By playing, you agree to uListen's Terms and Privacy Policy and YouTube's Terms of Service.

$13 Billion CEO Thinks Most People Shouldn't Start A Company

Today I walk with Shishir Mehrotra, CEO of Superhuman, the company formerly known as Grammarly.

Shishir MehrotraguestLuba Yudasinahost

Oct 1, 20261h 15mWatch on YouTube

Summaries, chapters, transcripts, and FAQs on this page are made by uListen, not YouTube.

CHAPTERS

Jump to publisher timestamps

Chapter summaries

Why ownership and control can distract from getting an idea into the world
Shishir Mehrotra frames a core belief: if you have a great idea, the priority should be getting it to the world as fast as possible, not maximizing ownership or control. He also previews his view that many people shouldn’t start companies without deep obsession for a specific idea.
- •Grammarly is described as much larger and broader than people assume, beyond “just grammar.”
- •A founder mindset can show up inside existing platforms, not only in new startups.
- •Being overly attached to ownership and control can lead entrepreneurs astray.
- •Starting a company is portrayed as a painful path compared to other ways of shipping an idea.
Lego robotics with his daughters: high performance, small teams, and “hive mind”
Shishir explains seven years of Lego robotics with his daughters, including making world championships multiple times. He draws parallels between team size in competitions and effective team structures inside companies.
- •The Lego robotics competitions involve timed tasks solved with Lego-built robots.
- •Their team stayed small, typically a handful of kids, and still reached world championships repeatedly.
- •He coached the team and describes an intensely “Lego-obsessed” household routine.
- •He connects success to smaller, more functional teams rather than large, chaotic ones.
- •He describes an internal ritual at Superhuman: hackathons where teams above five become coordination-heavy.
Growing up in an academic family and chasing many possible careers
The conversation shifts to Shishir’s upbringing in a highly academic family with frequent moves due to his father’s career. He describes being naturally curious and repeatedly imagining different future professions, including briefly wanting to be a taxi driver.
- •He grew up in the US and moved multiple times as his father progressed through academia and later worked at NASA.
- •He was drawn to many possible careers, often changing his mind as a kid.
- •He considered being a taxi driver in college to observe life and learn the city, which his mother opposed due to the commitment involved at the time.
- •He values being in roles that expose him to diverse domains and people.
- •The host shares her own experiments doing delivery driving to understand product experiences firsthand, which Shishir supports as a useful ritual.
“Books are at least 50 years old”: choosing fundamentals over vocational learning
Shishir recounts a pivotal college moment: his father threatened not to pay for college if he majored in computer science, pushing him toward disciplines with more enduring foundations. He argues the real payoff is learning how to think, not the specific subject matter.
- •His father objected to computer science as an undergraduate major because it changes quickly and can become overly vocational.
- •The rule given was to major in something where the books are “at least 50 years old.”
- •Shishir ended up with math and computer science, emphasizing fundamentals and thinking skills.
- •He describes “peeling the onion” in deep subjects as a way to learn how concepts fit together.
- •He mentions guiding his children toward fundamentals in a similar spirit.
A startup without planning it: MIT competition momentum and unintentional founding
Shishir explains how his first company began almost accidentally through an MIT entrepreneurship competition during grad school. He emphasizes that it didn’t feel like a glamorous default path at the time, and that momentum from interest and team-building carried them into starting a company.
- •A project in grad school became an entry into MIT’s entrepreneurship competition (the “50K Challenge” at the time).
- •Their strong showing attracted attention and investor outreach, turning a school project into a company quickly.
- •He notes startup culture didn’t feel as compulsory or glamorous in 1999 as people might assume.
- •He describes both companies he started as coming from “fairly unintentional paths.”
- •The host prompts a discussion on intentionality versus curiosity-driven paths.
Why most people shouldn’t start a company: obsession beats “founder identity”
