E406: Why AI Won't Transform Most Enterprises for 10 Years

30:13That's 50 % off your first year at monarch.com with code invest. And they had a request and the request was handled incorrectly. There's drastic consequences to that. Or let's say that you were working in a larger threaded warehouse network and some of the temperature settings were set incorrectly. That's going to affect the state of produce in that warehouse. Look, these things have pretty dastic and dire real-world consequences if you get them wrong. So this is not something that could be vibe-coded. This is not something where even 99 % accuracy is good enough. It has to be 99.9 % accurate.

30:46And also you have to know when you're the 0.1 % of times you're going to be wrong. You have to know when and flag that. Do you just wake up and decide I want to go into logistics and supply chain management? No. My entire life I've had this view that you basically, there are two ways to kind of build great companies. One is you go for the most obvious market opportunity that's cool and trendy, like building a great AI coding solution, and you go and you do it. The second way of building a great business is that there's a particular part of the world or a particular part of the economy that's underlooked for whatever reason.

31:21And I remember thinking about this. I think supply chain falls into this world because I think there's a lot of properties where Silicon Valley is not really doesn't really spend any time thinking about like how something manufactured, how does logistics work? Because you're like isolated. It's a completely separate world. So as a result, what ended up happening for a long time was you would have. Great commercial people go and try to build these supply chain tech companies, but they don't know how to build great software. You would have these software people that would try and they had no idea how the industry worked.

31:50so for that's why you have a lot of shitty software a lot of crappy software that's been there for the last two and a half decades so when i looked at that i was like my entire life i've never been the smartest person i've never been the the person that's won by doing the obvious thing i've always gone to the corner that like no one ever wanted to go to and i remember like in my last job i was working with one of the largest freight forwarders in the world and spending time with their operations this is a pretty tech forward flight freight forwarder i remember looking and seeing that even despite how tech forward they were, they were still tracking containers manually.

32:23They had way more operations people than engineers. So I was like, there probably is a gap here that no one's going to be looking and solving. So I left that job and I basically spent six months visiting the office of all these trucking companies from and were freight forwarders and warehousing companies across Yuba City, Stockton, all these parts of the industrial parts of California. And I went to the first customer's house. The guy had like 150 he pieces a paper on his desk and he was typing the shipment documents off his desk one by one by one on a Sunday and he's like, I'm doing this instead of watching the 49ers play the playoffs, I'm typing in shipment data every weekend evening and I'm like, that's crazy that you're still doing this and that's where I felt that here is such an important part of our economy, software is really shitty and I feel like after spending that time I had a pretty, I kind of appreciated the nuances of it so that's what made me want to go start this company.

33:16It's fascinating that you go after the space. I had a three and a half hour dinner with one of the chairmen of largest investment banks in the US. And he's had this phenomenal track record as investor, over 20 % returns for 30 years, did a bunch of deals with Warren Buffett. And he said that he looks for two things when making investments. One is things that are boring and two are things that are hard. And ideally the intersection of both things. When I think about things that are boring and things that are hard, Perhaps the new definition of that should be supply chain because it's just so boring and such a space that a lot of people don't want to go after.

