Qasar Younis on Why Physical AI, Not Software, Will Change Most People's Lives
Lenny's Podcast
Qasar Younis is co-founder and CEO of Applied Intuition. Host Lenny Rachitsky introduces the company as a $15 billion business that has grown quietly for close to a decade. It adds intelligence to vehicles such as cars, tractors, planes, submarines and mining rigs, and Rachitsky sums it up as roughly "Waymo or Tesla but without the hardware." According to the introduction, 18 of the top 20 automakers are customers, along with major construction, mining and trucking companies and the Department of Defense. Younis was born on a farm in Pakistan, grew up in Detroit, worked as an engineer at GM and Bosch, founded two earlier companies, and was COO of Y Combinator.
The conversation keeps returning to one argument. Younis believes the AI developments most visible in Silicon Valley, such as coding agents, chatbots and humanoid robot videos, will matter less to most people over the next decade than putting intelligence into existing machines on farms, in mines, on construction sites and on highways. He also covers fear of AI, the recent market sell-off, China, staying out of public view while building a company, company values, reading, and taste.
An Industrial Revolution Frame for Abundance
Asked what the future looks like if things go well, Younis compares AI to the Industrial Revolution. Someone in the late 1800s could list plenty of harms: child labor, emerging monopolies, abuse, wars. Yet today is nearly unimaginable without what that era produced, including broad access to health care, material goods, and things now taken for granted like heating and cooling a home.
He mentions a YouTube channel built around letters from German prisoners of war brought to America in the early 1940s. From their trains, the prisoners were amazed that every town they passed was lit up and that cars were everywhere. Younis states that 80% of German towns during World War II had no electricity. His point is that we tend to assume technology is evenly distributed when it is not.
The positive version of AI, in his view, is that things currently available to the wealthy or well-connected become available to everyone. His example is a personal coach tailored to you, as opposed to a chatbot giving generic answers. He also believes that solving "impossible problems like cancer" will be directly tied to this AI boom. His conclusion is that net suffering should go down significantly, as it did after the Industrial Revolution, and he calls himself a fundamental optimist on this.
For physical AI in particular, he points to mobility. People with their own car, working limbs and senses take the ability to drive to the store for granted. For someone disabled, or someone who can't afford a car, nearly free mobility is a big deal. If you live in Rwanda two hours from the nearest hospital, he says, free self-driving cars matter "in a very very true way." He argues that much of the negativity about AI comes from people "living in a very very good existence," while stressing that he is not naive about downsides.
His last example is personal. When his family came from Pakistan, long-distance calls were so expensive that they relied on handwritten letters. Now you can contact almost anyone on the planet basically for free. That has obvious downsides, but it also lets people stay in touch with those they love. He thinks AI can bring a similar kind of near-free abundance to many more people.
Fear Comes From Not Understanding
Rachitsky brings up anxiety about AI: jobs, robots, viral videos from China of humanoids wielding nunchucks, nervous markets. Younis first separates two things people tend to blur together. One is societal anxiety about a technical shift. The other is public investors reacting to particular stocks.
On anxiety, his thesis is that "the core root of fear is misunderstanding." If you are worried about what AI will do to your job, he says, the best response is to spend time understanding it, and you will quickly see its limits. He cites YouTube videos in which people try to get Gemini, and then ChatGPT, to understand a cup held upside down. If the revolution is coming, he jokes, the AI overlords first have to figure out the top and bottom of a cup.
He treats the nunchuck videos in a similar way. He isn't claiming they are fake. His description is that the humanoids are pre-programmed and that such a video cost $15 million to produce. Viewers, though, fill the gap with an impression of sentient beings acting on their own, when what they are seeing is motors programmed to do a specific thing. If you want to be impressed by robots, he suggests visiting a car factory, where very advanced robots have moved extremely fast for 25 years. Nobody feels anxious about those, he argues, because people understand them: a welding robot was programmed to make a weld. When people can't explain how a robot learned to do something, they substitute fear.
He grants the larger concern that technology can harm society. When WhatsApp emerged, for example, some companies disappeared and some people were hurt. He suggests dropping the word "AI," which he calls emotionally loaded, and just saying "technology." Technology can be used for good or bad, and his advice is to "get to know it, then actively make the technology be used for good," whether you're a founder or an employee at a large company.
Rachitsky notes that self-driving cars are essentially robots. Younis agrees and says that however you cut the statistics available from self-driving companies, they are far safer than human drivers. He predicts that within 20 to 30 years, people will look back on human driving the way we now look at child labor, which was once normal. People drove tired, impaired, extremely stressed, or in the middle of a traumatic experience. He says more than 30,000 people will die in U.S. road accidents in the next year and quotes the Stalin line that one death is a tragedy and a million is a statistic. Anyone who has talked with a family after a fatal accident, he says, stops worrying about AI robots and sees that human driving doesn't make sense, "not for any other reason than literally people die."
Rachitsky says he now uses Tesla's self-driving constantly after it improved a lot a few months earlier. Younis points out that driving isn't Rachitsky's job, and asks listeners to imagine a commercial truck driver or a mine worker getting "a helping hand" with a dangerous task. When self-driving trucks come up, people immediately ask about trucking jobs. Younis sets aside the fact that there aren't enough people who want those jobs, and focuses on the fact that people die in trucking accidents. He attributes the reflexive fear to a "monkey brain" shaped by hundreds of thousands of years in the wild, where a rustle in the bush meant a snake. On autonomous mining, he says those are awful jobs where people die, "and the best evidence is that people don't want to work in them."
Why the Market Sold Off
On the market side, Younis says openly that he has no specific research and is giving his guess. His MBA at Harvard exposed him to high finance, and it was also his first close look at extreme wealth, including private jets, which he found eye-opening coming from a modest background. From a distance he had assumed hedge fund and public equities investors made buy and sell decisions using deep, even theoretical, analysis. That isn't how it works, he says. They are smart and hardworking but lack the fundamental edge people assume, which he says is one reason retail investors have become such a significant part of the market.
His hypothesis for the recent sell-off in software stocks goes like this. Investors hire AI consultants or developers and ask them to build an app in a week. The consultants return something that looks like Figma or another web app. The real company would say it's only a surface imitation, without the depth or the integrations with other systems. The investor sees something built in weeks that took the company 500 engineers and a couple of years, reads posts on X about vibe coding your way to replacing billion-dollar companies, and the market prices in the risk. Younis adds that within the past 24 hours a "very calibrated" investor, whom he doesn't name, told him this is the time to buy because these companies aren't going away. The hedge fund manager, he jokes, isn't selling ServiceNow out of concern for society.
A Spectrum of Robots, and Why Hardware Timelines Are Hard to Call
Asked how close we are to robots being part of everyday life, Younis says to think of robots as a spectrum. Roombas are robots, and so is the machine that makes your coffee. The real question is how fast we move up the spectrum toward robots that handle many tasks with little guidance.
He offers a thought experiment. Imagine the conversation happening in 2006, the year before the iPhone, when everyone had flip phones. Asked when the "Star Trek phone" would arrive, he would have guessed one to five years. Within about five years, Uber, WhatsApp, Instagram and Snapchat were used by many millions of people. Nobody in 2006 could picture Instagram, he argues, because it depended on phones with app stores, cameras on both sides, general availability, and comfort with social networks that didn't exist yet. It was before Twitter, Facebook was still small, and MySpace was a different kind of community. So change can come quickly, but the form it takes is hard to predict, because intelligence built into generally available hardware is what shapes the use cases.
He expects the earliest use cases to be the ones with the most "bang for buck": a car that drives itself, or a mining vehicle that becomes intelligent. The engineering to build a giant machine that moves dirt was done over the past 50 or 60 years, so adding "a little bit of intelligence" builds on all of that. He acknowledges he runs a physical AI company and says he isn't "just pitching my own book." Emotionally, he says, we love humanoids because we're monkeys, but practically, the big change comes from putting intelligence into things that already exist around us. New applications will follow once that happens.
The Next Five to Seven Years in Cars, and Beyond
Younis then lays out what he expects. He says every car company in the world is working on something like Tesla FSD, without exception, and many are working on versions that will become fully autonomous with a cheap sensor suite.
He simplifies, and says so, into two approaches. The Waymo approach uses many sensors, heavy compute and maps. The Tesla approach uses few sensors, no high-fidelity maps, and cheaper compute. The Tesla-style product, called "L2++" in the industry, will be available everywhere because it is cheaper and doesn't need HD maps. The Waymo-style product works better within a geographically constrained area. In five years, he expects both L2++ and L4 to be much more common, and not only in the Bay Area and parts of China but globally.
His analogy is navigation systems, which people once paid thousands of dollars for. He says L2++ is at that stage now, with people willing to pay thousands for a semi-automated car. He expects it to become close to free fairly soon and says this downward price pressure is already visible in China. Five to seven years out, he expects every car to have some level of autonomy, with buyers getting FSD as standard. Then the change becomes quiet and taken for granted, the way CarPlay and Android Auto brought free navigation and apps to cars without feeling like a revolution, though he considers that shift huge. He expects full autonomy to follow as something everyone assumes, with a clear drop in injuries and deaths.
He stresses that the consumer car is only the easiest example. The same pattern applies in construction, mining and defense, where humans work alongside large physical machines. Treating the machine "not like a sentient being but almost like a physical agent" of what you're trying to accomplish would unlock things that are hard to picture now.
He says he loves Moltbook and what he calls "the OpenClaw revolution," and the hosts joke about lobster claws on set. Still, he thinks that world reaches only a small slice of society, mostly developers. His "barometer of impact" is sitting at a gate in the Detroit airport and asking how many people there have heard of OpenClaw. For them, he says, AI's impact will show up in the physical world. He concludes that the real impact of AI over the next 5 to 10 years will be in farming, mining, construction and self-driving trucks.
AI Arriving Just in Time for Labor Shortages
Rachitsky raises a theme from earlier episodes, including one with Marc Andreessen: AI may arrive just in time, as populations shrink and age. Younis agrees and says these industries need autonomy and "it couldn't come soon enough." He says the average farmer is in their late 50s, around 58, and asks what that means in 10 years, when many will have retired, and in 20 years, when the problem is bigger. He says every vertical looks like this.
He offers a hypothesis for why McDonald's or a local quarry says it can't hire even though the people haven't gone anywhere: the tradeoff is no longer worth it. In the 1980s and 1990s, long-haul trucking meant a working-class family accepted that a parent would be away for days or weeks. Today the same family can choose Uber or DoorDash, because you can turn off the app and pick up your kid. He sees intelligence in the physical world filling those gaps, not whole industries disappearing overnight. He doesn't believe in a near-term future where entire industries are fully automated, because replacing an industry with robots is still too complex. It may happen one day, he says, but society will be very different by then.
