Jake | The People Manual

Nvidia shares fell more than 8% after hours on Wednesday after reports that its board was weighing Jensen Huang’s future. That selloff raises a question every founder-led company should ask before a crisis forces it to answer. How much of the business can run without the person everyone calls when the stakes are highest? As of Nov. 12, Huang has publicly acknowledged being unfaithful and pledged to cooperate with Nvidia’s independent investigation. The board said on Nov. 8 it had no reason at that point to believe national security or company secrets had been compromised. Independent directors have since retained outside counsel, and Reuters reported that the board met to weigh Huang’s future. The outcome is still publicly unconfirmed. Three ways to read what’s happening. 1. Jensen Huang and the blind spots of power In lab experiments, people put in a high-power mindset were less likely to spontaneously take someone else’s perspective. In a founder-led company, there’s another mechanism worth considering. The more authority someone has, the more it can cost a colleague to say, “I think you’re misreading this.” One possibility is that a powerful executive sees a decision as private and manageable while overlooking how it looks to people with less power or more to lose. Motivated reasoning can reinforce that gap. Once you want to believe a situation is under control, the evidence that lets you keep believing it gets easier to accept. Huang’s motives are unknown. His apology shows that he publicly accepted responsibility for his infidelity. The board has a different set of questions to answer. What was disclosed, when, and to whom? Did anyone have a safe way to raise a concern? If the separately reported relationship with a direct report is substantiated, could that employee or others on the team have questioned a decision without risking their careers? 2. Nvidia investors and the price of an unknown successor Reuters reported that some Nvidia customers fell while chip rivals AMD and Broadcom rose. One concern investors raised was whether a successor would keep financing AI infrastructure customers on the same terms. Decision neuroscience makes a useful distinction here. In lab studies, choices with missing probabilities can engage valuation systems differently from choices with known odds. That’s ambiguity. Investors can model the effect of a smaller financing budget. It’s harder to price who will set that budget, which commitments they’ll honour, and how counterparties will react. A crowded trade can amplify the move. Some investors may sell because their view of Nvidia’s strategy changed. Others may cut exposure because volatility rose or because they hold the same AI bets elsewhere. The price alone can’t tell you which is driving the move. If Nvidia gives customers and investors credible answers on leadership, the product roadmap and financing commitments, some of that uncertainty could ease. If customers start delaying commitments or the company changes its funding strategy, the selloff could start to reflect a change in expected growth. Those are different developments to watch. 3. Nvidia’s board and the risk of deciding under pressure Organisational research suggests that threat can narrow the information a group considers and concentrate control. Neuroscience research also finds that acute, uncontrollable stress can impair working memory and flexible thinking. In a boardroom, the practical risk is that one urgent question crowds out the others. Directors could lean too heavily on decades of success with Huang. They could also give Wednesday’s share price too much weight. I’d ask every independent director to write down, before debating, what evidence would justify keeping him, placing him on leave, or replacing him. Then ask who could run Nvidia tomorrow under each scenario, with explicit authority over the chip roadmap, customer commitments and capital allocation. Nvidia’s independent directors have already hired outside counsel. They can test the succession plan while that investigation proceeds. Customers need to know who can make a commitment. Engineers need to know who can settle a roadmap decision. The board needs to know whether those answers hold if Huang is unavailable. @elonmusk, I’d be curious where you think this read is wrong. If Jensen were unavailable for 90 days, which would be hardest for Nvidia to carry forward: the technical roadmap, the trust of its biggest customers, or the judgment about where to put billions of dollars next? What would you look for to tell?

Psychology or Nonsense? "Small talk is killing your happiness." That claim came from a real study. Then the scientist behind it helped knock half of it down. In 2010, University of Arizona psychologist Matthias Mehl and colleagues published "Eavesdropping on Happiness" in Psychological Science. They had 79 college students wear an audio recorder for four days. It caught 30 seconds of sound every 12.5 minutes, 23,689 clips in total. The numbers looked damning for small talk. The happiest students had about a third as much of it as the unhappiest (10.2% of their conversations vs 28.3%). They also had twice as many deep conversations (45.9% vs 21.8%). That result got plenty of attention outside the field. In 2018, Mehl's own team, led by Anne Milek, checked it again. They pooled the original students with three new groups: women with breast cancer and their partners, healthy adults in Atlanta, and people who had just separated from a spouse. 486 people in all. Small talk? Nothing. Pooled across everyone, its link with life satisfaction was r = .05, a correlation you can't tell apart from zero. The authors said the original small-talk result looks like a false positive. The deep-talk half held up. People with more substantive conversations were more satisfied with their lives (r = .23, and still .18 after accounting for personality). A bit smaller than first reported, but there. Verdict: half nonsense. Small talk is off the hook. Having too few real conversations still looks bad. One honest catch. Both studies are correlational, so nobody knows the direction yet. Happier people might just pull others into deeper chats. Example: a coworker asks "Good weekend?" Don't skip it to seem deep. Answer it, then go one layer down: "Pretty good. What's been taking up most of your head lately?" #psychology #Neuroscience

