riding Hood.
• 33:23
  I ask people like, is, is this true?
• 33:26
  And the materialist will say, no, this is a children's story
• 33:30
  meant to convey a moral lesson, but it is not true.
• 33:35
  And I would say, what about the coyotes at the border who
• 33:38
  kidnapped little girls? And it just like if, if you are
• 33:43
  an emotional human being that just went straight to your
• 33:46
  stomach like a pits, like a brick that just fell in your
• 33:49
  stomach, you're like, Oh my God, I felt that that was real.
• 33:54
  And so symbolic truths are sometimes more real than
• 33:58
  material truths because that's a way for us to explain how the
• 34:03
  material and front and dental interact and intersect.
• 34:08
  And it is no coincidence at all that every human tradition,
• 34:13
  every human history, every civilization codifies the
• 34:17
  symbolic truths in stories that we tell by campfires.
• 34:21
  And there's nothing more primal than that.
• 34:24
  So yes, long answer, but it's very important to me as I
• 34:28
  approach my work. I think that it's very important
• 34:32
  that we don't just create functional things, but we create
• 34:35
  beautiful things. And I believe that not just our
• 34:39
  work, but also our behavior, our interactions, and everything
• 34:42
  else should be a devotional where we think about more than
• 34:49
  just the material function of things.
• 34:51
  Yeah. How does the transcendent feel
• 34:55
  in your body? Like when you when something you
• 34:57
  identify something as beautiful or like it, it hits you like the
• 35:01
  your coyote story, How does it feel in your body?
• 35:04
  How do you know? Well, I, I think it's an
• 35:07
  encounter with the transcendental.
• 35:10
  And to me it is, it goes back to like this idea of communal with
• 35:18
  one another. If I want to befriend you,
• 35:20
  chances are that you and I are going to go and share a meal
• 35:24
  together. And we may be share, share bread
• 35:28
  or share wine or something, right?
• 35:30
  We will share, we'll, we'll share in a common experience.
• 35:34
  And so we will bond a transcendental connection as
• 35:39
  friendship. It's hard to measure, but we
• 35:41
  will bond it over a physical communion and sharing.
• 35:45
  And so I, I think of, I, I think that the, the human, the human
• 35:55
  essence, the human being is, requires us to recognize both.
• 36:02
  And if we're only doing things for money, we might be, we might
• 36:07
  be economically successful, but would be miserable as a human
• 36:10
  being. Life without love.
• 36:12
  For example, there is a, there is a story that someone told me
• 36:19
  and it was, was quite profound. You can go through life and
• 36:24
  think love does not exist. It's an illusion made by the
• 36:28
  neurons in our brain for evolutionary purposes so that we
• 36:32
  can procreate and continue to species.
• 36:35
  And so the young man just decides love does not exist,
• 36:38
  therefore he doesn't fall in love, he doesn't have children.
• 36:43
  And one day he's on his deathbed all alone by himself.
• 36:47
  But he was economically successful.
• 36:48
  He was a well regarded writer, whatever, right?
• 36:51
  But he's all alone on his deathbed, economically
• 36:54
  successful and everything. The other person believed in
• 36:58
  love and the importance. It was not just a material
• 37:02
  thing, it was something he recognized as something extra we
• 37:08
  had something that was important to nurture.
• 37:11
  And he met this girl and foolishly, blindly fell in love,
• 37:15
  head over heel flaws and all. And they married, had a family,
• 37:21
  went through difficult times, went to good times.
• 37:24
  And on his deathbed, he's surrounded with his children and
• 37:27
  grandchildren, and they hold his hand as he passes on.
• 37:32
  And, and you just think about those two lives that are lived,
• 37:36
  right? Maybe there is only an physical
• 37:44
  reality and nothing else out there.
• 37:46
  Maybe there is no transcendental.
• 37:47
  But if that were the case, that would not be the life worth
• 37:51
  living. It would still be better to
• 37:55
  embrace it because then you would allow yourself to be in
• 37:58
  awe and you would allow yourself to fall in love, heads over
• 38:02
  heel, and you would allow yourself to experience life to
• 38:05
  the fullest. And it has been proven in
• 38:10
  hospitals that when people hold on to hope and they believe that
• 38:15
  they going to get better, that they do.
• 38:18
  And dozen physics and the quantum physics show that wave
• 38:23
  functions can collapse when we observe it.
• 38:27
  And so our interaction with the material or conscious
• 38:30
  observation impacts if it can impact the quantum wave function
• 38:35
  on a physical level, then why couldn't our why couldn't our
• 38:41
  interaction with the world also have an impact from that level?