For people outside the AI world who feel anxious, he recommends history books about the Industrial Revolution. Unlike earlier eras, many people were reading and writing during it, so you can read accounts from both those harmed and those who benefited. Taken as a whole, he says, it was very positive, though the downsides should be mitigated.
What he argues against is the urge to "pump the brakes" on technology. His view is that this would make the American economy stutter, and that would hurt the lowest end of the labor market most, so efforts to protect the most marginalized would end up harming them. He points to the gap between Europe and America over the past decade and says American growth has come from Mountain View, Sunnyvale and the Bay Area, meaning frontier technology, not Detroit. He does say a future that doesn't take care of the average American worker would bring much bigger problems, so any solution has to account for that. It just shouldn't be "pump the brakes."
China: Comparing Companies to a State Is a Category Error
Rachitsky notes that fear of China is one reason people oppose slowing down, and asks about Younis's contrarian view. Younis bases it partly on House of Huawei, which the company read together. He calls Huawei an amazing company that makes great technology. Of its couple hundred thousand employees, he says about a quarter are Communist Party members, and its aim isn't profit or shareholder returns. It is a private company but functions as an extension of the state, and he says its name means "China's ambition." He compares that to a company named "MAGA" with a quarter of its staff from one political party and a mission of expanding that movement, and says at that point it isn't really a company.
His first point is that Americans project their understanding of markets and companies onto China, assuming Huawei must be like Apple because both make phones. He suggests reframing it as OpenAI competing with the Chinese government instead of with DeepSeek, and Apple competing with the Chinese government instead of with Huawei. "Chinese government" may not be the best label, he says, but these aren't independent, privately owned, for-profit groups building products for a market.
His second point is about sustainability, and he uses EVs as the example. Chinese EVs are widely praised, followed by what he calls fairly shallow analysis about how bad Munich, Detroit, Tokyo and Seoul supposedly are. He calls Rivian a Chinese-EV-like company in America that makes great products but loses a lot of money doing it, so it isn't highly valued. He doubts it would make a list of the top 50 or 100 Bay Area companies, and says that's not because the products are bad or the people incompetent but because the EV business is hard. The home team is judged as a business and punished by public investors when it doesn't make a profit, while the Chinese players aren't really companies. If America combined Tesla and everyone else and ignored profit, he thinks it would produce some very impressive products.
On the philosophical question of whether one side has to fail for the other to succeed, he says open, free markets let everyone succeed, as a century of evidence shows. The current difficulty is how China fits in, since its markets aren't open and free. That doesn't require hostility, doesn't mean China is incompetent, and doesn't mean it isn't worth serious attention, but it also isn't a one-to-one comparison. He adds that a nuanced five-minute explanation like this will never reach the person waiting at the Detroit airport, who only hears "China bad."
Why Younis Stayed Silent for a Decade, and Why Most Founders Shouldn't Copy It
Rachitsky describes Younis's recent first tweet, a simple note saying they would start tweeting. It got around 2 million views and a reply from Elon Musk. Marc Andreessen quote-tweeted it as "the best AI CEO nobody knows," and Elad Gil called Applied Intuition the most successful quiet company in AI. That runs against the usual advice to build in public.
Younis says the quiet was deliberate and that, left to himself, they'd do it forever. The company takes more inspiration from Berkshire Hathaway than from Silicon Valley favorites. But he immediately cautions founders against copying this. He could stay quiet because people in the ecosystem already knew him personally, so he didn't need a public brand to stay on Elad's radar. His first two companies were much less known, before YC. The company's values, he says, come down to two words, "radical pragmatism," so check whether advice fits your situation before following it. He cites investor and friend Naval Ravikant's view that fame is a powerful tool. If you don't have a network, a following is a great way to recruit employees, investors and customers.
The rational case for staying quiet is focus. Every minute spent on a podcast, a post on X, or anything written for the public is a minute not spent on customers and product, and those are what produce results. He also offers some armchair psychology: people act first and rationalize later, and neither he nor co-founder Peter Ludwig gets much emotional satisfaction from public activity. Growing up as a Pakistani immigrant in Warren, Michigan, "with a little bit of a weird name," made them skeptical of the mainstream. They think many founders share that outsider feeling, which became hard to square with being COO of YC, about as inside a position as exists.
Andreessen, who pushed them to go online, and Gil told them to set that baggage aside and be pragmatic: public communication gets the message out. That message can be narrow, about what's happening in physical AI, or broad, about how society is changing. Younis says they have seen both extremes of the economic spectrum, and Andreessen argued those ideas were worth sharing beyond promoting the company. Debating ideas about how technology is changing society is something Younis says they can get behind. Later in the conversation, they describe it as a responsibility that grows with the company's influence: sharing ideas first with AI founders, then technology leaders, then the wider world. They note that VCs used to be anonymous "blobs of ominous sounding" firms, and that a16z and a few others, including John Doerr, created the model of the individual investor spreading a set of ideas.
Early Traction and When to Hard Reset
From watching hundreds or thousands of companies at YC, Younis says good companies tend to show traction fairly early and then sustain it for a decade or more. For founders about two years in and struggling to raise money or build something customers love, measured in attention or dollars, the heuristic is this: if what the market tells you isn't pointing you toward a more and more specific path, consider resetting.
Often, he says, the foundation is the problem. He compares it to a house where a cup of water keeps sliding off the table, and you keep adjusting the table when the whole house is off-kilter. The foundation could be the co-founders, the market, your stage of life, or how much effort you're willing to put in. You may not know which one it is and still have to hard reset.
He tells Applied Intuition employees who leave to start companies, and he says there's a growing founder alumni group, to treat their first three years as a zero and drop the expectation of success. Being a founder is a craft. If your first woodworking table wobbles, nobody tells you to go work at Crate & Barrel. They say it's your first table. Founders who expect greatness right away miss the real payoff of a first attempt, which is learning and building the skill for the second and third. He doesn't think it's random that his third company is the most successful, and notes that some funds focus almost only on repeat founders.
He also acknowledges why this advice is hard to hear when you're in "the proverbial war." Founders have to hold two conflicting ideas: the company is deeply important and deserves everything, and it also isn't that important. Staying out of public view helps here. When he became COO of YC, he asked president Sam Altman not to announce it for a year, so that if the partners didn't want him in the role, he wouldn't face public questions about serving only six months. Pivoting is easy before you raise money and hire people. After you have ten employees who joined for a specific mission, telling them "turns out this is wrong" is like ordering troops off a hill they were just taking, and you lose credibility. When you're very public, the startup also becomes your identity. That's why an early company value read "our best work is done alone and quietly," which Younis says they believe for practical reasons and not because being under the radar is cool.
Applied Intuition's Values
Younis calls "radical pragmatism" and "our best work is done alone and quietly" meta-values that sit above more specific operating principles. The tactical advice is to set values once you have a bit of traction, and not to invent them "as philosophers." Write down the five to ten real reasons you're succeeding, and those become your values.
Applied Intuition's list begins with "speed above everything." Younis later gives the exact wording as "move fast, move safe" and says the behavior evaluated under it is decisiveness. Others include never disappoint the customer, technical mastery, and high output matters. Less obvious ones include "laugh a lot," which has been there since the beginning. On intense work, humor keeps people grounded and gives them perspective, and it allows softer feedback, "it's not the best" instead of "this sucks." Another value is "half the work is follow-up": taking notes and following up "is the business." Younis and Ludwig meet every week with all new team members and mostly talk through the values in depth.
When Rachitsky mentions the Stripe Press book Maintenance, Younis ties it to their roots in automotive engineering, which they call an exercise in quality. People call rockets complex, but even at high capacity a rocket launches every couple of days, while a car comes off the line every 30 seconds, has to be very cheap, and competes globally. A factory runs on safety and maintenance, not on anything especially complex. Operational strength means watching a handful of things and doing them very well. Younis cites the saying that a man who cannot command himself is not fit to command others, and argues that when you maintain yourself, your team and the company, good products come out of that system. They encourage founders to see a company as a machine and tune it, like car enthusiasts who detail under the driver's seat. Younis details their own cars.
Rachitsky mentions The Score Takes Care of Itself. Younis says Joe Montana is an investor, and the company's Series D announcement was titled "the valuation takes care of itself."
A Self-Cleaned Office and No Raised Capital Spent
Visitors often assume the clean office means a large cleaning staff. Younis says employees clean it themselves in a weekly "cleaning zen" where everyone tidies their own area, modeled on Japanese schools where students clean. Younis lived and worked in Japan, which also inspired the no-shoes policy. There are still cleaners, but employees are expected to notice their surroundings, and not get so caught up in being "a Stanford software engineer" doing AI that they won't clean their desk.
Younis then makes what they call the company's "claim to fame," which they admit "almost sounds like it's made up": in nearly ten years and with over a thousand engineers, Applied Intuition has never spent any of the money it has raised and operates as a functioning business. Younis thinks this is somehow connected to cleaning the office and sees "a direct line" between working quietly and alone, keeping your desk clean, and writing good software, but says plainly they don't know what that connection is.
They also encourage founders to take the best of Japan, Germany, China, Detroit and Silicon Valley. They interpret Steve Jobs's "great artists steal" line as a less generous way of saying: be humble, learn from everything around you, and be well-rounded.
Read Old Books, Especially in Areas You Know Nothing About
Younis quotes Charlie Munger's line that he never met a highly successful person who didn't read all the time. Their rule is to read old books, because time has filtered out the noise. A person might read 50 to 100 books in a lifetime in the best case, they say, so don't waste any on low-quality material. Read the pillars of human thought and work out how they apply to your business or technology. They "absolutely" believe reading The Autobiography of Malcolm X makes you a better founder, while admitting it isn't a simple if-then. Understanding society and history somehow leads to better products, and they say they can't explain how. Rachitsky compares this to LLMs, where lots of data leads to something nobody fully understands.
Younis says they deliberately mention unexpected books because everyone already knows High Output Management. They are currently reading a recently published vibe coding book that the whole company is reading, which they admit breaks their own rule, and are nearly finished with The Emperor of All Maladies, which they say changed how they think. They consider that the real test of any material. The more obvious recommendations are Sam Walton's Made in America, which they note Walton wrote near the end of his life, Colin Powell's My American Journey, and Jared Diamond's Guns, Germs, and Steel and Collapse. Their method for picking a book is to find an area they know little about, such as Roman history, which led to SPQR, or philosophy, Jainism, or feudal Japan, and read the best book on it. Rachitsky contrasts this with Andreessen's "barbell" of today's news and books from ten years ago.