Shishir lays out his strongest view: starting a company should not be the starting goal. He argues that without a deep, persistent obsession with a specific idea, the inevitable roadblocks will knock founders off course and lead to aimless pivots.
- •He challenges people who say they want to start a company but can’t name an idea they’re obsessed with.
- •Starting from scratch is described as the most painful and hardest way to get an idea into the world.
- •He calls the startup impulse a “curse” when an idea won’t leave your head.
- •Weak attachment leads to over-pivoting when customers, investors, or employees push back.
- •He says meaningful things take time and require sticking through hard moments.
The Vinod Khosla email: signaling, speed, and moving the company to Silicon Valley
Shishir tells the story of raising capital after receiving an unexpected email from Vinod Khosla, whom he didn’t initially recognize. He describes how Vinod’s apparent interest changed other investors’ behavior immediately and led to a rapid relocation to Silicon Valley.
- •He received an informal-looking email from Vinod Khosla and initially didn’t know who he was.
- •A Boston investor advised him to look Vinod up, emphasizing Vinod’s prominence.
- •$500,000 appeared in their account quickly, which Shishir attributes to the market reacting to Vinod’s signal.
- •He describes Vinod as think-big, extremely fast-moving, and highly persistent.
- •A condition of Vinod’s involvement was moving the company to Silicon Valley, which happened on a compressed timeline.
What Centrada was building—and a lesson on timing being “too early”
Shishir explains Centrada as a data center automation/utility company with a vision similar to what cloud computing later became. He reflects that the category’s enabling technology wasn’t ready yet, making timing a decisive factor in outcomes.
- •Centrada aimed at on-demand computing, resembling how cloud services are later understood.
- •He says their cohort of companies was “five to seven years too early.”
- •He points to missing enabling technology at the time as a key blocker.
- •He describes how some companies in that space ended up building software for data centers instead.
- •The story is used to underline that execution isn’t the only variable—timing matters substantially.
Joining YouTube when it was losing money per view: the “pirate” team and a bold thesis
Shishir describes joining YouTube after the Google acquisition, when it had a bad internal reputation and serious business challenges. He explains that a small group held a contrarian belief about YouTube’s future and operated with a “pirate, not navy” mentality.
- •YouTube was net negative on every view, meaning more usage made the business worse.
- •The company also faced major legal and perception challenges, and many people didn’t view it as serious.
- •He names a small group of leaders who believed YouTube could become much bigger, including his boss Salar Kamangar.
- •He shares a simple thesis he used publicly: YouTube would do to cable what cable did to broadcast.
- •He contrasts “pirate” team dynamics (entrepreneurial) with more established “navy” operating styles.
Manufacturing signal under existential pressure: milestones, board scrutiny, and TrueView’s purpose
Shishir details the stress of operating under repeated shutdown conversations and the need to define measurable intermediate milestones. He also shares how coaching influenced his decision to stay and finish shipping skippable ads, which later became a major revenue driver.
- •They defined intermediate success milestones, including reaching $1B in revenue and profitability in at least one country.
- •He describes quarterly board meetings with Google leadership where shutdown was openly discussed.
- •He recounts a moment meeting a creator who said they “work on YouTube,” reinforcing the ecosystem responsibility felt by the team.
- •He explains learning from Dean Gilbert, including a three-list framework: goals, marginal utility, and long-term purpose.
- •Skippable ads (TrueView) are positioned as a purpose-driven project, with a theory of rebalancing incentives by charging only when ads aren’t skipped.
- •He notes TrueView went from zero to a billion dollars in about 18 months.
Learning ecosystem thinking from Susan Wojcicki: the Partner Program and revenue sharing
Shishir reflects on what he learned from Susan Wojcicki about building non-zero-sum ecosystems. He uses the YouTube Partner Program decision—opening monetization to creators and setting a revenue split—even while YouTube was losing large sums, as the key example.
- •He says Susan Wojcicki hired him and was his first boss at Google/YouTube.
- •He credits her with ecosystem thinking learned through AdSense and multi-party incentive balancing.
- •YouTube opened the Partner Program beyond major media companies so creators could monetize.