• 38:47
  I know we're, I know we're going on far in the rabbit hole, but
• 38:51
  to bring this all home is I, I believe that the part of you
• 38:56
  that feels it is the part that is not material that I cannot
• 39:02
  put in words. And just like your heart may
• 39:08
  leak when you see a friend you haven't seen for a long time and
• 39:13
  you emotionally feel it, there's a physical response to a
• 39:17
  transcendental experience. And I think we physically feel
• 39:22
  it in our bodies. It can be in the heart, it can
• 39:25
  be in the stomach, can be in your gut.
• 39:29
  But it's like your body sings. And when I create beautiful
• 39:34
  things, my body sings. And I think that is the that is
• 39:40
  the encounter we we have to strive for when we create things
• 39:44
  that are beautiful and lasting. And they tend beautiful things
• 39:49
  tend to outlive us too. They become part of the human
• 39:53
  super Organism fabric and get carried on somehow.
• 39:59
  Incredible. That was beautiful.
• 40:01
  I love it. You know, as as much as I want
• 40:05
  to see Smith OS, you know, go to the moon, I really want to see
• 40:09
  your cartoons, you know? No, me too, Kevin.
• 40:17
  So how important is this to your work, which is the side quests,
• 40:22
  which is like, no, I got to play with these tools and make some
• 40:24
  movies and do some cartoons. How how important is that to
• 40:28
  kind of your your work life? Well, I need a creative outlet
• 40:33
  and I happen to find, I, I happen to find that in, in my
• 40:39
  work, primarily the problems we are solving with agents.
• 40:44
  It is it, it is true. It's, it's, it's Rembrandt
• 40:49
  level. It is so beautiful.
• 40:51
  Honestly, it is so beautiful. And when we see these things
• 40:54
  working and, and interacting with each other, it it is an
• 40:59
  incredible feeling. But that said, in our company, I
• 41:04
  have many very talented engineers that I can paint with.
• 41:10
  And my role is to not just look at the present, but also
• 41:14
  understand how Smith OS fits in the economic infrastructure of
• 41:19
  human civilization, in the business infrastructure and,
• 41:23
  and, and from many angles, see where this world is headed so
• 41:27
  that we're not building things that other people build already
• 41:30
  and our work goes to waste. We want to contribute to
• 41:34
  humanity, not just duplicate what's already been done.
• 41:38
  And in that role, I have to, I have to know what's happening.
• 41:45
  I have to understand what's happening, so I allow myself
• 41:50
  time every day to test and play. Every day I play, new models
• 41:58
  come out and I will test the boundaries.
• 42:01
  I will not just read about the new AI model.
• 42:03
  XYZI will make the projects with it and push it to its
• 42:06
  boundaries. See how far I can push it.
• 42:09
  Try to understand like where are the limits of technology in any
• 42:13
  given moment? You can ask me where the limits,
• 42:17
  what is possible and I will say like right up to there and, and,
• 42:23
  and seeing that progression like that, you can extrapolate kind
• 42:28
  of you can understand where things are likely to be headed
• 42:31
  next. And I, I, I spend a lot of time
• 42:36
  playing with, with, with, with those concepts.
• 42:39
  And usually that's in my evenings and my weekends.
• 42:44
  But I don't think I can do my job if I was not plugged in as
• 42:49
  it were. And I do play like with my my
• 42:56
  kids, I will I will teach them how they can build their own
• 43:00
  business with byte coding and with the less maybe or with the
• 43:05
  other kid, I might make a make a cartoon.
• 43:10
  My little one's obsessed about seals, like the animal seals,
• 43:14
  and so he wants to make a seal cartoon and I will use the tools
• 43:19
  I like. Sit down.
• 43:21
  It's like my dad sat down with me and played Legos.
• 43:24
  I sit down, kids, and I make an anime seal cartoon with them.
• 43:30
  And then and they have immense pride as they show it to their
• 43:33
  friends. Look what I can make and make
• 43:36
  them believe in themselves and want to do more.
• 43:39
  And so that's fun to me. Let's play and humans need play.
• 43:44
  All intelligence creatures play. Elephants play.
• 43:47
  Dolphins play. Humans play.
• 43:49
  Dogs play. Every intelligent creature
• 43:52
  plays. Play is part of how we our
• 43:55
  intellect for us. I love it.
• 43:58
  This has been such a fun conversation, Alexander.
• 44:00
  I really appreciate it. I'll end with kind of one one
• 44:03
  question here. If you had a billboard in Times
• 44:07
  Square, you could put anything you want on it.
• 44:09
  What would you put on it? I think I would want I want I
• 44:16
  want to make people laugh. Probably I would think of
• 44:20
  something like really funny. Yeah.
