CEO Management Tactics: Bezos, Thiel, MrBeast

Martin Basiri runs a company where the reward for good work can arrive before the meeting even ends. Basiri, founder of the international-student platform ApplyBoard and now the CEO of a company called Passage, told Shaan Puri over breakfast that when someone on his team solves a problem on the spot, he simply says, "Option drop, 50,000 options," and grants them stock in the company right there (07:51). If it is 10 p.m. and a dozen people are still in the office, he might say it to the whole room at once (08:10). No committee, no quarterly review. Just an instant, concrete signal that the behavior he wants just happened.

That story opens an episode of My First Million built around a simple premise: the people running the biggest companies in the world do things that would get a normal manager reported to HR, and those things often work. Basiri's options still have a strike price, so employees have to buy them, not a free check (09:10). But the psychology behind it is ordinary and well established. Immediate, specific, tangible rewards shape behavior faster than a values poster on the wall (08:29). What is unusual is a CEO willing to hand out equity like a bonus round.

The Machine and the Squeak

The other tactics follow the same shape: strip out delay, force clarity, make the cost of ambiguity visible. At OpenAI, employees can email an internal address, friction@openai.com, to flag anything slowing them down, a policy, a broken process, a pointless approval step (11:47). A staffer is assigned to triage the inbox and clear the blockage. Puri calls it "a janitorial crew for a bureaucracy" (12:13), which is a tidy way of saying that big companies generate their own sludge and someone has to be paid to mop it up.

Siemens' CEO takes the opposite approach to meetings. He answers emails with one word, "okay" or "no," and holds no recurring one-on-ones with his direct reports. If something needs more discussion, the person has to come find him in person, right then (15:02). His explanation is blunt: "I don't entertain people, and they don't need to entertain me" (15:53). Puri's read is that most recurring meetings are a kind of theater, a chance to perform progress for a boss who does not fully trust the team to just do the work, and for the team to perform confidence they do not always have (16:05). Cutting the show does not cut the work. It just removes the audience.

Jeff Bezos ran a version of the same idea at Amazon. Customers could email him directly, and when a complaint looked like a real problem, he would forward it to the relevant team with nothing attached but a single question mark (17:15). Teams initially pushed back with data, noting a 98 percent success rate on whatever metric was in question (18:15). Bezos's answer, as Puri tells it, was that the metric could be correct and still incomplete. A single unhappy customer is not proof the system is broken, but it can be a signal that the system is not capturing something real.

"Amazon is a machine, the machine might be broken, and this anecdote is a little bit of a signal." — Shaan Puri, quoting Jeff Bezos [18:15]

Peter Thiel, running PayPal in its early years alongside people who would go on to build Tesla, LinkedIn, and YouTube, enforced something he called single-tasking: every person on the team owned exactly one problem, whether it was fraud coming out of Ukraine or the virality of the PayPal button on eBay (20:21). If someone tried to bring up anything outside their one problem, Thiel would reportedly just walk out of the room, no explanation needed (21:16). Puri credits Thiel with a phrase he says he still does not fully understand but keeps repeating anyway: "focus is convex." The claim, unpacked a little, is that going from 80 percent focus on a task to 90 percent does not add 10 percent of output. It roughly doubles it (22:14), the same way the difference between the 80th-percentile swimmer and Michael Phelps is not really about swimming a little faster but about capturing a wildly disproportionate share of attention and reward at the very top of the curve (22:54).

Jack Welch's version of ruthlessness at General Electric was more literal. Every division had to be number one or number two in its market within a year, or it got fixed, sold, or shut (24:08). And every year, without exception, Welch fired the bottom 10 percent of performers across the company (24:33). Puri is careful to note this is contested. NVIDIA's CEO Jensen Huang, by contrast, has said he does not fire people; he raises the intensity until people either rise to it or leave on their own (26:05). Two very successful leaders, opposite instincts. The lesson is not that one method is correct. It is that the range of things that can work at the top is wider than most people assume.

MrBeast's approach to training key hires might be the most extreme of the set. New hires in central roles shadow him everywhere, in one case even serving as his literal alarm clock, from the moment he wakes up to the moment he goes to sleep, for six months straight (29:11). By the end, the idea goes, that person can act as a working copy of MrBeast's judgment, freeing him to be effectively in two places at once. Puri recalls a room full of multi-billionaire CEOs reacting to this story with something between admiration and alarm: "I think that's awesome and illegal" (29:45).

What Greatness Actually Costs

The second half of the episode turns the lens around. Puri retells a conversation the writer and Wired co-founder Kevin Kelly had on David Perell's podcast, where Kelly was asked what he gave up to live the life he built. Kelly's answer was direct.

"To have this life I have, I had to give up on greatness." — Shaan Puri, quoting Kevin Kelly [32:22]

Kelly defines greatness narrowly: being remembered by people outside your own family for more than one generation (33:12). By that standard, he says, he almost certainly will not qualify, and he is at peace with that trade. Puri connects this to a broader pattern: the people who reach the very top of extreme fields, sports, invention, business empires, tend not to be balanced, well-rested people. As one of the hosts puts it, "There ain't normal people on the top of Mount Everest" (34:46). Even the winners of ordinary success rarely get remembered. Puri points out that almost nobody today could name the eightieth-richest American of the 1930s (36:21), which suggests that money alone, even a lot of it, rarely buys the kind of memory Kelly is talking about.

Out of that realization comes a distinction the hosts land on as the real takeaway of the episode: achieving greatness is not the same as being great. Building something the world remembers for generations is a long, narrow, and often costly path. Being great, on the other hand, is available constantly, in how someone handles a hard conversation, finishes the last three reps of a set, or responds to a four-year-old at the park who wants to talk about a caterpillar (39:34). "There's a big difference between achieving something great and being great," Puri says (38:24), and he argues the second version is both more attainable and, day to day, more satisfying.

The episode closes on a smaller, quieter observation that complicates all the hardcore stories that came before it: ambition tends to dip for a lot of parents somewhere between their kids turning three and turning thirteen, the window when children are out of diapers, old enough to have opinions, and still willing to spend time with their parents (40:35). The hosts note this is not universal, but common enough that a tweet about it went viral, and that for most people the ambition does eventually come back. It is a useful corrective to the rest of the episode. Bezos, Thiel, and Welch built systems designed to extract more focus and output from other people. But the hosts, both CEOs themselves, end by admitting that they are not living that way, and mostly do not want to. The stories are worth collecting not because they are a blueprint, but because each one quietly asks the same question the elephant chained to a small stake never gets to ask: what exactly is stopping me from doing that, and is it actually as fixed as it looks (02:27)?