Listening to Naysayers, Then Deciding Fast
Rachitsky asks about Younis's practice of encouraging leaders to listen to dissent. Younis describes the ideal as a company where competing ideas get tested without emotion and the best idea wins regardless of who proposed it. They say founders are often the obstacle, because media and tribal instinct tell them strong leaders hold firm views and weak leaders get pushback. Applied Intuition takes input from customers, employees, competitors, investors and society, and Younis credits that as one reason for its success. The instruction isn't to disagree with everything, which would be contrarianism for its own sake. It is to "speak up." Someone who worked at Zoox, Waymo, Tesla or a Chinese company may have exactly the relevant experience when the team is debating something unfamiliar, like literal space, and they need to feel able to say it even if they're the most junior person or lost the last debate. Younis quotes Gandhi's line about truth standing the test of time and says "best idea" may be more accurate than the Bay Area term "truth-seeking." Once the best idea is found, you commit fully.
As a counterexample, Younis describes watching from inside Google as Facebook rose. They describe Google in the late 2000s and early 2010s as Silicon Valley's "apex predator," at a time when the MacBook Air had just launched, AWS was young, Nvidia was near bankruptcy, Microsoft was run by Steve Ballmer, and Twitter was small. Google had the best engineers and, in Younis's account, about a billion dollars a month in cash flow, and was 15 to 20 times larger than Facebook, which had maybe a thousand people. It still couldn't compete, because "how does a gorilla learn how to fly? By not being a gorilla." Google could only have won social by being a social company. A company's momentum and culture create a "wall of sound" that drowns out signals that the market is changing. Younis says small companies can suffer the same thing when a founder's view is ten degrees off: right market, maybe even the right problem, but the company still goes astray. "We're so scared of failing" that they listen when someone says the company is five degrees off course, and once that becomes the culture, they argue, it's hard to lose.
Asked how a company that's bad at this can put it into practice, Younis says the answer is a middle path between charging ahead and endless uncertainty. After the debate, you walk the chosen path with confidence. The company evaluates, pays and promotes managers against its values, including the decisiveness behind "move fast, move safe," so openness and quick decisions are built into the system on purpose even though they pull against each other. Balancing them is why a founder "get[s] paid the big bucks." At some point, often sooner than you'd expect, no more information is coming and you have to decide.
Taking Emotion Out of Decisions
Rachitsky describes a technique of asking what you would decide if nobody's feelings would be hurt, and then figuring out how to handle the fallout. Younis calls that the obvious version and goes further. Being attached to an idea because it's yours is itself emotional, tied to ownership. They describe emotions as reactions and frameworks built from life experience that weren't optimized for, say, a product review. A useful heuristic is whether different people in the company would reach the same decision. You want the "raw image" without the emotional filter so decisions can be classified consistently.
Rachitsky mentions Michael Pollan's new book, which argues emotion, not more intelligence, gave rise to consciousness. Younis says it underestimates human thought to reduce it to associations among ideas, facts and words, and quotes a saying that technical mastery is mastering the complex, which computers do well, while creativity is mastering the simple. They add they might "eat my words" if the best artist in three years is an AI. They return to their approach: take in broad inputs, remove the filter, build a culture that gives no weight to where an idea came from, and then decide as the leader. "And by the way, you got to be right." Founders like to take credit, but starting a company or having a vision isn't enough. The evidence is whether the company becomes a sustainable standalone business. Younis adds that all their advice is aimed specifically at founders of venture-backed AI companies in the Bay Area.
Taste Comes From Wide Exposure, Including Life at the Bottom
In the last topic, Younis says many Silicon Valley CEOs lack great taste, both artistically and in running a company, down to HR policy details, and immediately says they don't claim to have it either. The main cause, they believe, is limited exposure. Their example, not aimed at anyone in particular, is someone who grows up in Cupertino, attends Berkeley, starts a company right after school, and has never been an employee.
Younis spent more than a decade at organizations with over 100,000 employees, such as General Motors and Bosch. Deep inside those companies, you learn how bad it can be to be an employee: bureaucracy above you, leaders who don't know what's going on, outdated tools. That experience matters when you later set policy and build culture as a leader, and many founders "never had the frankly the fortune of being at the bottom of the totem pole."
Taste, as Younis defines it, means understanding people and life and being able to judge what's good. One heuristic is whether a founder would be good at many things beyond running a Bay Area AI company. They think someone who has backpacked around the world for a few years will somehow be a better founder, and admit there's no peer-reviewed research to support that.
Asked how listeners can help, Younis asks for book recommendations they haven't come across, since some favorites came from random suggestions, along with unusual research. They also want opinions, new or old but from a different angle, on how AI will affect mines, farms, construction sites and robotaxis, and they are now reachable on X.
You decided to join Twitter recently, put out your first tweet. Marc Andreessen quote tweeted it and said, "This is the best AI CEO nobody knows."
Our best work is done alone and quietly. Every minute you're writing something for public consumption, you're not focusing your very limited time that you have on your customers and your product.
You're building a lot of the future that we're going to be living in. What does the next couple years look like?
Us solving some of these impossible problems like cancer are directly going to be related to this AI boom. Net suffering in humanity overall should go down significantly.
A thread that has emerged on this podcast is that AI is coming just in time to save us.
The real impact of AI in the next 5 to 10 years really is going to be in farming, mining, construction. These industries, they need autonomy and it couldn't come soon enough. If you look at farmers, the average age of a farmer is in their late 50s. What does that mean in 10 years from now?
There's a lot of anxiety about what AI is going to do to the world.
The core root of fear is misunderstanding. If you at home are very anxious about AI, the best thing that you can do is spend time to understand and you will quickly see the limitations. Get to know it, then actively make the technology be used for good.
Today my guest is Qasar Younis, co-founder and CEO of Applied Intuition. You have probably never heard of Qasar or Applied Intuition. This is the most important under-the-radar AI company and CEO that I've ever come across. It's a $15 billion company that has been growing quietly over the last decade. What they do is they add AI to vehicles like cars, tractors, planes, submarines, mining rigs, and a lot more. 18 out of the top 20 automakers are customers, as well as the biggest global construction, mining, and trucking companies. Also, the Department of Defense. They're basically Waymo or Tesla but without the hardware. Qasar himself was born on a farm in Pakistan, grew up in Detroit, started his career as an engineer at GM and then at Bosch. He then went on to start a couple companies before starting Applied Intuition. I love everything about this episode and I'm so excited to bring it to you. Don't forget to check out lennisproduct.com for an incredible set of deals available exclusively to Lenny's Newsletter subscribers. Let's get into it after a short word from our wonderful sponsors.
This episode is brought to you by Omni. Many product teams today are in the process of debating how to ship AI analytics. The hard part is obvious. Having an LLM guess at SQL in production is a huge mess and just a bad idea. Omni takes a different approach. They have a semantic layer built in so that when you embed their analytics, the AI actually knows your business definitions, not just your raw tables. You can test queries, validate the reasoning, and lock down permissions before anything hits production. If you want AI analytics in your product without building the whole stack from scratch, check out omni.co/lenny for a free 3-week trial. Companies like Perplexity, dbt, and BuzzFeed use Omni to ship analytics their customers can trust. That's omni.co/lenny.
My podcast guests and I love talking about craft and taste and agency and product market fit. You know what we don't love talking about? SOC 2. That's where Vanta comes in. Vanta helps companies of all sizes get compliant fast and stay that way with industry-leading AI automation and continuous monitoring. Whether you're a startup tackling your first SOC 2 or ISO 27001 or an enterprise managing vendor risk, Vanta's trust management platform makes it quicker, easier, and more scalable. Vanta also helps you complete security questionnaires up to five times faster so that you can win bigger deals sooner. The result, according to a recent IDC study, Vanta customers slashed over $500,000 a year and are three times more productive. Establishing trust isn't optional. Vanta makes it automatic. Get $1,000 off at vanta.com/lenny.
Qasar, thank you so much for being here. Welcome to the podcast.
Thanks for having me.
You're basically building a lot of the future that we're going to be living in. And people may not even realize this. And there's two sides to this. On the one side, let me ask you this question. If things go really well, what does the next couple years look like for people with the emergence of AI, with physical AI? What's a vision of the future?
Let me take the broader AI point and then the micro, the more specific one on physical AI. Macro, I think about this like the industrial revolution, right? So if you're sitting, let's say, in the late 1800s, we can focus on a lot of bad things that happened because of the industrial revolution, right? You have child labor and you have monopolies emerging and you have abuse of, you know, wars end up happening. But it's also an almost unimaginable present without the values and without the kind of benefits we got out of the industrial revolution, which is broader access to health care like we've never seen before. Like access to goods, material goods, things we take for granted like heating and cooling your home.
There's this great YouTube channel that focuses on POW letters from Germans who are seeing America in the early 40s, and they're writing letters back to Germany about what they're seeing as they're basically, you know, prisoners of war, and they're kind of blown away that the towns that they roll by in these trains as they're going to their POW camps are all lit up, or that there's cars everywhere. 80% of German towns in World War II did not have electricity, and that's kind of a mind-bending thing because we just assume all this technology is, you know, equally distributed.
So the positive version is these things that, let's say, folks who are wealthy or folks who have access to technology have, everybody has access to. The fact like simply having somebody who's a coach to you, and having that coach very specific to you, not a generic ChatGPT that's giving fairly generic answers. I think that's a very powerful thing. I think us solving some of these impossible problems like cancer are directly going to be related to this AI boom. So I think net suffering in humanity, just like the industrial revolution, overall should go down and should go down significantly, and I'm a fundamental optimist in that view that technology will bring that positivity.
In physical AI specifically, again, when you have your own car and you have your limbs and you have your senses and you can drive, you take these kind of things for granted. You jump in your car and you go to the store. For somebody who maybe is disabled, or somebody who doesn't have the money to afford a vehicle, access to mobility that's nearly free or is free is a big deal. And that simple example of making self-driving cars free for everybody and how that would change the planet. You live in Rwanda and you are two hours from the nearest hospital. That matters in a very, very true way.
And so I think a lot of, let's say, the negativity around AI comes from people who, frankly speaking, are living in a very, very good existence. And when you live on the other edge of society... Yes, and I'm not some naive person who thinks that there's no downsides of technology. We can discuss that. But I just see there's a lot more positive.
So when you ask that question, what's the next, forget 3 to 5 years, what's the next 20 years? These things that we take for granted that are bad suddenly are not there. And I think certain diseases, certain inaccessibility to basic services suddenly start going away.
One last example is you take just the fact that you can message people basically for free. For people old enough, this was not the norm. We came from Pakistan. We couldn't even communicate back to Pakistan because the long distance was so expensive, and so it was handwritten letters. Today, you can basically contact anybody on the planet basically for free. There's obvious downsides to that. But there's lots of upsides to that, which is being in touch with people that you care about and you love basically for free. And so I think AI has the ability to bring this abundance to many, many more people at a near-free cost.
