All podcasts / Lenny's Podcast / Summary

The real AI revolution isn't software. It's farms, mines, and trucks. | Qasar Younis

2026-03-08 - 84 min - source - Read full transcript

Lenny Rachitsky (host)Qasar Younis

Key insights

The biggest AI impact over the next 5-10 years will be physical, not software - specifically farming, mining, construction, and trucking.

Qasar argues these industries have decades of engineering already built into their machines, so adding intelligence to existing hardware (a mining vehicle, a tractor) delivers the most 'bang for buck' compared to novel software categories or humanoid robots, which he says draw disproportionate attention mainly because human brains are wired to find them emotionally compelling.

physical-ai-autonomy

Autonomy is converging on two competing architectures - Tesla's cheap-sensor, no-map L2++ approach versus Waymo's sensor-and-map-heavy L4 approach - and Qasar expects both to become globally ubiquitous within five to seven years.

He compares this to the earlier transition from paid nav systems to free CarPlay/Android Auto: once autonomy is bundled into every vehicle by default, drivers will stop consciously registering which company built it, the same way nobody thinks of in-car navigation as a discrete purchase anymore.

physical-ai-autonomy

Physical-world automation is arriving 'just in time' rather than displacing workers who want those jobs, because the workforce for the dangerous, unwanted jobs is already aging out or refusing to take them.

He cites the average US farmer's age (late 50s) as a signal of an impending labor gap, and reframes trucking and mining jobs as ones people are actively avoiding today (citing the more flexible gig-economy alternative of driving for Uber/DoorDash) rather than jobs autonomy is stealing from willing workers.

physical-ai-autonomy

Roughly 30,000 annual US driving deaths are the strongest argument for pushing autonomous vehicles now, and Qasar predicts society will eventually judge human driving the way it now judges child labor.

He argues the anxiety framing around self-driving job losses obscures the actual stakes: self-driving systems are already statistically far safer than human drivers, and the harm of continuing to let tired, distracted, or impaired humans drive is a bigger moral cost than the transition pain of automating it.

physical-ai-autonomy

Applied Intuition stayed almost entirely out of public view for close to a decade, guided by an internal value that reads 'our best work is done alone and quietly.'

Qasar frames every hour spent on public-facing content (X posts, podcasts) as an hour not spent on customers and product, and says he and co-founder Peter Ludwig get little emotional payoff from being public, which reinforced the strategy independent of its stated rationale.

founder-visibility-strategy

The advice to stay quiet and avoid building in public is conditional on already having an ecosystem network - founders without one may need visibility to recruit talent, investors, and customers.

Qasar explicitly warns founders not to copy his approach uncritically: he could build a $15B company under the radar partly because he was already known in Silicon Valley from his Y Combinator tenure and personal relationships with people like Marc Andreessen and Elad Gil.

founder-visibility-strategy

Comparing Chinese tech giants like Huawei to US counterparts is a category error, because Huawei is structured as an extension of the Chinese state rather than a profit-maximizing private company.

About a quarter of Huawei's roughly 200,000 employees are Communist Party members, and Qasar says the accurate framing is 'OpenAI competing against the Chinese government,' not against a company - so the metric of success for a state-backed entity that doesn't need to turn a profit is fundamentally different from a Western public company being judged by investors.

china-ai-competition

Praise for Chinese EVs as proof China has 'won' on automotive tech ignores that the comparison isn't apples-to-apples, using Rivian as the counterexample.

Rivian makes a well-regarded product but loses significant money and is valued far below top Bay Area companies specifically because it is judged as a real business; Chinese EV makers aren't judged the same way, so the perceived gap overstates Chinese competitive advantage rather than reflecting a fair business-to-business comparison.

china-ai-competition

Applied Intuition's culture requires that the best idea win regardless of who proposed it, operationalized by mandating that everyone - including the most junior or previously overruled person - voice their specific relevant experience before a decision is finalized.

Qasar says company failure often traces to good ideas simply not being surfaced or not being adopted once surfaced, or to a company's accumulated momentum in one direction making it structurally unable to hear that the market has shifted (his example: Google, despite having the best engineers and huge cash flow, structurally could not become a social network to compete with early Facebook).

decision-making-culture

Company values should be derived retroactively from what's already driving success, not chosen upfront as an abstract philosophical exercise.

Qasar's method: once a company has some traction, write down the actual five to ten reasons it's succeeding, and those become the values. Applied Intuition's resulting values include speed above everything ('move fast, move safe'), never disappoint the customer, and 'laugh a lot' - the last one used deliberately as a way to give critical feedback without it landing as pure negativity.

decision-making-culture

A useful reset heuristic for founders stuck without traction around the two-year mark is whether market feedback is narrowing the path at all; if it isn't, the underlying foundation - not just the current tactic - may be wrong.