33:55And it seems so difficult from an execution standpoint. I think that's right. If you go for things that are boring and hard, inherently the set of people that are going to go solve it kind of continues to dwindle and dwindle and dwindle. But I will say that one thing that's interesting is that I think the supply chain industry at face value is far more interesting than like categories that I think people think are sexy. because if you think about it that things that you build at pallet affect like 40 like we work with affect all a huge percentage of the food supply chain in the united states or how do the leading defense contractors move their inventory across the country or how do retailers have stock every single day within their stores how does e-commerce fulfillment happen globally like these problems actually have consequences in the physical world so if we build our technology right we actually do impact actual physical good movement and storage globally and that to me is far more interesting than going and building a another social app that gets teenagers addicted to social media or going and building another chat bot that helps e-commerce brands like basically personalized marketing campaigns like it's like all these sort of things that people are working on are just not that impactful so it's like why would you not want to go work on something like this it's a packaging question yeah oftentimes we have a media side to the business in the podcast i have to think about how to package certain things that maybe on the surface look boring how to make them very interesting has that been the secret to how you've been able to recruit such great people to your company the thing i tell people is that look first of all the things that you're probably going to go work on at a lot of these companies which are like a lot of other companies are just honestly not that impactful versus if you come to pallet if we were actually able to automate this 11 trillion dollar logistics industry and create like two three percent more efficient we're actually driving a lot of abundance right from reducing the cost of manufacturing to reducing the cost of distribution these are things that can actually have effects on even the rate of inflation globally so you're working on something far more impactful secondly it's way better to be the smartest team in an area that's not to be attract great talent because it's like being that great poker player in a game where there's like not a bunch of good poker players versus a one dollar table yeah like imagine being the best player but the pie is still big so like what you have is this counterintuitive dynamic where you could be the best poker player in a place where the pot's still big versus all your other friends from MIT and Stanford are going and building coding agents and competing with each other you don't have that dynamic here and the The final thing I would say also is that this is a business that's going to be fairly durable and sticky.

36:41It's not like overnight there's going to be someone else that comes and learns all the nuances of how to work with an international global freight forwarder, builds a distribution network, deploys a system at scale, navigates the enterprise change management. It's really hard to take a pallet deployment out once it's there because it's so critical, so sticky that this business is going to be a lot more. If you had to kind of stand the test of time, and is Palik going to be around in a decade? The answers are probably yes. Can most AI companies say that? Probably not. We were just talking about my interview with Balaji Srinivasan.

37:14Interviewed him for three hours. He coined this term, earned secrets, which is the more you work on a problem, the more nuanced you start to learn about it. And he also coined this term, the idea maze, which is sometimes there's these decade-long journeys of the founder just going out and figuring out little parts of the business and after 10 years they're by far more qualified than any other person planet to go after a specific problem set i can really agree with them i have this thesis on business building and building these great businesses and that it's really hundreds of small decisions dot compound on daily basis and oftentimes they're extremely unremarkable, but as they start to compound, they start to give you competitive advantage.

38:00Do you subscribe to this belief? That's probably true. And that there's like, I think especially today, there's a lot of tiny decisions from like who you hire for every single function. Like how do you think about your marketing best practices, your talent team, your engineering team, to what segments the market you go after, to which investors you pick, to how you think about pricing. All these tiny decisions do compound. And it's not like any one of them makes a difference on its own but every tiny tiny detail just really important and you're thinking about your job as a ceo you have to be a great architect of each of these tiny decisions and what you find is a lot of people have like really big picture visions and they often neglect the tiny details like what's the type of talent i want in every single function in my company and that costs them down the line and i think the best founders are remarkably good at appreciating those tiny details and getting all the tiny decisions right alongside the big ones and i think that's what creates like an enduring organization especially today like where the rate of software trends to zero it's really easy to kind of go and build decks it's really easy to kind of a lot of the execution pieces are getting increasingly automated so the thing that matters is like taste holistically and making the right decisions across 100 different things in a given year have you evolved where you spend your day-to-day the way i think about it is pretty different i think that at a given point in time in a company there is one bottleneck.

39:23And what you actually want to do is ignore it, spend 80 % of your time on unblocking that bottleneck and go all in on it and diagnose it and move on to the next thing and move on to the next thing and move on to the next thing. And the bottleneck always switches. It could be first, maybe your product, you need to work on your product a bit to get it to product market fit. Then it maybe becomes you need to iterate on your sales motion. Then it becomes you need to make a better deployment motion to handle that scale. Then it becomes you need a lot more leads and you need to get your demand generation right then it might become well you need to architect a talent org to go and scale that team again then it might go back to we need another product for act two and to expand the potential of the company and you keep going through these cycles and what you as ceo need to do is just go super deep on solving that bottleneck and just move on to the next thing the times i've seen myself not do well is when i try to do a lot versus when i just really isolate focus on the bottleneck and make it clear that this is the bottleneck that's what moves the business radically forward.