On the flip side of this, as you pointed out, there's a lot of anxiety about what AI is going to do to the world and jobs, robots. There's these videos coming out of China with these robots with nunchucks, like the stock market.
I feel, you know what, I feel the nunchuck union is up in arms with that.
How dare they. Yeah. But it's scary, and the markets are reacting more and more to just like, oh wow, these companies are maybe not going to survive long term. Again, being at the center of this and building a lot of the stuff that'll get us there, how do you envision the next couple years playing out? Are you optimistic? What keeps you optimistic? Any advice to people to help them kind of stay calm through this period?
So those are two separate things: anxiety around a technical shift, and then the public investors reacting to specific stocks they held. We have to separate those things. So let's talk about them separately.
The first one, the core root of fear is misunderstanding. And I think if you at home are very anxious about the impact of AI in some variant on your own job, the best thing that you can do is spend time to understand it and you will quickly see the limitations. There's some great videos on YouTube which are like trying to get Gemini to understand what a cup is by just holding it upside down, and then they do it with ChatGPT. So it's like, if the revolution is coming, the AI overlords have to first understand the top and bottom of a cup. And so you realize that you can see the video of nunchuck-wielding humanoids which are pre-programmed and that cost $15 million to do that video. Yeah, that is true. It's not fake. I'm not implying it's fake, but it's also not what your brain kind of fills in the gaps with. You see nunchuck robots and you just feel like, oh, these are sentient beings that are going at their own volition, rather than it's a bunch of motors that have been programmed to do a certain thing.
If you really want to be impressed, you go to a car factory, and we've been doing that for 25 years, where you have very, very advanced robots moving extremely fast to build things. And why don't we have anxiety about the car factory but we have anxiety about the nunchuck robots? It's because the human being doesn't like that gap. We understand the gap of a welding robot. You see, oh, okay, that's a robot. It's been programmed to make this weld. But as just an individual human being living in the world, you don't know how that robot was made to do that nunchuck thing. And so you substitute that with anxiety and fear. And so I would really implore you to learn more about the technology, and you start seeing the edges.
Now, does that take away from the most fundamental thing that you're getting at, the string that you're pulling at, which is, is society going to be fundamentally harmed, and is this net-net bad for society? I think in any technical shift, the emergence of WhatsApp just as an example, there are people who are damaged by that. Literally companies that go away, but also humans who are damaged by the advent of that technology. And so I think as members of society and as leaders in society, we can kind of move that funnel in whichever way. Remove the word AI. AI is such an emotional word because it's wrapped in these things you don't know, and so that fear then kind of deforms. So let's just say technology. So I think it's up to us to recognize this technology can be used for good and technology can be used for bad, and I think that's where really the focus is. So get to know it and then actively make the technology be used for good as a participant, whether you're a founder, all the way to an individual employee, citizen of a large company.
Then on the second part of the question, about public investors and stuff, I don't have any particular research on this, but this is my guess of what's actually happened. Beyond being an engineer, which is maybe my core identity for lack of a better word, I also did an MBA at Harvard. I didn't come from a very wealthy upbringing, and this was the first time, when I went to Harvard, I saw that world, that world of people having private jets and stuff. It was a really eye-opening experience for me.
But the real world I was exposed to was high finance and how high finance works. And you might think, as I did from far away, that folks at hedge funds or at large public equities funds are extremely nuanced and thoughtful and they are on whiteboards with extremely deep and maybe even theoretical math to figure out should they buy or sell Figma. And that's not actually how it works. Really what these folks are, and in this specific case I think what's happening is, they buy and sell stock. They are smart people and they do work hard, not to take that away, but they don't have a fundamental edge that you would assume that somebody who sits in a skyscraper in New York has. And by the way, that's why retail investors have become such an active and significant part of the market.
So those folks have gone to AI consultants and have gone to people who are literally developers at these firms, and they'll do something like, hey, why don't you build me this app in a week? And then this consultancy will come back with an app which kind of looks like maybe a Figma or some other web app. And if the company was sitting there, they would say, no, no, this just looks like my app, but this is actually not my app. It's not as deep. It doesn't have all these things. There's integrations with all these other systems. But the public investor, they're like, yeah, but it only took a few weeks or a month to build this. It took you 500 engineers a couple of years. This AI thing could be real, and the things I'm reading on X about just vibe coding your way to replace billion-dollar companies, that might be the case. And the market immediately prices in that risk, and that's where that selloff comes from.
That doesn't necessarily mean all of those... I mean, just within the last 24 hours, I can't say his name, but a very, let's say, calibrated investor said this is the time to buy because these companies are not actually going away. And so I think anxiety within society and the selloff are two very different things. They're motivated by different things. They're part of the larger AI narrative, but I wouldn't conflate those two things. It's not that the hedge fund investor is like, I'm worried about society, sell ServiceNow. It's different than that. At least that's my impression.
This is the alpha we've been talking about. Time to buy. This is not investment advice. Well, that's really good advice.
I think the real advice is to fight fear. And I feel that anxiety,
Especially when I go to Michigan and, you know, outside of people in the Silicon Valley bubble, it's like just try to learn a little bit about the technology that you're afraid of and you'll start seeing some of the edges.
I love your point about how self-driving cars are essentially robots. We don't call them that, but they're robots. Absolutely. And you see a nunchuck-wielding robot, a self-driving car doing bad things could be very dangerous already. And so that's a really good reframe that if you just think of it, it's just another robot.
And it's been really good for us. And by the way, the self-driving thing as an example, you know, whichever way you slice the statistics that are available from self-driving companies, they're supremely, supremely more safe than human drivers. And I do believe in 20 or 30 years, not that much longer, we'll look back and we'll kind of be like, it's kind of like we think about child labor. You know, in post-industrial revolution, that was a normal thing. You would send kids who were in middle school to go work. It happens in third world countries today. There isn't a lot of emotion behind it. It is not considered to be exploitative because you have no choice. You know, everyone's just, and I think we'll look back in 25, 30 years like people were just like tired, under the influence, you know, extremely stressed, going through a traumatic life experience, and then they jump in to [laughter] a car like that. That is crazy.
And just for everyone, everyone should really emotionally think about just in the United States, over 30,000 people will die in the next year from these accidents. Like the old Stalin line, it's like, you know, one death is a tragedy, a million is a statistic. And we just let the statistic kind of go over our head, like, oh, it's 30,000 people. But if you ever have talked to a family of somebody who went through a tragedy like a car accident, it's unbelievable. And you suddenly, all the fear of AI robots goes away and you really see that human impact and you realize, like, actually us driving doesn't make sense. And it's not for any other reason than literally people die.
I've become a huge, I have a Tesla and I just use self-driving all the time now. Just like a few months ago, it got very good. And it used to be nerve-wracking and now it's like, wow, this is much better.
And you're not doing driving as a job. Imagine if you're a commercial truck driver or you work in a mine or you work, you know, like, a little bit of intelligence, a helping hand in that very dangerous task. It's incredible. And I think there's something about the human brain where, you know, when you bring up that reality of like self-driving trucks, immediately people are like, well, what about the trucking jobs? Now, needless to say, we don't have enough people who want to do that job. So, leave that fact to the side. I think the fact that you really focus on is the fact that people die from trucking accidents. Like, we can't, you know, throw out the baby with the bathwater.
And so I implore everybody, you know, who thinks about AI broadly and physical AI specifically to always recognize that your monkey brain is programmed because of thousands of years of being in caves, you know, hundreds of thousands of years of living out in the wild and being in the cave, that when you hear the rustle in the bush, you think it's a snake because that's what our ancestors were programmed.
Now, when something new enters our psyche, your view isn't, well, if mines became autonomous, well, wouldn't that lose jobs? It's like those are awful jobs that people die in. And the best evidence is that people don't want to work in them. Like, that's the best evidence. Like, nobody's clamoring to go work in a mine in a remote area. And so intelligence can help make that reality much, much better.
People are seeing AI advance in all these different ways on the software side. They see all these models being released. It's driving 100% of people's code. Now, what's really cool about you is you see the hardware side of this. And I think one of the biggest changes to our lives will probably be robots walking around doing things for us. Do you have a sense of just how close we are to just robots around us day-to-day?
So I would think about the framing here matters, again, on a spectrum. So there are robots around us, like Roombas, you know, like the ones that clean your carpet while you're sleeping. There's robots around you when you make a coffee. That's an automated machine that is taking an input and doing a bunch of things based on, you know, what you need. So what you're really talking about is how fast can we go up that spectrum to where you have a robot that can take on lots of tasks with little guidance.
And the way that I would think about this is, let's say this podcast is happening not in 2026 but 2006. And you're asking me the same question about mobile and you say, well, mobile is coming. This is, remember, pre-iPhone, which comes out in '07. Everyone has got those flip phones. So we have mobile. It's not like a completely, you know, so we have some robots around us already, but it's like, okay, so when are we going to, and you asked me in 2006, when are we going to get that Star Trek phone that can do everything? And I think at that time, I would say, because I don't even know the iPhone is coming a year later, I would say, well, Lenny, I don't know. Maybe it's one to five years. And it's not 5 years later that Uber, WhatsApp, Instagram, Snapchat are all products and they're being consumed by many, many, many millions of people.
So what happens when you think about sitting in 2006, and why can't your brain figure out that Instagram is coming? Instagram is very hard to even conceive without phones that have an app store, have cameras on both sides, are generally available, that lots of people have, and the fact that people are comfortable being on social networks in 2006 is still an early thing. This is pre-Twitter, and Facebook is not that big, and MySpace is, but it's not the same type of private kind of community.
And so the point I'm making is I think it can come pretty fast, but the way and the form factor it will come in is hard to pick, just like it's hard to figure out Instagram is going to happen, because the intelligence in that particular type of hardware, which will be generally available, that's a key word, generally available, is really going to impact the use cases.
So I think the most obvious use cases that will come early are going to be use cases where you get the most amount of bang for buck, and the bang for buck is a car that drives itself or a mining robot, which is a mining vehicle which is now intelligent. And the reason is all that, you know, let's say engineering required to make this giant machine that moves dirt has already been done. It's been done over the last, you know, 50, 60 years. So then you're just putting a little bit of intelligence into it and leveraging everything else that the companies and people have developed.
So I think, and I'm not just, you know, pitching my own book, I mean we're a physical AI company, I continue to believe that our brain emotionally loves the humanoid concept because we're monkeys, and [clears throat] but actually, just more pragmatically, it's actually just putting intelligence into things that already exist all around us. And then once that happens, then new applications will emerge, which I think we'll talk about in five to seven years, which we'll start seeing.