- •They debated and then chose a revenue share model, described as a 55/45 split.
- •He describes a value that emerged from this approach: “even playing fields invite play.”
- •He frames revenue sharing as a long-term flywheel investment rather than a short-term optimization.
From YouTube to Coda: the “cursed with an idea” phase and why he tried not to found it
Shishir explains how Coda began as an obsession during a break after YouTube, with sketches by his bed and a conviction that productivity tools hadn’t meaningfully evolved. He also describes actively trying to place the idea inside existing companies before concluding that no one else would build it.
- •He took a six-month break after YouTube and found it unusually difficult because it felt low-stakes.
- •He started brainstorming with Alex (later his co-founder), initially trying to help Alex find an idea to pursue.
- •Coda’s core observations: productivity metaphors (docs/spreadsheets/presentations) felt antiquated and were being used as “application surfaces.”
- •He initially didn’t want to start another company and tried to get established companies to pursue the idea instead.
- •He says companies wanted to hire Alex but dismissed the product vision as too hard or unlikely.
- •He concluded the “marginal utility” signal was strong: if they didn’t do it, nobody would.
Selling Coda to Grammarly: AI at the edge, founder mindset, and purpose over pride
Shishir explains how Hemant Taneja connected him with Grammarly’s founders, revealing a strategic alignment around AI-enabled productivity. He describes Grammarly as an “AI superhighway” at the edge of where people work and frames the acquisition as a new beginning rather than an exit.
- •Hemant Taneja urged him to meet Grammarly’s founders before closing a financing round.
- •He learned Grammarly had tens of millions of daily active users and over $700M in revenue at the time discussed.
- •The founders’ framing: Grammarly isn’t about grammar; it’s about bringing AI to the edge across many applications.
- •He describes Grammarly as an early “agent” that operates unobtrusively across apps, with infrastructure underutilized by only doing “grammar teacher” tasks.
- •He accepted becoming CEO as a purpose-driven decision, using the same goals/marginal utility/purpose test he learned earlier.
- •He rejects “startup founder pride,” emphasizing outcomes and speed-to-world over ownership.
Bundling theory and Superhuman’s agent-driven future: context, surfaces, and proactivity
Shishir discusses his interest in bundling economics and how it influences product strategy at Superhuman. He then goes deep on AI agents: auto-drafting emails, bringing tools to the “surface where you work,” and the importance of context plus proactive action.
- •He references writing “Four Myths of Bundling” after discussions with Daniel Ek.
- •He describes bundling as minimizing super fan overlap while maximizing casual fan overlap.
- •He lists multiple Superhuman products and acquisitions as pieces of a broader bundle strategy.
- •He shares an AI draft milestone: over 60% of auto-drafted emails are sent, with about half sent unchanged.
- •He explains the trio he wants from agents: ubiquity (where you work), context/knowledge, and proactivity (acting without being asked).
- •He contrasts standalone prompting with in-line collaboration as the step from “useful” to “life-changing.”
- •He describes the engineering challenge of working across many different applications in different ways.
Leadership growth, learning loops, and ending on “the happy version” of an inventor
Shishir reflects on what he’s still learning as a leader, especially as AI changes team dynamics. He closes by describing how he learns—getting outside the building, learning from peers and boards—and shares a personal metaphor about becoming the version of an inventor who makes ideas real in the world.
- •He believes AI will make teams bigger and teamwork problems harder due to mixing human and virtual teammates.
- •He says he’s writing a book called “Rituals of Great Teams.”
- •He identifies a personal growth area: translating optimism into language and accountability systems that work for different personality types.
- •He credits learning through external exposure: customers, partners, peer leaders, and board roles.
- •He names people he learns from and describes a perspective about seeing a person’s reflection in the products they create.
- •He ends with a childhood book series metaphor: wanting to be the “happy version” of an inventor who ensures ideas happen in the world.

Publisher timestamps

Get more out of YouTube videos.

Read along on your phone, search it beside the video, or ask your favorite AI app what was said.