40:17How do you figure out what's the current bottleneck? If you look at your business and ask you, why are you not 100 times or 10 times bigger? There'll be one answer that's crystal clear. You just keep on asking this day after day? Every day. Every week, you should ask yourself, why are you not 10 times bigger? Then you'll find an answer. It's like, well, we need a lot more leads. It's like, well, go figure that out. Or it might be that, well, our reps are not converting fast enough. Why don't we enable them to do them better? Or it might be that, well, we don't have enough folks to go and deploy.

40:44Well, that means like we should go automate some parts of our deployment motion or go and hire more folks in deployment. So it's like whatever that bottleneck is, you need to go ask yourself, what is it? And go and solve it. Is that also how you figure out when to fundraise when capital becomes a bottleneck? My hot take in the modern world is that capital right now is kind of like either super available or it's not available. And I think the right way to think about if you're running your business well, there will always be capital. And if you're not running your business, there won't be capital.

41:13and I think it's one of those binaries that if you focus on building a great company, the capital will follow you. And do you think about making your business anti-fragile to capital itself? I do think about that, but I do think that like - Is that possible? It's possible, but I think if you want to win a market and you want to go super fast - There's trade-offs. There's trade-offs. Like you might need to blitz and take advantage of a point in time where you want to go and become the number one company really quickly. So having capital to accelerate is powerful. The core business on its own, like ours, could definitely be on a per-union economic basis, accretive, run profitably.

41:48But capital still could be a competitive advantage to be able to accelerate your roadmap, accelerate your international expansion, accelerate your ability to go to different verticals. I think about bottlenecks. I think about the Elon model where he'll go into one of his companies and for 14 straight hours, he meets with every single individual for five minutes discussing their one bottleneck. and what's underappreciated about this model outside of just having that the ability to have that many meetings in one day focusing on people's hardest problems is that the meeting itself is the competitive advantage meaning everybody is forced to distill their bottleneck to their level in other words the way that elon is always able to focus on the bottlenecks is he essentially crowdsources the bottlenecks from the employees.

42:37And those employees both have the carrot and the stick, the carrot being that they will have Elon's time for five minutes a day. And the stick being that if they come in unprepared, not thinking about their bottleneck and not understanding to the most granular level what it is that's keeping them from succeeding, then they may get fired on the spot. Here's a lot to learn from how Elon runs his companies. I think that the focus, the extreme maniacal to performance management, he's almost unemotional about it is something that 99.9 % of even good CEOs are not able to do and I think that's why he's exceptionally successful is because he can do a couple of things really well which is one be unemotional almost to a degree that kind of comes off as maybe a bit people might call him a lot of different terms that are that but in reality just an extreme psychopathic but that's just an extreme devotion to performance management the second thing is that he is able to cut through all the noise because there's lots of ways people can kind of go and you have like a vip manager level that sandbags a vp level that sandbags but he just kind of eliminates all the bullshit and keeps high performance culture and the third is at scale he manages intensities that very few startups have like even startups might not have the same intensities as large organizations and like that's impressive to do at scale and i think his style of management is like what hbs would call like oh this is like anti-management practices but I do think there's something true to that style of management that more CEOs need to emulate.

44:06It's been about a year since founder mode has come out. What do you think still rings true in terms of founder mode and do you apply founder mode to your business? What rings true about founder mode is that, like, you can't just be this person that sits, like, 200 feet above the ground or 2 ,000 feet above the ground and you're just, like, making these decisions and having your VPs report to you on progress. and you're like yeah yeah yeah let's just like well let me just let them run the company well i can do press i can do pr i can do all this stuff like that does not work because like fundamentally the main reason that a ceo's jobs exist is to drive performance of their organization and to drive intensity and by default like if you do not hold those standards and you don't keep those standards to an utmost high no matter even the best people need a little bit of pressure to kind of go ahead, right?