Qasar frames this using a wobbly-table analogy: if you keep adjusting the table (the product) and the water still slides off, the problem may be the foundation itself - which could be the co-founding team, the market chosen, or the level of personal commitment available at that phase of life - not something fixable by another product tweak.

decision-making-culture

Qasar's reading philosophy is to deliberately read old, well-regarded books outside his existing knowledge, on the theory that time filters out noise and that unfamiliar domains produce the most new framing.

He describes picking his next book by identifying a whole area of human knowledge he knows almost nothing about (citing how he chose SPQR to learn Roman history) and finding the best available book in that space, rather than reading recent releases or staying within a comfortable domain - arguing this diverse-input approach is analogous to why diverse training data makes an LLM's understanding richer.

reading-for-taste

Books referenced

House of Huawei: The Secret History of China's Most Powerful Company - Eva Dou - Applied Intuition read this as a company; Qasar cites it as evidence Huawei isn't a normal profit-driven company but an extension of the Chinese state (about a quarter of its ~200,000 employees are Communist Party members).
Maintenance: Of Everything, Part One - Stewart Brand - Referenced by Lenny on the undervalued importance of maintenance work; Qasar ties it directly to his own obsession with operational hygiene (cleaning the office, well-tuned systems) as a founder discipline.
The Autobiography of Malcolm X: As Told to Alex Haley - Malcolm X and Alex Haley - Qasar's example of a 'wacky' recommendation that changed his framing on life, illustrating his read-old-books-outside-your-domain philosophy.
High Output Management - Andrew Grove - Named as the obvious/classic pick Qasar deliberately avoids leading with, preferring less predictable recommendations.
The Emperor of All Maladies: A Biography of Cancer - Siddhartha Mukherjee - A book Qasar was reading at the time that he says changed how he thinks - his bar for what makes reading worthwhile.
Made in America - Sam Walton - Qasar's pick for most influential founder book, written by Walton on his deathbed.
My American Journey - Colin Powell - Named alongside Walton's book as a formative read, though not on Qasar's public reading list.
Guns, Germs, and Steel: The Fates of Human Societies - Jared Diamond - Cited as part of Qasar's interest in connecting human history from 'cave people' to the present tech industry.
Collapse: How Societies Choose to Fail or Succeed - Jared Diamond - Companion pick to Guns, Germs, and Steel, same author, on the same reading list.
SPQR: A History of Ancient Rome - Mary Beard - Example of how Qasar picks his next book: identify a whole area of history he knows little about (Roman history) and find the best book in it.
A World Appears: A Journey into Consciousness - Michael Pollan - Raised by Lenny in the context of a trend that emotions, not just intelligence, may be what produces consciousness - discussed alongside whether AI/LLMs could become conscious.

Companies

Applied Intuition - Qasar's company: a $15B AI company that adds autonomy software to cars, tractors, planes, submarines, and mining vehicles for 18 of the top 20 automakers, major construction/mining/trucking firms, and the Department of Defense - built for a decade with almost no public visibility.
Y Combinator - Qasar was COO of YC before starting Applied Intuition; drew on seeing hundreds/thousands of companies to form his views on early traction and founder resets.
Huawei - Used as the central example that Chinese tech giants shouldn't be benchmarked against US companies like Apple, because Huawei functions as an arm of the Chinese state rather than a for-profit company.
Rivian - Used as the counter-example to 'Chinese EVs are so much better than Detroit' narratives - a US EV maker with a good product that still loses money and is undervalued because it's judged as a real business, unlike subsidized Chinese EV makers.
Tesla - Used to illustrate the low-sensor, cheap-compute L2++ approach to autonomy (versus Waymo's sensor-heavy approach), and as Qasar's own personal example of trusting self-driving day to day.
Waymo - Contrasted with Tesla's approach as the sensor-and-map-heavy autonomy strategy that performs better in geographically constrained areas.
Google - Qasar's account of working there during Facebook's rise, used as a case study for why market-leading companies with the best talent still can't out-execute a structurally different competitor.
General Motors - Where Qasar started his engineering career; source of his firsthand experience with large-organization bureaucracy that later shaped his leadership philosophy.
Bosch - Qasar's second engineering employer before business school and his first startups; another data point in his 'spend time as a low-level employee' argument.
Berkshire Hathaway - Cited as the model Applied Intuition consciously emulated over the 'Silicon Valley darling' build-in-public model.

