CHAPTERS
- 0:00 – 3:26
Professor Jensen and the Whiteboard
David Senra explains that he reread The NVIDIA Way to extract only how Jensen runs his company. He covers Jensen as a teacher and his insistence on the whiteboard as the main tool in meetings, tied to constant reinvention.
- •The episode strips the book down to about twenty ideas on how Jensen works
- •Jim Sinegal's line on spending ninety percent of your time teaching
- •Colleagues call Jensen 'Professor Jensen' for his whiteboard explanations
- •The whiteboard forces rigor and transparency because there is no place to hide
- •Michael Dell's view that a company must reinvent itself before a new competitor does
- 3:26 – 9:47
Complacency Kills
Senra describes Jensen's fear of complacency and his inner voice that says nothing he does is good enough. He compares this with Andy Grove, Herb Kelleher, Warren Buffett and Steve Jobs, and gives examples of Jensen's urgency such as saying 'kill Intel'.
- •Jensen downplays his own early history and says NVIDIA wasn't great on day one
- •Andy Grove's mantra: only the paranoid survive
- •Kelleher and Buffett on success breeding complacency and decay
- •Jensen told his company in '97 to kill Intel, which was 860 times larger by revenue
- •Steve Jobs's approach of not dwelling on past success
- 9:47 – 13:18
A Flat Organization with Sixty Direct Reports
Jensen runs a flat organization with sixty direct reports and no one-on-one meetings. Senra explains how this speeds decisions and information flow, and that Jensen refused to hire a COO.
- •Flat structure fights slow decision-making and weeds out those who need direction
- •Jeff Bezos on great people leaving when decisions are slow
- •Jensen compares managing his E-staff to working with smart AI agents
- •He refused board advice to hire a chief operating officer
- •The company is like a race car the CEO knows how to drive
- 13:18 – 17:11
Public Criticism for Company Learning
Senra explains Jensen's habit of criticizing publicly so the whole organization learns from one person's mistake. He gives the NV30 example and compares it with Steve Jobs's lesson to Jony Ive about putting the work first.
- •Jensen criticizes in front of everyone to counter turf protection and politics
- •He called out engineers at an all-company meeting over the NV30 chip
- •Jony Ive's story of Steve Jobs saying he was vain for wanting to be liked
- •Jensen says feedback is learning and the company is optimizing for learning, not for avoiding embarrassment
- 17:11 – 20:33
Tortured into Greatness: Self-Criticism and Suffering
Jensen applies self-criticism to himself and says he would rather torture people into greatness than give up on them. Senra quotes his Caltech and Stanford talks on pain, suffering and resilience as sources of character.
- •Jensen says he looks in the mirror each morning and says 'You suck'
- •Praise is seen as a distraction and looking back as the deadliest sin
- •At Caltech he named endurance of pain as his superpower, not intelligence
- •At Stanford he said high expectations mean low resilience
- •Greatness comes from character formed through suffering
- 20:33 – 23:05
Speed of Light
Jensen judges the speed of work against the theoretical maximum allowed by physics rather than past performance or competitors. Each project is broken into tasks with times that assume no delays, which also guards against internal rot.
- •Speed of light is the absolute maximum speed with no delays, queues or downtime
- •Employees are measured against it, not against what others do
- •Senra compares the idea to Elon
- •It links back to the fight against complacency
- 23:05 – 27:04
Unapologetically Extreme
Jensen is extreme in work ethic, desire to win and recruiting. Senra shares the urinal story, Jensen's low tolerance for complaints about hours, and comparisons to Larry Ellison and Bill Gates.
- •The Kenneth Hurley urinal story
- •No one will outwork Jensen, and working is relaxing for him
- •Olympic training analogy for complaints about long hours
- •Second place is the first loser
- •He told an SGI engineer he would put SGI out of business
- 27:04 – 30:12
The Top 5 Email Idea
Jensen asks employees at every level for emails listing the top five things they are working on and seeing in their markets. This gets him unfiltered information from the edge and lets him catch weak signals early, such as the early machine learning developments.
- •Leaders can be surrounded by people who don't tell the truth, as in Daniel Ek's story
- •Each email is five action-word bullets, tagged by topic for searching
- •Jensen reads about a hundred a day
- •He wants to intercept signals when they are weak
- •An early machine learning signal helped push investment in GPU acceleration
- 30:12 – 33:28
Blunt Communication and LUA
Jensen's communication is short, blunt and direct, like Napoleon's orders and Steve Jobs's demo reviews. When someone rambles he says 'LUA', meaning listen to, understand and answer the question.
- •Jensen's emails are short like a haiku
- •Napoleon: long orders are the enemies of speed
- •Steve Jobs spoke just four sentences in an iPad demo meeting
- •LUA means listen, understand, answer
- •Easy-to-understand ideas are easy to spread
- 33:28 – 36:55
Mission is the Boss, Pilot-in-Command, and Continuous Strategy
Jensen treats the mission as the boss, names a single accountable pilot-in-command for each project, and rejects periodic planning. Senra compares these ideas with Elon's algorithm, Henry Singleton and Michael Bloomberg.
- •Every project has a named pilot-in-command who reports to Jensen
- •Nobody can hide behind a team name
- •Talent is a general pool that can be redirected to new missions
- •Strategy is action, with continuous planning and no five-year plan
- •Singleton and Bloomberg held similar views on flexible planning
- 36:55 – 40:43
Ship the Whole Cow and Go to School on Everybody
Inspired by The Innovator's Dilemma, Jensen repackaged rejected chips into cheaper products to defend the low end. Senra then shows Jensen's learning habits, such as attending a conference himself, and his command of details.
- •Ferrari-grade chips and the fear of low-end competitors
- •Failed-quality parts become cheaper product lines
- •Jensen attended an academic conference himself to learn
- •Mitch Rales taking notes and Bill Walsh on champions behaving like champions
- •Jensen recalled biographical details of all fifty new hires at a party
- 40:43 – 44:14
Create the Market and Choke Them with Gold
Jensen avoids commodity competition by creating new markets, as with robotics, where no customers means no competitors. He also rewards top talent with special stock grants.
- •Rising ASPs while everyone else's fall
- •Robotics: no customers means no competitors
- •Comparison to Edwin Land and instant photography
- •Medici line 'I will choke you with gold'
- •Jensen reviews stock allocations and awards special grants outside annual reviews
- 44:14 – 45:26
Prioritize Highest-Priority Work First
Jensen starts with his highest-priority work so his day is a success before he gets to the office. Senra pairs this with Larry Ellison's view of focusing on a handful of important things.
- •Jensen keeps a clear priority list and starts at the top
- •Remaining time goes to helping others
- •Ellison argues against doing all thousand things
- •Jensen tells others to focus on the most important activity and not be beholden to a schedule
- 45:26 – 55:00
Swarm Your Greatest Opportunity: CUDA and AI
Senra traces how Jensen spotted GPU use beyond graphics in 2002, built CUDA at great cost during the financial crisis, educated the market, and overruled lieutenants to go all in on deep learning. He closes with the book's last paragraph and a plug for his new show.
- •Researchers hacked GPUs for non-graphical work, and Jensen hired the researcher who spotted it
- •CUDA made GPUs easier to program, and making things easier expands the market
- •In 2007-2008 gross margin fell from 45 to 35 percent and the stock fell over 80 percent, but he stayed the course
- •NVIDIA educated the market through classes and a textbook, as Intel once did
- •Jensen overruled lieutenants on deep learning and prefers 'strong, self-reinforcing network' to 'moat'
- •The book ends on adversity as the best teacher, with Jensen saying 'I love Nvidia'