44:53You just need to like set the standards above. And I think that like is one thing that's true about Founder Mode is that you need to be fundamentally hands-on. You can't be hands-off. No great CEO has ever been hands-off. Like there's very, very, very few CEOs. I can even think of one that's been hands-off and successful. The second part I would say is that it's really hard to go and assess the people you manage unless you're involved in the details. Like how do you know that the marketing team is hitting their maximum efficiency if you have not seen the marketing plan and you're not paying close to how they're generating leads and you don't have an opinion on that how do you know your sales team is enabled if you're not watching calls and understanding what's happening at the ground level how do you know your product teams are shipping at a maximum cadence it's like you can't just let your team just go and go off and like you don't have any control you should have an opinion about is this good is this bad is this not otherwise there'll be too much drift the third thing to have an opinion on strategy you need to be involved in the details otherwise you're like strategy is frankly made based on like a bunch of like she said this he said that but so you need to be on the details so i think about if you're not involved in any details you're probably not running your company well if you're not setting those intense standards you probably aren't running your company well but let me copy out that with something i do think that a lot of people are using founder mode to justify micromanagement it's a difference micromanagement is i'm going to be involved in every single detail i'm not going to go hire good people i'm just going to go and fix these problems myself versus founder mode is go inspect figure out the problem is and figure out the systematic way to go fix it like you should set up systems that can make this company run on itself and you should go and solve them but if you're still the bottleneck for solving all problems with the company your business is fucked it goes back to what you're saying which you're constantly looking for the bottlenecks to solve but it would be the wrong strategy to sit with everybody and do all their work for them from day to day.

46:45That would be the most extreme is if you're doing routine tasks, you're not focused on the routine tasks, you're focused on these bottlenecks that are like a friction to the entire class. Yeah, if you're solving the same bottleneck again and again and again, that means like you're doing the solution and your team's not learning and you don't have a system to go and solve it. You need to put systems in place to go and solve it. Like maybe the leader's wrong. Maybe the system in place doesn't make sense and you need to work with your team to re-architect the system. Like whatever it is, figure out the structural solution versus don't figure out the band-aid.

47:15The backdrop behind Palette and any company that relies heavily on AI is that the foundation models are changing sometimes on a daily basis. How do you build a business that relies so heavily on foundation models? What you have to do is you have to make sure your product benefits from every sort of foundation model like evolution, right? So like the way we think about Palette is we have all this context about these businesses that's stored, right? For example, what are their SOPs for running quotes? What are their standard operating procedures for tracking order shipments? What are the ways that they handle orders from Walmart versus Target?

47:53So there's all this institutional context. There's an execution layer built on top of the models. So every model release should make our product better because we have our context, we have our execution layer, we have our data layer, we have our data integration layer. So in many ways, we're like a next generation system of record that contains context, but also does these actions. So every model release arguably makes our product more performant, but we're not at the risk of the models because they don't have all the institutional context, ingredients, the data that we have. So they can't come and displace us overnight.

48:26And I think that's the right way to architect a next generation AI company. And the wrong way to do it is to basically go and write all this deterministic code. And then when the model upgrades, all that code and that logic you wrote is kind of useless. you have to rely on the models to kind of go and do these computational intensive tasks. But you have to set up the architecture and the execution engine and the context layer that takes advantage of them. So another way, you're managing your information, not writing specific code for a specific model. You're more managing the data infrastructure so that when that new model comes out, it's not going to be obsolete in terms of what you could execute.

49:07Yes. So it always improves what we can execute. If you go back to the day before you started Palette, what would have surprised you the most about building an AI company? The rules of the game have changed in a lot of ways, and you have to rethink everything from first principles in terms of team building. So let me give you some counterintuitive takes. Prior world, everyone's like automated SDR outreach and marketing are automated. Let's go and have a bunch of outbound SDRs. Let's go have them call. let's go and use all these email prospecting solutions that's a new way of getting meetings but in a world where this is the cost of that incremental channel has gone down to zero what's happened is that there's been a return back to in person of like events and introductions and the old school tactics of sales and marketing are kind of back in play like people love having for a while like now more and more companies are doing like sports partnerships and lots of big aggressive brand marketing and trying to get their name out there and doing more of these old school brand building tactics and doing a lot of in-person events in a way that wasn't in the past secondly we shifted to this world of product management under the google and facebook paradigm where product managers became different like became these people that would go and do user research and go and identify functionality and go and build stuff out but in this modern world we do not have any product managers palette we have agent product managers who are like a hybrid of what was traditionally customer success in PM.