Techniques and frameworks

Radical pragmatism - Applied Intuition's summary description for all of its company values; the underlying test for any piece of founder advice, including staying quiet, is whether it's practically justified, not aesthetically appealing.
Move fast, move safe - The literal wording of Applied Intuition's speed value; used to operationalize decisiveness as a specific, assessable, compensated behavior for managers.
Values derived from what's already working - Qasar's method for writing company values: once you have early traction, write down the actual five to ten reasons you're succeeding, and those become the values - not an abstract philosophical exercise.
Best idea wins / mandatory dissent - Applied Intuition's culture for surfacing disagreement: everyone, regardless of seniority, is required to voice their one specific piece of relevant experience before a decision is made, removing ego from whose idea it was.
Removing emotion from decisions - Both Qasar and Lenny describe versions of stripping emotional/ownership bias from a hard call - asking what the decision would be if nobody's feelings were at stake, then separately handling the people impact.
Read old, unfamiliar books - Qasar's reading heuristic: prioritize time-filtered older books over new releases, and deliberately pick topics you know nothing about rather than staying in familiar territory.

Summary

Qasar Younis, co-founder and CEO of Applied Intuition - a roughly $15 billion company that adds autonomy software to cars, tractors, mining rigs, planes, submarines, and other vehicles for 18 of the top 20 automakers plus major construction, mining, trucking, and defense customers - joins Lenny for a wide-ranging conversation prompted by Qasar's recent, uncharacteristic decision to join Twitter after building the company almost entirely out of public view for nearly a decade. His central argument is that the AI conversation is overweighted toward software and underweighted toward the physical world: the biggest impact over the next five to ten years will land in farming, mining, construction, and trucking, because those industries already have decades of engineering built into their machines and need only have intelligence layered on top. He frames this as arriving "just in time" rather than as a threat, citing the aging farmer workforce (average age in the late 50s) and the fact that dangerous jobs like long-haul trucking and mining are already going unfilled, not being stolen from willing workers.

Qasar spends significant time defusing AI anxiety by separating two distinct phenomena that get conflated: public fear about job loss (which he attributes to unfamiliarity, urging listeners to actually study the technology's limitations) and stock market sell-offs (which he attributes to hedge funds pricing in the risk that AI-generated demos threaten incumbent software businesses, a separate and more mechanical dynamic). He's bullish that self-driving technology in particular is already statistically safer than human driving, and predicts that within a couple of decades society will view human driving the way it now views child labor - a practice tolerated out of necessity that eventually becomes unthinkable, given roughly 30,000 annual US driving deaths.

A second major thread covers Qasar's deliberate anti-visibility strategy and the company culture behind it. Applied Intuition's early, informal value was "our best work is done alone and quietly" - every hour spent on public content is an hour not spent on customers and product. He's careful to caveat this advice: it worked for him partly because he already had a Silicon Valley network from being COO of Y Combinator, so founders without that starting position may genuinely need visibility to recruit talent and investors. He connects this to a broader operating philosophy - remove emotion from decisions, let the best idea win regardless of who's most senior, and derive company values retroactively from what's actually driving success rather than picking them abstractly. He illustrates organizational blind spots with his own experience at Google watching it fail to compete with early Facebook despite vastly more resources, arguing that accumulated momentum in one direction can make an organization structurally unable to hear that the market has shifted.

The conversation also covers a pointed rebuttal to comparisons between US and Chinese tech companies, using Huawei as the example: with roughly a quarter of its employees being Communist Party members and its stated purpose oriented toward national ambition rather than shareholder profit, Qasar argues it should be benchmarked against a government, not a company like Apple - a framing he extends to arguments that Chinese EV makers have "beaten" Detroit, using money-losing but well-regarded Rivian as the fairer American comparison point. The episode closes on Qasar's reading philosophy (deliberately read old, well-regarded books in domains you know nothing about, because time filters out noise) and a candid take that most Silicon Valley founders lack real taste because they've never worked as a low-level employee inside a large, dysfunctional organization - an experience he credits with shaping how he now builds and leads Applied Intuition.

Notable Quotes

"The real impact of AI in the next five to ten years really is going to be in farming, in mining, in construction, in self-driving trucks." - Qasar Younis

"Our best work is done alone and quietly." - Qasar Younis

"You're not comparing companies to companies. This is not apples to apples... imagine instead of thinking OpenAI is competing against DeepSeek, you say OpenAI's competing against the Chinese government." - Qasar Younis

"A man who cannot command himself is not fit to command others." - Qasar Younis

"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." - Qasar Younis