So let's just move forward five to seven years and let's see what reality exists, and then maybe we can try to jump into the future from there. I think generally speaking, every single car company on the planet right now is working on a product that's like a Tesla FSD product. Every single car company there, without exception. Many, many companies are working on versions of that that will become fully autonomous within a cheap sensor suite.
So the fundamental difference, just to simplify it all, the Tesla approach versus the Waymo approach, just to really keep it simple, is the Waymo approach is lots of sensors and lots of compute and maps, and the Tesla version is very few sensors, no maps, or, you know, no high-fidelity maps, I'm just generalizing here, and cheaper compute, for lack of a better word. And the Tesla version of a product, this in the industry is called an L2++ product, is going to be available everywhere because it's literally cheaper and it doesn't require HD maps. The Waymo product functions better in a geographically constrained area.
So you fast forward 5 years, both of these types of technologies will be much more ubiquitous. L2++ and L4 will be much more ubiquitous, not only in the Bay Area or in parts of China, but really globally. There are companies working on this globally.
So now, I don't know if you remember, but nav systems used to be a big deal in cars. You would pay thousands of dollars, and nav systems were kind of the thing that everybody wanted. We're at that moment for L2++ systems where people are willing to pay thousands of dollars for a semi-automated vehicle. It will not be a long time, you're already seeing this happen in China, where the downward pricing pressure for that autonomous product, for lack of a better word, will make it close to free.
So now you fast forward 5 to 7 years and every car has some level of autonomy. So now you have to mentally live in that reality that everybody who's buying a car, they just get FSD with it. Now you start seeing a different world, because now the average person isn't wondering, is self-driving in a car? They use it all the time. They don't wonder about navs.
And so what you have in nav systems is CarPlay emerges and Android Auto emerges, and it's very natural. People are like, "Oh, I have my phone. I just plug it in." And it wasn't a big revolution, but the CarPlay and Android Auto revolution is actually huge. It brings free navigation and free applications to your car and it's fairly ubiquitous. And so I think the next thing that happens in five to seven years is then full autonomy becomes the thing that everyone expects. And with all of that, you will see a clear decrease in injuries and death because you have some intelligence helping you drive.
Now again, I'm using the consumer vehicle analogies just so people can understand it, but this is the same in construction, it's the same in mining, it's the same in defense. In every one of these verticals, there's these big physical machines that humans are interacting with, and teaming up with that machine is the future. The productivity unlock from you looking at a machine not like a sentient being, but almost like a physical agent of something you're trying to accomplish, unlocks things that I think are very hard to think about.
So I love, you know, things like Moltbook, and I love the, let's say, OpenClaw revolution that's happening, for lack of a better word, but I think the big impact, that's still such a small part of society. My barometer of impact is like, you go to, you know, the Detroit airport and you sit at a gate and you look around and you're like, how many people here are using OpenClaw? And it's like, you might be the only person who knows what that is. Whereas everybody, they're living their lives. And so to them, actually, the impact of AI is going to be in this physical world. I see you also have a... Yeah, there you go.
Check it out. I'm a convert.
There you go. Perfect.
Peter's coming on the pod soon, so I got some lobster claws.
Yes. So I think the real impact of AI in the next 5 to 10 years really is going to be in farming, in mining, in construction, in self-driving trucks. That's where you're going to have a real impact. I mean, I love the stuff that's happening on these platforms, but it's still segregated to, frankly, developers and a small, very, very small part of society.
I wasn't planning to spend so much time here, but this is extremely interesting and I think it's important for people to hear from folks like you about where things are heading because, as I said, everyone's just like, what is happening? What is going to be my future? The jobs piece is really interesting, and a thread that has emerged on this podcast recently is that people are afraid AI will take their jobs, but in reality AI is coming just in time to save us, because populations are declining, people are aging, and we need something to help us there. I know this is something Marc talked about and you're really close with him. Help us feel better about just how AI isn't going to take our jobs and is actually going to save us.
Yeah, I think honestly speaking, these industries need autonomy. I mean, it couldn't come soon enough. Frankly speaking, people are not fighting for those trucking jobs. If you look at farmers, the average age of a farmer is in their late 50s, 58 or so. What does that mean 10 years from now? That means many of those farmers are going to be retiring, if they're not already retired, and in 20 years we have an even bigger problem.
That, by the way, is every vertical, and, you know, my hypothesis here, sometimes people say, like, you know, McDonald's can't hire, or the mine, the local quarry can't hire, and where are all the people? The people are still here. I think the tradeoff is just not worth it anymore. In the 1980s and the 1990s, doing a long-haul trucking job was what the family had to sacrifice: the father not being there for days and weeks on end. And today that same working-class family can make that decision and say, you know what, I will drive for Uber or DoorDash, and I'm willing to do that because I can turn that app off and pick up my kid, and I prioritize that.
That is where I think this kind of intelligence revolution in the real world is really going to fill those gaps in, rather than an entire industry suddenly gone and it's just automated. I don't believe that future, mainly because the realities of actually replacing an entire industry with robots, that's still too complex. One day it will happen, but it's not happening anytime soon, and the entire society will be different by that point. And again, use the industrial revolution as a good version of that.
The earlier question, if I'm somebody who is not in the AI ecosystem and I have this anxiety, how would I deal with it? Reading history books is a great way to really understand how society deals with this. And there's a lot of literature, because the industrial revolution doesn't happen, like, in the dawn of Christianity, where not many people are writing and not many people are reading. Lots of people are writing, lots of people are reading in the last 150 years, and you can read both the people who were impacted by the industrial revolution and the people who were benefiting from it, and writ large, it's a very positive experience.
And that doesn't mean, again, there are no downsides. We should mitigate the downsides. But the thing that we can't do, and this is maybe specifically as America, or as the global population as a whole, there's this impetus to say, like, we've got to pump the brake on, again, don't say AI, say technology, pump the brake on technology. The issue then is the American economy really ends up
stuttering, and that impacts the lowest end of the labor market way more than anybody else. And so in the attempt to help the people who are the most marginalized, we actually hurt them the most. The statistics between Europe and America have been pretty explicit, but in the last decade basically the American economy has now grown at a much higher pace, and that growth hasn't come from Detroit, Michigan. That growth has come from Mountain View, and it's come from Sunnyvale, it's come from the Bay Area, which is another way of saying it's because of new frontier technologies.
So putting brakes on frontier technologies because we're afraid of unintended consequences will actually have real intended unintended consequences on people who we're trying to help the most. And the reality is very, very fundamental. In a future that does not take care of the average worker and the average person in America, we'll have much bigger problems. So we need a solution that takes that into account. But that solution isn't just pump the brakes: AI is bad, or frontier technology is bad, or technology is bad, or whatever thing that you don't like. I think that that'll have really, really bad consequences.
One of the reasons that we don't pump the brakes is just fear of China and competition with China. The nunchuck robots being a recent example of like, oh. I know you have kind of a contrarian take on just how much of a threat China is and how they're approaching things.
The summary version of this is, I think the way we... we recently read as a company this book, House of Huawei, which is just a really great, interesting book. And Huawei is a really amazing company for the reason that it makes great technology, but the couple hundred thousand people that work at Huawei, about a quarter of them are members of the Communist Party, and Huawei's goal is not to grow profits or shareholders. It's a private company; it's really an extension of the state. So literally the name Huawei means China's ambition. So imagine if you had a company called, you know, MAGA, and half of the company or a quarter of the company was a certain political party, and they said our goal isn't to make profits, our goal is just the expansion of... It's not even a company anymore. It's something else, right?
And so I think we incorrectly, when we, specifically speaking of Americans, think about China, we impart our understanding of markets and companies onto China. So we think Huawei, since they make phones, they must be just like Apple. It's like, no, no, no, actually they're not, that's not like Apple at all. And so I think the first thing I would implore everybody who thinks about China, especially with anxiety in America, is you're not comparing companies to companies. This is not apples to apples. This is very, very different.
And so imagine instead of thinking OpenAI is competing against DeepSeek, you say OpenAI is competing against the Chinese government. Instead of Apple competing against Huawei, Apple's competing against the Chinese government. And you can even remove the word... Chinese government is the best word to define what this organization is. But it's not a for-profit, privately owned, independent group of people who are working on projects together to build products to market. So that's the first very important thing. You cannot treat China like another America or another Europe or another whatever.
Number two is, if your goal isn't to make profits, you can do incredible research and it can be extremely compelling. But like we've seen, if the system is not sustainable, that's also not a company; that's not sustainable.
Let me give a very stark example of that. Chinese EVs are really lauded as being this exceptionally interesting product, right? And then you constantly get the stream of, I would say, fairly shallow analysis which says, look how good China is and look how bad Munich, Detroit, Tokyo, Seoul are, the other epicenters for automotive globally. There is a Chinese EV-like company in America. It's called Rivian. Makes great products, but they lose a lot of money making those products, and therefore the company is not very highly valued. I think if you said top 50 or top 100 companies in the Bay Area, I'm not sure Rivian would even make that list. And it's not that the products are bad or the people at Rivian are incompetent or they're not working hard. It's just the business is a tough business. The EV business in automotive is a tough business.
So how can we hold these realities? So we say, "Look how amazing these Chinese EV companies are and look how bad the home team is." It's just because the home team is being assessed for being a business. It has to make profits, and because it doesn't, it gets hammered by public investors. The other thing is not even a company. Now, if we do apples to apples, America just has to build great EVs. That means Tesla and everybody else combined, and we don't care about profits. I think America would field some very good products and there would be wow products. So the comparisons are really, really off, and I think that creates a misunderstanding.
I think then maybe the most philosophical question: can China succeed, and does that mean America has to fail, or vice versa? If you believe in open and free markets, you believe everybody can succeed in those markets, and that's been proven for over a hundred years. And I think what we're experiencing right now is how does China play in that ecosystem, because I said open and free markets, and those are not open and free markets. But that doesn't necessarily mean that you have to have an antagonistic relationship. It certainly doesn't mean that China is incompetent, and it certainly doesn't mean that it doesn't warrant our attention and our, let's say, focus. But it's also not a one-to-one comparison. I think we should be very careful in implying it's a one-to-one comparison.
And by the way, that like five-minute explanation is never going to get to the average person sitting at an airport in Detroit, Michigan, waiting for their flight. All they consume is China bad. It's not like that. It's not that simple. It's way more nuanced.
This episode is brought to you by Lovable. Not only are they the fastest growing company in history, I use it regularly and I could not recommend it more highly. If you've ever had an idea for an app but didn't know where to start, Lovable is for you. Lovable lets you build working apps and websites by simply chatting with AI. Then you can customize it, add automations, and deploy it to a live domain. It's perfect for marketers spinning up tools, product managers prototyping new ideas, and founders launching their next business. Unlike no-code tools, Lovable isn't about static pages. It builds full apps with real functionality, and it's fast. What used to take weeks, months, or years, you can now do over a weekend. So if you've been sitting on an idea, now is the time to bring it to life. Get started for free at lovable.dev. That's lovable.dev.