50:32So they work to deploy our agents in the field. And we have engineers. And what we find in that is that basically all the technical decisions are like which sort of models result in the best voice performance. What are the nuances of deploying with a certain freight forwarder? Like you don't need these like traditional spec style, workflow style product management in a way that you did in the past. So that function has kind to become obsolete. And then the third thing I would say is that you sort of don't have a good mental model anymore of how to build out your leadership bench. Like, for example, there's a lot of leaders who learned how to build a great SaaS company.

51:10That does not mean that they would be great at leading your AI deployments team, or they would be the best at leading your engineering team because they were trained in this old school paradigm. So now you have a choice of, do I bet on my hungry? The number of people that have experience with LLMs is like less than like under four years, right? There's not that many people with lots of experience. The 21-year-old might actually have more experience today than the 37-year-old that looks senior. So in that world, who do you promote into your leadership bench and what does the leadership bench of the future look like?

51:43So these are the kind of counterintuitive decisions on the people side you have to deal with in a way that I didn't think was true of the last generation. One of our portfolio companies, Lagora, they talk about how they prepare their teams that the entire code may need to be rewritten in a week or two weeks and making sure that they have this mental flexibility. And I think about oftentimes maybe that's why there's so many 20-year-olds and 22, 23-year-old engineers is there's something about the mental plasticity of somebody younger and the lack of ego that they have not yet developed that maybe a 30, a 35-year-old engineer might have been more ossified in the way that they approach software development.

52:25this is partially true but i do think that there's also value in those 30 year old engineers that are still mentally flexible and that have a good understanding on how to build scale systems and have good taste on system design so i think that what you're really looking for more than age or any factor is how much does this person approach problem solving from first principles the moment you hear someone in an interview be like i've done this before and i have experienced immediate at red flag that sounds that signals to me that like i think there was a case at palette where someone once told me that they have 20 years of experience and i was like i don't give a shit i don't give a shit that you have 20 years of experience in this field it's changed like what i care about is if you come to the right answer not the fact that you had 20 years of experience in the past and like you have to like value first principles thinking Zayder Amman, who's CEO of Flex, which is a newly unicorn company, he said that his biggest red flag on an interview is when someone says best practices.

53:30That's when he knows that somebody is not a first principle thinker. I mean, there are some cases of best practices. Like how do you deal with a client meeting or something like that? There probably is some set of best practices for things, but I generally agree with them. If you could go back and give yourself one timeless piece of advice before starting Palette, that would have dramatically helped you accelerate the success of the company. What would that be? Have more confidence in yourself. Like, I think the thing that I find that especially young founders do is that you think that you've hired these leaders.

54:06They have more experience than you in something. So maybe they know something you don't and your gut feels off about something. you probably might be right because you just have so much more context on your business and your market and your end state that like you should trust your own gut instinct more than you do and i think that sometimes you feel like is like do i know what i'm talking it's like you probably have a good sense on what's right and i think enough founders they don't trust their gut instincts and something feels wrong they need to like listen to that gut instinct more and be like yeah this is clearly not good enough i'm gonna push the matter i'm gonna push the matter i'm push the boundary.

54:38It's a knowledge paradox. The more you know about something, the more you realize that you don't know anything about that one topic. And the second order effect of that is that you're less confident. So the people going around with extremely high confidence tend to be more ignorant than some of the people that are going around with humility. There's a balance where you have to be reality, be a truth seeker and understand what's actually happening. But if you have the right piece of information and you're not deluding yourself, trust your opinion more. There's a reason you started this company and you got it here.