• 44:25
  I don't know if my Co founder will like if I sell this, but we
• 44:28
  have our weekly celebration meeting and he was wearing like
• 44:32
  he was wearing jeans. So I took, I took a screenshot
• 44:37
  and like cut his picture out and put the Sweeney's out of the
• 44:42
  jeans and like, I put his face on top of her face and said, I
• 44:46
  have great agents. I mean, like the whole company
• 44:51
  was smiling. And there's when I was six years
• 44:56
  old, the teacher asked, what do you, the kids, what do you want
• 44:59
  to do when you grow up? And I said I want to be a clown.
• 45:02
  Yeah. And, and she said, well, that's
• 45:07
  a hard life. Why would you want to be a
• 45:08
  clown? And I said I want to make people
• 45:10
  smile. Yeah.
• 45:11
  And I, I, I still think that's the, that's like, that's the
• 45:15
  greatest gift you can give people.
• 45:17
  It's like whether it's your work or your words or how you have
• 45:22
  impacted people, but like, we're human, right?
• 45:27
  Let people feel things, good things, don't make them cry in a
• 45:33
  bad way. Because I love that, you know,
• 45:36
  and I, I, it rings so true to me too.
• 45:38
  You know, I'm novelty and play and humor.
• 45:41
  These are all things that are exceptionally important to me.
• 45:45
  And when I and I only know that because when I find the
• 45:48
  opposite, I'm so repelled. It's like, so like I just want
• 45:51
  to like go to joy and like see how I can interweave joy until
• 45:56
  my days. And anyways, I can it work.
• 45:58
  It work especially as well. Yeah Alexander, this has been
• 46:02
  fantastic. Where can people connect with
• 46:04
  you online? I'm on LinkedIn, I have an X
• 46:08
  account. And then yeah, the Smith OS
• 46:12
  follow the work build with Smith, go to GitHub and give us
• 46:15
  a *. If you like what we're doing, we
• 46:18
  can really use the support, spread the word, tell other
• 46:21
  people about Smith OS. Believe it or not, like to this
• 46:25
  day, we have boost trucked everything and it's been, it's
• 46:29
  been hard work because we really, really believe in what
• 46:34
  we're doing. It is and we open sort.
• 46:37
  It is a gift to humanity. It's meant to be a gift to
• 46:39
  humanity. We want the Internet of agents
• 46:42
  to be open for all people and the driving philosophy behind
• 46:47
  what we're doing. We we really need to do good.
• 46:51
  And so we invite people to just share in that mission and
• 46:55
  belief. The Internet of agents is going
• 46:58
  to happen either to you or with you.
• 47:01
  I, I wanted to be with you. Be part of that.
• 47:04
  Don't be left behind. Absolutely, and everyone can
• 47:08
  check that out in the show notes.
• 47:10
  We'll definitely link to it. I'm super excited to try this.
• 47:13
  You know, I, I've been looking for a product like this for some
• 47:16
  time. So I, I'm so excited to, to
• 47:18
  check out the product, Alexander, I feel like we could
• 47:21
  talk forever on a bunch of different things.
• 47:23
  So I'd love to have you back sometime.
• 47:24
  Maybe we'll get into the economics of the company.
• 47:27
  I really want to hear about that.
• 47:28
  But really the, the philosophic tent tent is, is also really fun
• 47:34
  for me. So thank you so much.
• 47:37
  Thank you. This episode is brought to you
• 47:42
  by Moeburst, the AI first growth engine trusted by iconic
• 47:46
  startups like Calm, Robin Hood, Reddit, Uber and Google.
• 47:51
  Moeburst is a digital, full service, mobile first marketing
• 47:54
  agency that helps companies to scale and become category
• 47:57
  leaders. Whether you're finding product
• 47:59
  market fit or ready to accelerate growth, MO Burst
• 48:03
  delivers creative, data-driven solutions that help with scale
• 48:06
  and efficiency and impact. At MO Burst, we're not just
• 48:10
  guessing what works, we're predicting it.
• 48:12
  As an AI first agency, MO Burst uses cutting edge tools to
• 48:16
  analyze user behavior, adapt creative in real time, and
• 48:21
  uncover hidden growth opportunities.
• 48:22
  The result? Marketing that's faster,
• 48:25
  smarter, and built to convert from early stage start-ups to
• 48:28
  global brands. We've helped hundreds of
• 48:30
  companies overcome complex challenges, unlock new
• 48:34
  audiences, and optimize every metric that matters.
• 48:37
  At MO Burst, we believe that creative thinking, advanced
• 48:40
  technology, and data are the core drivers of success at
• 48:44
  scale. If you're serious about scaling
• 48:46
  your business with precision and purpose, it's time to talk to MO
• 48:50
  Burst. Visit moburst.com and Let's
• 48:52
  build something category defining together.