So you decided to join Twitter recently. Put out your first tweet. Your first tweet was just like, "Hello, I'm going to start tweeting." That tweet got like 2 million views. Elon replied to you. Marc Andreessen quote tweeted it and said, "This is the best AI CEO nobody knows. Follow for free alpha." Elad Gil, famed investor, describes you as the most successful, most quiet company in AI. And to me this is really interesting, because most founders are told build in public, build a following, be loud, get out there, talk all the time about what you're doing. You did the opposite. You were very under the radar, stayed quiet, build, build, build, build, and then decided later, okay, now it's time to talk about our story. So I think this counter-narrative is really interesting, and I think will inspire a lot of founders to not feel like they have to do this. What was your philosophy of staying quiet and then starting?
Yeah, it's a great point. So number one, it was intentional, and I think if it was up to me, we would do that forever. I think we're very much inspired by folks more like a Berkshire Hathaway and less like, let's say, a Silicon Valley darling.
I'll tell why I changed the views, but before some founders go and take that advice immediately without really thinking about it: I can do that because I'm known in the ecosystem. You know, I know these folks personally, and so I don't need to have a brand out there that is getting Elad to remember me and think about me. My first two companies were a lot less known, before I really came to YC. So, you know, all of our company values can be reduced to these two words: radical pragmatism. So before you take the advice, make sure it applies to your situation.
One of the reasons... and Naval, who's one of our investors and a friend, says fame itself is like a tool, and it's powerful. Now, if you don't have a network and you can get a following, that's a fantastic way to recruit people to your company, to recruit investors to your mission, and then of course customers. But for us, I think that wasn't a hard requirement 10-plus years ago.
The other thing is, I think Peter and I... you know, the old saying about life is kind of like you do things and then you rationalize the thing that you do, right? So I think fundamentally Peter and I, my co-founder Peter Ludwig, we don't get a lot of emotional satisfaction out of doing very public things. And I think if I was really to play armchair psychologist and really try to get to the root of why, beyond the rational view, which is focus on your customers, focus on the product. Every minute you're doing a podcast, every minute you're doing an X post, every minute you're writing something for public consumption, you're not focusing your very limited time that you have on your customers and your product. And ultimately that's the only thing that's going to produce and yield results. But the reality of the situation today, in 2026, is even a company like us that's known, or somebody like me that's known in the ecosystem, you still want to get that broader message out, and that's what I talk a little bit about on X.
So it is definitely contrarian, but it's not just contrarian for contrarian's sake. It plays a little bit to our own psychology.
And then I would say, just to finish that thought there: I'm an immigrant, came to the US from Pakistan when I was a kid, have a little bit of a weird name. I grew up in Detroit, in Warren, Michigan specifically, for all those at home. And when you feel that you're a little bit on the edge of society, or you're not maybe in the mainstream, and this resonates with some people, it doesn't resonate with everybody, you feel very skeptical of the mainstream because you're just on the outside for so long. And I think you can trace a bunch of founders' psychology to this feeling of being an outcast, actually. And so then you find yourself in a situation where you're the COO of YC, and the narrative of I'm an outsider is like, I don't know if there's anything more inside than being the YC COO, right? So I think that reconciliation over my career also has had to happen, which is like, maybe that's just kind of a weird kind of thing.
And so when I talked to Marc Andreessen, who really pushed me to go online, or Elad, or whoever it is, their view is leave your baggage and your trauma in the background, and let's think more pragmatically. And the pragmatic thing here is, whether I like to do these types of things or not, fundamentally it helps get the message out. And the message can be something very small and myopic, like what's happening in physical AI and machines becoming intelligent, or much larger, which is what's happening in society through this fundamental change that we're going through.
I've had the rare privilege or experience of seeing the full economic spectrum. I've really seen the extreme ends of both sides, and truly, I really mean that. And so somebody like Marc, who's close to our company, says, well, those are some ideas that are worth getting out, beyond just promoting your company or something like that. And that I can actually get behind, which is the debate and discussion about ideas and what's happening to our society because of these technical changes. And so here I am.
Amazing. Okay, so there's a few threads I want to follow there. One is, you were, as you said, COO at Y Combinator. You saw a lot of startups up close. This is your third startup on your own. Something that I hear you talk about is that successful companies almost always show traction very early. A lot of founders hear, like, no, just keep trying, fighting, and maybe we'll be the next Figma, Notion; four years in, we'll figure it out. What's your experience there, and what's your advice to founders who aren't seeing traction early?
Nuance. I mean, if I was starting another company, I'd call it Nuance, right? I think what you're saying is correct. I continue to believe that good companies tend to have traction fairly early and then just sustain it for a decade plus.
To the founders that are toiling: let's say you're listening and you're about two years into your company, and you're maybe having a tough time getting money and building that first product that consumers or businesses really love, either through attention or dollars. Two years is a difficult time. The heuristic that I would use is, if the information I'm getting from the market is not informing me on a more and more specific path, I would consider resetting. And what I mean by reset is, oftentimes, and this is wearing my YC hat, seeing hundreds and thousands of companies, oftentimes the co-founding, literally the foundation upon which the house is built, is not correct. It's like, imagine you built this house, and every time you put a cup of water down, it slides off the table and falls on the ground, and you keep adjusting the table. It's like, maybe the foundation is actually wrong. The whole house is off-kilter. And that foundation might not only be your co-founders. It could be the market that you're in. It could be the phase of life that you're in and the amount of effort that you're willing to put into that thing in order to make it successful. There's a bunch of reasons that a company can fail, and you have to be able to somehow say, I don't know what the reason is, I'm just gonna have to hard reset here.
One thing I would tell founders, and I tell... Applied is creating a founder class in itself. People who've worked at Applied Intuition are now starting their own companies. We have a thousand-plus engineers, and over time they're starting their own firms. And I say to all of them, just imagine the first time you're going to do a startup, for the first three years it's a zero. Just rid yourself of the expectation that it's going to be successful, and that you're really a craftsperson. If this was a woodworking podcast and you said the first table that you built was wobbly, you wouldn't say, well, go work at Crate & Barrel. You'd say, that's the first table. We're going to...
Keep at it. Being a founder is its own muscle, and you want to exercise that muscle. But I think a lot of founders, especially early in their founding career, put such an incredible pressure on themselves to make it great out of the gate that they actually miss the thing that you're getting in that first round, which is learning and building that muscle in the second, third time. And I think it's not random that my third company is the most successful company. I think that's... you see that more often than not. There are funds which are almost exclusively focused on multi-time founders, right? For this reason.
What I love about that advice is often the best ideas come from when you have low expectations. You're just playing around. You're just tinkering. You're not like, I'm going to build the next great, I don't know, Google. It's just you having fun. And that's like how I found this world that I'm in right now, this path. And OpenClaw is a good example of that.
I think why that advice is so difficult is if you hear this and you're in the proverbial war, you're like, "What the hell are these people talking about, having fun? This is hard." And so you have to hold these contrasting or conflicting views in your head, which is like, it's deeply, very, very important and you should give it your all. And it's also not that important. And that's a really hard thing to reconcile and keep in balance.
And the way that you approached this company, where you stayed quiet like that, I think helps a lot, where you're not...
Absolutely. Yeah, it's that, like, you know, even at YC when I became COO, I told Sam—Sam Altman was president—and I told Sam, let's not announce this for like a year, because, you know, if the partners don't want me to be COO, it's not a successful thing. I don't have the pressure of the public scrutiny of, you know, why were you COO only for six months or something like that. And I think you have to be very honest with yourself as a founder and as a human being that those things matter. What people think about you matters, and it impacts yourself, having the spotlight on you. You know, I always say it's very easy to pivot before you raise money and before you have employees. Nobody cares. The moment you raise money, and more importantly the moment you hire employees, the employees join a very specific mission. And you walk into the office and there's 10 of them. You say, "Guys, turns out this isn't wrong. We're going on a different mission." Imagine if this was war. It's like, "What the hell? We just were attacking that hill and now we just say that hill is not important. Like, how do you know the next hill is important?" And you as a leader lose a lot of credibility. And it's not only for the superficialness of being a credible leader. It's the practical nature of: when you're very, very public, the startup becomes your identity, and then suddenly you're having to reconcile that actually that thing is not correct. So it is easy. We have these core values in the company, and early in the company I used to have this line which says our best work is done alone and quietly. And I deeply believe that, and so founders, I would think of it that way. But it's for pragmatic reasons. It's not just because it's cool to be under the radar. It just allows you to maybe work in a bit more peace.
I love these core values you've shared so far. The last one, the best work is done alone quietly, I'm so on board with that. Radical pragmatism is the other one you shared earlier. Are there a couple more there? These are gems.
Yeah, those are, I would say, the meta values. We have very specific, let's say, operating principles, and this is as tactical as advice I can give to founders. Come up with your values when you're getting a little bit of traction, and the reason I say that is it's early enough where... And the way you come up with the values is not like, what values should we have, like as philosophers. No, no, you should figure out why are we being successful. Literally write down the five to 10 things that are the reasons you are being successful, and those become your values. And so we did that, and our first one was, you know, speed above everything, and it was like us being fast. The second one is like, you know, never disappoint the customer, technical mastery, high output matters, all the way down to ones that are not obvious, like laugh a lot. That's been our core value from the beginning of the company's history. And it's like, when you're working on intense things, if you don't have the ability to keep grounded, have perspective... Laughter and humor also is a way to get subtle feedback, and a slightly different taste than "this sucks." You can say it's not the best. And that is slightly in. And so you're really creating the framework in which people are learning how to behave with each other within the company. And so today the values really serve us as almost like guiding principles. And so, you know, we do new team meetings every week. Peter and I, we meet all the new team members, and we're almost always just talking about the values in some level of detail, in depth. Yeah. Another value: half the work is follow-up. Just taking notes and following up. That is the business. It's not more complex than that.
Laugh a Lot is my new favorite company value. Yeah, sounds like a wonderful place to work. And then on the last piece there, there's this book that just came out by Stripe Press about maintenance and how valuable and underappreciated the maintenance part of work is.
Absolutely. Absolutely. Yeah. I think if there's a takeaway that you get from, let's say, a bit of my philosophy on where, you know, we started the conversation around why... being promotional has all these negative connotations in it, so I'm careful using that word, but why not be promotional? It's because there's costs to everything. And so if you can focus on the craft and making the product really, really good and really listening to your customers, you have a much higher likelihood of success, and then you can always then go and scale there. A part of that is the thing that you're talking about, maintenance, or another version of it.