55:08You were right more often than wrong if you got it to a certain scale. Who do you go to for first principles thinking and to test whether you're on the right path? The best thing that I do for myself is rather than any one person is I play devil's advocate with myself and I'm like what if the opposite was true and I aggressively counter my own opinion and try to like play devil's advocate with the other argument and if i find that the devil's advocate perspective doesn't convince me that that's how i try to run it and run it because i find that when you go and you ask five people for their opinions three people agree with you two people disagree with you you're going to be like and they're all smart you're just going to be like back to confusion and analysis paralysis that i find that usually yeah like most people ask too many people for their opinions you know the charlie monger inversion yeah charlie monger when he approached a problem, he figured out that it was much easier to figure out how something would fail versus how it would succeed and then do the opposite of why it would fail.

56:06So you might look at Palette and you might say, well, what are the failure modes in Palette? Well, maybe we don't get customers. Maybe we aren't able to recruit. Maybe we're not able to scale. And then he inverted that to figure out what he needed to do. Somehow there's something about the human mind that makes it much easier to see why something would fail versus why it's going to succeed. Yes. Although I used to run my company a lot from a fear of failure, and I found that there are downsides to thinking a lot about failure. I think that it can take you pretty far, but the thing that gets you all the way there to the end is a desire to really win.

56:43How were you able to make that shift? at some point i realized because i was thinking about all the reasons things weren't going wrong i would worry about things like making tough calls like being like like oh my god we like what if we get this marketing experiment wrong or what if i just spent two hundred thousand dollars on this big call and it goes wrong at some point i'm like well if i don't do these things like yeah we're not going to fail but if i want to win i need to go big and i need to go make bold calls The way I reaffirmed it is to me, failing is not being number one. So when you rethink that, you're like, everything is going to be number one.

57:17You just almost created a new mental algorithm, which is like, failing is not actually failing. Failing is being number two in something. And when you realize that failing is being number two and winning is just like being the absolute best, you start to think of failure in a slightly different heuristic way, which is like, if I'm not the number one company, if we're not winning every single deal, if we're not in every major enterprise account, I failed. So like when our team shows me our pipeline, I'm like, what are the biggest customers that we're not talking to? Like who is not on this list and why are we not talking to them?

57:45It's more my default stream of thinking. So that way I'm always thinking about how do we maximize our winning ability? And you start to incorporate opportunity costs. In other words, not taking action is default failing. So you must try to take action so that you have a chance of not failing. My other hot take is that like most people think that you have to be 100 % right. I think that you should try to get the right decision and 30 % of the time and be 70 % right. As long as you're making 70 % of your decisions directionally right, and you're doing them really fast, you'll be a better CEO than most.

58:17I'm a big believer in that quality is downstream of quantity. So one of the ways to increase your chance of success is to do more and to do it quicker, get faster feedback loops. I've still struggled not to beat myself up when I do something, it doesn't work. I get the learning and I improve the system. And I'm trying to work on how do I maybe talk to myself in a way that's more productive versus criticizing myself for making a wrong attempt. One of the things I did for myself was Huberman is a big advocate of hypnosis. So I actually did a very active period of my life where I just did hypnotherapy.

58:57I tried to rewire my own subconscious patterns. And that actually helped me like kind of get rid of a lot of that negative dialogue in my head. The second thing is that when you beat yourself up, like the amount of period you like waste in like kind of that defeated like back and forth thought pattern is just causing you a loss of productivity that I view it as like, look, the longer I introspect and I keep thinking about the past, like the past is the past. We're not going to fucking fix it. Like we got to keep going forward and like otherwise we're still going to lose. If I keep thinking about the past, we're going to lose for sure.

59:26So I made the mistake. Let's just take the learning and let's just keep going. Otherwise we're going to lose. So like you have to always keep re-reforming it. Like what's the past is the past. We have to just keep moving on or we'll lose. I have to ask, how did you do hypnotherapy? I looked around and I found a guided hypnotherapist. And every single week it was like, how do I get over my fear of failure? I hate failing. So we just did a bunch of prescriptive sessions on trying to get rid of my fear of failure so that I could just make calls that I felt like were unpopular. Talk about first principles thinking.

59:56Yeah. Shant, this has been an absolute masterclass. Thanks so much for stopping by.