And, you know, my roots are in automotive engineering, and automotive engineering is actually an exercise in quality. That's really what it is. You're building these very complex machines at scale. You know, people talk about rockets being really, really complex. You only got to send up a rocket, even at the highest capacity, like once every couple of days. You're making a car every 30 seconds, and you have to make it extremely cheap, and it's globally competitive. So you really get into the nuance and minutiae of how a factory runs, and a factory is about safety and maintenance. There's not a lot of complex things. It's just, you know, when you break down what is being operationally strong: operationally strong is keeping an eye on a handful of things and making sure you're doing them really, really good. And I'm one of those believers that, you know, there's this adage: a man who cannot command himself is not fit to command others.
And it's like that maintenance aspect is a part of that, right? It's like if you maintain yourself and your own work, you maintain your team, you maintain the company, the products almost come out of all of that system. And I think a lot of founders don't think about their company as a system, or almost as a machine. But I would implore you to do that, because then you really focus on the craft of the machine and building the machine and making it more hygienic and making it more well tuned. Just like, you know, you'll meet people who really love cars and they really obsess about the maintenance of cars. Like, they will detail underneath the driver's seat. As somebody who details my own cars, nobody's gonna look at that. But it's under that same ethos of really caring a lot about the craft. And frankly speaking, since you have a limited amount of time, it's hard to really care about X and also make sure your company's hygienic. And there's different reasons at different points of your company that you should do different things, but that's kind of a little bit of the ethos.
I love how it keeps coming back to just staying quiet, just working on the thing and not talking about it. Your last point there makes me think of The Score Takes Care of Itself. Classic.
Yeah. So Joe Montana is actually one of our investors. In our Series D post, the post was "The valuation takes care of itself." Very much we fall into that category.
And it's like, you know, sometimes people will come to our office and they'll say, oh, it's such a clean office, you guys must have this giant cleaning staff. And it's like, actually, we clean our office. Just like in Japanese schools—as I mentioned, I lived in Japan—the students clean their own schools. We have a cleaning zen every week and everyone cleans the area around them. And I think it's important that there's something about this ethos of also not getting so wrapped up in your own narrative of, like, I'm a Stanford software engineer and I do AI. It's like, clean up your desk. So there's some basic things like that. And I don't know what that larger philosophy is, but it is a philosophy that we kind of drive towards. And I think, you know, our claim to fame, which is kind of a crazy reality, is we've never spent any money we've ever raised in the history of the company, which almost sounds like it's made up. It's true. The company is almost 10 years old, a thousand engineers plus. And so we're a functioning business without using capital that we've raised. And I think it's somehow connected to us cleaning the office. I don't know how.
Saving all these cleaning costs. It all makes...
We still have people, you know, clean, but our employees also are aware of their surroundings. And I think there's a direct line between be quiet and alone, and clean your desk, and well-written software. And I don't know what that thing is, but it all falls in the same arc.
I know you also have a no-shoe policy for that same reason, to get things...
Yeah. Yeah. And also influenced by Japan. I worked there and we had a similar office setup. The other way to think about this, maybe—again, I'm just trying to impart everything I've known to founders, because I feel like that's who... that information is so limited and everyone's kind of trying to make it up, frankly speaking.
The alpha.
Yeah. Exactly. I would implore you as a founder to really try to take the best of Japan and the best of Germany, the best of China, the best of Detroit, the best of Silicon Valley. And, you know, I think sometimes people take that Steve Jobs line and they really deform it, where they say, like, great artists steal. What he's really talking about is, the less magnanimous version of that is: be humble and learn from everything around you, and as a leader, be well-rounded. I think reading should be... you know, there's a Charlie Munger line where he says, I've never met anybody very, very successful who doesn't read all the time, and I very much fall into that category as well. And so if you unpack why that is—why does reading a physical book make you a better founder? Just ask that question in the most direct way. I'm not reading... you know, especially, my ethos of reading is: read old books. Don't read anything new. Read old books, because time has filtered out a lot of the noise, so you get a lot of signal. And in your life you're going to... a thousand books maybe you'll read, in the best case scenario. You're going to read probably 50 to 100 books, which is kind of crazy, for the average person. So you're just not going to read many. So don't read low-quality content. There are true pillars of human ideas out there. You consume those ideas, and then it's up to you to interpret how those ideas then reflect upon the business that you're leading or the technology that you're developing.
I absolutely believe reading a book like Malcolm X's autobiography will make you a better founder. And again, it's like the whole cleaning zen all the way to clean code. It's not directly one-to-one related. I think we always want these very simple if-then statements. But I think being a well-rounded founder, where you understand society around you and history around you, somehow makes you build a better product, and I don't know how and why, but I think it absolutely is true. I do see a connection there. And then people like Charlie Munger, who is obviously not an AI founder, also believe that, and so I think there's some pattern there.
It's interesting, this is the same... it's a metaphor for LLMs. You feed all this data, somehow they become almost conscious. How does that happen? No one fully knows. It's so interesting how similar you are to Marc Andreessen in your way of thinking and the way you consume content. Like, there's...
We're both bald.
There's a thread here. Here's important ingredients to being really successful.
Yeah. I mean, Marc... you know, we're fortunate enough to choose our investors, and that's a true privilege. I didn't have that in my first company. We spent years... because we didn't raise a dollar. So I certainly appreciate it. But, you know, if I've ever had a mentor, Marc would fall into that category. And so, you know, before Applied, we debated and talked a lot. And I think Marc is also like that, right? He really consumes content actually outside of this little industry that we're in, and I think it makes him actually a better investor.
Yeah. We'll point people to your website. You have a list of the books that you recommend and love, and it's very long and very not what you often see. I can't help but just ask: are there a few books that have most influenced your thinking, most influenced your life?
Yeah, I've been thoughtful about that list. And the reason I use books like The Autobiography of Malcolm X as an example is I know that's not on the top of the list. If I say High Output Management, classic Andy Grove, you guys know that. So book here. Yeah. Yeah. Exactly. It's partly the theatrics of also entertaining you, but also giving you new information as a listener. I think, like, the books I'm currently reading—this is kind of a random slot of books—I'm reading the vibe coding book that came out; our whole company's reading that, which is a new book, and it kind of goes against the grain of my heuristic. But The Emperor of All Maladies, the cancer book, fantastic book. I'm almost done with it. I think it changes the way I think. You read that and it changes the way... And I think that's the ultimate test: when a piece of material changes your existing framing on life. This is good in the LLM use. This is somehow related, in the sense that diverse data makes your understanding of the world more rich and nuanced, and therefore it's better. But yeah, so like I'm always
…inspired to give more wacky kinds of examples rather than the obvious ones. But the obvious ones, if I really wasn't being theatrical: I think Sam Walton's book Made in America is an unbelievable book. It's very, very good. He wrote it on his deathbed. My American Journey is also very good, Colin Powell's book. It's not on my website, but it's really good.
I'm somebody who tries to connect some of these dots from us being cave people to now living in Silicon Valley, working as a venture-backed AI company founder. And so books like Guns, Germs, and Steel really are top of that list. A fantastic, fantastic book. Or Collapse, also same author.
So yeah, but my point to founders is read the stuff. You can still go to physical bookstores. Read the stuff that is both old and well regarded and you know nothing about. Often when I'm trying to find the next book to read, like I remember the way I picked up SPQR, which is the book on Roman history, was I was like, I don't actually know a lot about Roman history. I know the high-level stuff. So think about all the ideas in the universe, from philosophy to history to Jainism to the rise of Japan as a feudal state, areas you don't really know, and then just find the best book in that space. And I think you just start filling in the blanks. And so often that's kind of the way that I grow the ecosystem: what don't I know anything about, and let me find the best piece of material in that. And yeah, it doesn't work.
Oh, I like that. So Marc's philosophy is this barbell strategy of only today's news, like X, and 10-years-ago books. I love that you're like, no, just like upside-down T almost, just only.
Yeah, you know Marc was a heavy influence in me getting on X. So he's propagated that view, and the thing that I really do agree with him on is:
As our company becomes a larger, more influential and impactful company to society, it is my responsibility as a co-founder of the company and the CEO of the company to at least propagate my ideas to our physical AI founder community, and then larger, the technology leadership, and then the world at large. And so that's part of it. I think in that way Marc has really taken it. You think about VCs: not long ago you would never even know their names. They were like PE guys or hedge fund managers. You can't even think of names. They're just blobs of ominous-sounding Obsidian Corporation or something like that. And it's only a16z and a couple of other folks, John Doerr, who really kind of created the, hey, I'm going to be the individual investor, and I'm going to propagate a certain set of ideas, and that's going to create gravity within Silicon Valley and influence founders to then make certain types of companies, and of course they invest in those.
On this thread about reading to find almost areas you disagree with and haven't thought about. I know one of your approaches to management, and maybe one that's a value, is to encourage your leaders to listen to naysayers, to not create this positive reinforcement cycle. Talk about why that's so important, how you operationalize that.
So imagine I'm not the founder of the company and Peter's not the co-founder; it's just a generic company. The ideal situation with a generic company is one that you can put in lots of different competing ideas. The culture is one where you will shake those ideas out. There's not an emotion in it. Whoever brings the idea, the best idea wins. So why can't companies do that? Frankly speaking, a lot of times it's the founders. And the founders are told, by popular media and the way that human beings experience life and our tribal kind of outlook, that you have to have this hard view, and if everyone's not following you, then maybe you're a weak leader or something like that.
And I think we just don't believe that. We just don't believe in that philosophy. And maybe a more tactical way of saying that is we take inputs of the environment: our customers specifically, our employees, our competitors, our investors, and what's happening in society at large, and that impacts our strategy. And I think it's one of the reasons we've been very, very successful. We're not so arrogant to think that we just have the answers because we had the ambition to start a company. And I think that permeates into a very specific culture. It's also not being contrarian for contrarian's sake, right? One view I do have is emotions are generally not helpful in making rational decisions. They're almost the opposite of it. And sometimes passion and leadership are supposed to be, again, magnanimous or emotional. And we just don't believe that. And that's, I think, a bit more of our kind of Midwest roots showing for Peter and I.
And so it's not that we say, "Hey, disagree with everything in the room." What we specifically say is speak up. Speak up. Everyone has to speak up, because that one person with their one experience, because they worked at Zoox or Waymo or Tesla, or they worked at a Chinese company, whatever it is, that one idea they have in their head when the debate is happening about what we should do in space, like literally the space of space, and something maybe we don't know much about, that one person's one idea, they have to feel comfortable sharing it. Even if they're the most junior person, or they feel that they didn't get their way in the last debate, or whatever anxiety they might have, they have to share that opinion: guys, this is actually the right idea, or this is the wrong idea. And if you can create that environment where the best idea wins. Gandhi has this line: truth is what stands the test of time. And I think it's become a little bit of a meme in the Bay Area, truth-seeking as a culture, but it kind of is like that. We're trying to find the best idea. Maybe truth is the wrong word; maybe it's the best idea. Find the best idea and then let's go full bore against the best idea.
Let's maybe use a counterfactual. Why do companies fail when they have great talent and they have seemingly all the same components that an Applied Intuition has? It's because maybe the best ideas are not being surfaced, and certainly maybe they're not actually being adopted. Or more often than not, when I think about companies that have been very successful, they have momentum going in a specific direction, and that momentum, that wall of sound, overwhelms any new sound that's emerging, which is, hey, the market's changing, the market's changing. You just can't even hear that because there's all this momentum going in a particular direction.
A good example, I had front row seats for this when I worked at Google. It was the era where Facebook was emerging. And here was Google. And people don't remember Google in the late 2000s and early teens. Google wasn't just a company. It was the apex predator of Silicon Valley. Apple was just, you know, the MacBook Air had just come out. Steve Jobs was getting it in the right direction, but nothing like Apple is today. Amazon, AWS was still a young thing. Nvidia was teetering on bankruptcy. I mean, all these giant companies that you think of: Microsoft was run by Ballmer, Twitter was a small thing. But Google was this already larger-than-life number one company. Everybody wanted to work at Google, and there were not many companies with that stature. And then in the periphery, this little company Facebook starts emerging, and Google, who has the best engineers on the planet, is making a billion in cash flow a month, tries to fight this little company. And I remember Facebook at that time maybe had a thousand people, and Google was like 15x, 20x the size, with a lot of cash flow. And why couldn't Google fight Facebook? It's because Google is not Facebook. It's like that Confucian saying: how does a gorilla learn how to fly? By not being a gorilla. The way that Google would have won the social media wars is by being a social media company. It's just fundamentally not.
And so this happens in companies all the time, which is you're just going in one direction with momentum, consciously or unconsciously, because that's what all the employees are, that's what the culture is, that's what they work on. And then something changes in the market, and you just can't even move there. And I think that also can happen, surprisingly, at really small companies, where founders have a view and it's like, that view is the view it's going to be. And actually that can be just 10 degrees off from what was the correct path, and the whole company's kind of led astray. So they were in the right market, they might have even been solving the right problem, but they were just a little off. And so we're so scared of failing and so scared of losing that I will humble myself and listen to other people, and they say, hey, we're five degrees off course here, and it's like, okay, let's maybe fix the course. And then once that becomes your culture, then it's really hard to lose, because everybody's not about fulfilling a preset path. They're just about finding how to win.
This is exactly what I wanted to ask about. Everyone listening to this either is like, oh yeah, we're very open-minded, we're absolutely going to listen to everyone's opinions and decide rationally the right path. In practice…
Almost never happens.
Right. Or they're just like, we know we're not good at this. We're just too nice to each other. How do you do this at a company that isn't good at this? Does it have to be the CEO top down, in your experience? Does it have to be part of the culture? How do you operationalize it at a company that's not like you?
You know, the middle way is typically the right way, and it's hard to find the middle way because these are conflicting ideas. The guardrails, or the goalposts you just set: one side is we're just going to go, and the other side is we're maybe almost unsure. And once you do have that debate, you have to then confidently walk down that path. And again, this is conflicting. I just said be humble enough to listen to what's going on. But then once that decision is made.
In our values, that first value, the speed value, which specifically the wording is "move fast, move safe." That's specifically the wording. We assess our managers on adherence to those values. Literally, we compensate and promote against those values. They're not just abstract values. So the behavior we're actually looking at under speed is decisiveness. So we're setting up a system that is looking at these conflicting kinds of things. One is be open, and the other one's make decisions quickly, and you have to hold those in tension. This is why you as the co-founder or founder of the company get paid the big bucks. You've got to do that. You have to know when to bluff and when to hold them and when to fold them, as I say. So at some point, and that point comes sometimes faster than you think, you will not get any more information and you have to make a decision. So you're walking this very, very thin line.
Your point about emotions was extremely interesting, and I want to make sure people don't take away the wrong takeaway here. So what I actually found really helpful, which I think is aligned with what you were suggesting, is taking emotions out of it. The way I've used this in my work is, when you have to make a hard decision, pretend nobody's feelings would be hurt and emotions are not involved. What would you do if nobody cared, if they're totally great? So what would you do in that world? And then that tells you, okay, that's actually the right thing to do. And then it's, okay, how do I help people feel okay about this? How do I deal with the downsides of this path?
Yeah, I think that's the obvious version of it. I think maybe another version of it is, as an emotion, let's take another: you as the leader, or as the engineer who is getting direction, you already have some preset view that this is my idea. That's an emotional construct, and it's around ownership and feeling of ownership. So yeah, I really fall into that category of, maybe most fundamentally, what is an emotion? An emotion is this set of reactions, a framework that's been imparted in your brain through life experiences. And those life experiences have not been optimized for you to make a decision in a product review. And so the more you can pull that away. A good heuristic would be the same decision being made by multiple people in the company gets the same result. So you're removing a little bit of that almost filter. You can almost think of that emotion as a filter. So I like to have the raw image come through, the raw decision come through, so we can consistently classify it again and again. Not to get too abstract, but I don't know if that makes sense.
Yeah, it makes sense. And I have one last question, but there's an interesting trend I've noticed with people talking about AGI. The missing piece I've been hearing more and more is just emotions are what creates consciousness, potentially. Michael Pollan has a new book out about consciousness, and his take is it's not just more intelligence, it's actually emotions that led to the consciousness.
I think it's undermining, let's say, how complex human thought is to think that it's just the inputs and outputs of, or let's say association, for lack of a better word, of ideas, facts, words, letters. It's not just those associations. And creativity is kind of a little bit of that as well. The old saying of: technical mastery is mastering the complex, and I think computers do that really well, and creativity is mastering the simple. I'm sure I'm going to eat my words on this, as the best artist in like three years will be…
Yeah. Yeah. And so this again goes back to my philosophy of consume broad inputs, but then try to remove that filter, see things as honestly as you possibly can. Create a culture in the company that is also similar and doesn't put any weight on who the idea came from or where it came from, but then ultimately, as a leader, you decide. And by the way, you've got to be right.
That's the other thing I think a lot of founders just don't emphasize enough. Founders love to take credit for things. It's just human nature. Everybody does. But the reality is you have to be right. It's not enough to just start a company. It's not enough to have this vision of the world. You have to be right. And the evidence is: is the company a sustainable standalone business? Because we're talking specifically about venture-backed AI companies in Silicon Valley. I should have said this at the beginning: all my advice is specifically for that narrow group, which is founders of AI venture-backed companies in the Bay Area.
Speaking of that, last question. I've…
I've been wanting to get to this because it's an interesting spicy take that you have as a last question. I know you have to run after this. You have this view that a lot of CEOs in Silicon Valley don't actually have great taste. I'm excited to hear what your experience there and just what you...
Yeah, I also want to be careful to imply that I do.
Okay. [laughter]
I fall into that group. I think, you know, it is true because of a couple of reasons, both taste in the most, let's say, artistic sense and in the most specific, like running a company and like what should be the policy, like HR policy for point X. A lot of that is I think they're just not exposed to a lot of interesting good things, and that's been a theme in this whole conversation, is like just get more and more exposure.
It's very unfortunate when I meet somebody who, and I'm not thinking of anyone in particular, so if this is you and you're one of my friends, I apologize, I'm not talking about you. But it's like, you know, you grow up in Cupertino, you go to Berkeley, and the first thing you do when you come out of school is you start a company, and then that's all you do, you know, for 20... like you've never even been an employee.
And why I think that's so important: like, I spent over a decade working in large organizations, and like truly large, like more than 100,000 employees, like a General Motors or like a Bosch. And when you're in the back alley of that organization, the bowels of those organizations, you learn how bad it is to be an employee. Like the bureaucracy above you, leadership doesn't know what's going on, the industry, your antiquated tools, all this stuff.
Why that's so important to experience as an individual is then when you become a leader, you're making policies and you're creating culture and you have to keep that in mind. And a bunch of founders just never had, frankly, the fortune of being at the bottom of the totem pole. And that's just one version of how to... that doesn't obviously seem like, you know, reread, consume the photos of Brissant or Picasso or whoever it might be. But it's something similar.
There's something similar about you. You can sometimes meet founders, and maybe a good heuristic here is like there's some founders that would be good at lots and lots of things, actually, not just being a founder of an AI company in the Bay Area. And there's something about taste there, because what you're really talking about is like understanding humans and understanding life and then being able to discern with some judgment what is good and what's not good, because that's really what we're talking about.
And so if your life experience is very narrow, you could still be good and might have the ability to discern what's good and what's not good. But I think there's something about, like, if you've backpacked for a few years around the world, I somehow believe that's going to be a better founder. It's like, you know, I don't know how I can... there's no peer-reviewed research that I can point to that says that. So I think that's what I'm getting at. There is some developing of taste.
Yeah. Well, I feel like we have helped people build their taste, feed their model with more insights and different perspectives in this conversation. I feel like we could chat for hours, but I know you got to run.
Yeah, I'm not sure if there were any real takeaways other than...
Okay. [laughter]
We really went everywhere. I'm sorry if you had a particular line of questions you wanted to go down.
We went in all the perfect directions.
Okay, good. Good.
Qasar, thank you so much for doing this. Thank you so much for being here. Final question: just where can folks find you online? How can listeners be useful to you?
That's a great question. I mean, I love to hear what are books that I don't know. So that's always good. Some of my favorite books have been just randomly kind of recommended to me. So I'll take that. Of course, I consume research as well. And so if there's something particularly novel that's going on... obviously all the mainstream stuff we as a company and me as an individual are going to consume, but things that are a bit off the beaten path, we're always looking for that.
But yeah, and then if you have a particular opinion about specifically our domain, physical AI, and how AI is going to impact mines, farms, construction sites, robotaxis, all of that stuff, I'm always interested to hear new opinions on that, or even old opinions maybe with a different viewpoint. So yeah, you can see me online, of course. I'm always around as well. I'm always open to that feedback.
And you're on Twitter now. There you go.
Yeah, exactly. Yeah, follow me there. Yeah, exactly. That's the answer.
That's the call to action. [laughter]
Channeling my inner Marc.
There you go. Qasar, thank you so much for doing this and for being here.
Yeah, thanks for having me. It was a lot of fun.
Bye everyone.
Thank you so much for listening. If you found this valuable, you can subscribe to the show on Apple Podcasts, Spotify, or your favorite podcast app. Also, please consider giving us a rating or leaving a review, as that really helps other listeners find the podcast. You can find all past episodes or learn more about the show at lennyspodcast.com. [music] See you in the next episode.
Article published · Updated