Psych myths #5: "Don't worry, I'm good at multitasking." Nearly 70% of students in a University of Utah study rated themselves above-average multitaskers. Their confidence didn't predict their real score at all. The paper is "Who Multi-Tasks and Why?" by Sanbonmatsu, Strayer and colleagues, PLOS ONE, 2013. 277 students reported how much they juggled media and how often they used a phone while driving. Then they took the Operation Span, a test where you solve math problems while holding a string of letters in memory. Two jobs at once, scored for real. 193 of the 277 said they were better than average at multitasking. Only 54 admitted to being below. Somebody's wrong. Worse, the pattern ran backwards. The people who multitasked the most tended to score lower on the test, not higher. Same for phone use behind the wheel. And the strongest personality link was attentional impulsiveness, meaning trouble holding focus on one thing. The authors' read is uncomfortable. A lot of multitasking may come from struggling to block out distractions. Skill doesn't look like the driver. Example: a colleague answers Slack all through your project call and swears they caught every word. In this study, that kind of certainty told you nothing about actual ability. Try this on your next call. Close the inbox tab and say, "I'm off email for this one so I actually hear you." Then notice how much more you remember afterward. That's my advice, not something this study tested. Honest limit. This is one correlational study of college students, and the links are small. A 2018 PNAS review found many media-multitasking studies show no difference at all. Nobody has shown that multitasking causes weaker focus. #psychology #neuroscience

"That's nice" sounds supportive. In relationship research it counts as a bad answer. UCLA psychologists filmed 79 dating couples for a 2006 paper in the Journal of Personality and Social Psychology called "Will You Be There for Me When Things Go Right?" Each partner talked through one recent good thing and one recent bad thing. Then the person sharing rated how understood, validated and cared for they felt. Outside coders rated how the listener actually reacted. You'd think the real test of a relationship is the crisis. Who shows up when you lose the job. Two months later, the good news conversations told the researchers more. How partners handled the good news was more closely tied to commitment, satisfaction and love than how they handled the bad news. For the women, feeling heard after sharing good news strongly predicted how the relationship was doing at follow up. Feeling heard after bad news barely added anything. Only 4 couples broke up before follow up, so treat this part as a hint. Still, all 4 scored below the typical couple that stayed together on how their partner met good news. The team sorts reactions into four kinds. Their own example is a lawyer who just got the firm's biggest case. Engaged: "Wow, this is great news!" plus a real question about the case. Flat: "That's nice, dear." Undermining: "Are you sure you can handle it?" Ignoring: "What do you want to do for dinner?" In Gable's earlier 2004 work, only the engaged kind went with better relationships. The other three went with worse ones, and that includes the polite, flat one. The authors' reading is that a flat reply tells someone the news didn't matter much to you. Maybe they didn't either. Next time someone tells you something good, stop what you're doing and ask one real question. "Wait, how did that happen? Tell me everything." That line is my suggestion, not something the study tested. Honest limits. This is correlational, with young dating couples at one university. And a 2022 study found muted replies weren't linked to worse relationships in its Mainland China samples, so culture changes the picture. #psychology #neuroscience

"That's nice" sounds supportive. In relationship research it counts as a bad answer. UCLA psychologists filmed 79 dating couples for a 2006 paper in the Journal of Personality and Social Psychology called "Will You Be There for Me When Things Go Right?" Each partner talked through one recent good thing and one recent bad thing. Then the person sharing rated how understood, validated and cared for they felt. Outside coders rated how the listener actually reacted. You'd think the real test of a relationship is the crisis. Who shows up when you lose the job. Two months later, the good news conversations told the researchers more. How partners handled the good news was more closely tied to commitment, satisfaction and love than how they handled the bad news. For the women, feeling heard after sharing good news strongly predicted how the relationship was doing at follow up. Feeling heard after bad news barely added anything. Only 4 couples broke up before follow up, so treat this part as a hint. Still, all 4 scored below the typical couple that stayed together on how their partner met good news. The team sorts reactions into four kinds. Their own example is a lawyer who just got the firm's biggest case. Engaged: "Wow, this is great news!" plus a real question about the case. Flat: "That's nice, dear." Undermining: "Are you sure you can handle it?" Ignoring: "What do you want to do for dinner?" In Gable's earlier 2004 work, only the engaged kind went with better relationships. The other three went with worse ones, and that includes the polite, flat one. The authors' reading is that a flat reply tells someone the news didn't matter much to you. Maybe they didn't either. Next time someone tells you something good, stop what you're doing and ask one real question. "Wait, how did that happen? Tell me everything." That line is my suggestion, not something the study tested. Honest limits. This is correlational, with young dating couples at one university. And a 2022 study found muted replies weren't linked to worse relationships in its Mainland China samples, so culture changes the picture. #psychology #neuroscience

The famous "because" trick failed its rerun. And one kind of reason did worse than none. Harvard's Ellen Langer tested it at a New York copier in 1978. Ask to cut in for 5 pages with no reason and 60% said yes. Add "because I have to make copies," a reason that says nothing, and 93% did. Most retellings skip the second row. At 20 pages, that empty reason got 24%. Exactly the same as no reason. In 1985 Valerie Folkes reran it at UCLA and Sussex. No magic. Between 83% and 93% said yes with no reason at all. What moved people was the kind of reason. "Because I feel really sick" got 18 of 19. "Because I don't want to wait" got 54% in her biggest run, below the 83% who heard nothing. So stop leaning on the word. If you need to cut in, name the thing you can't control: "Sorry, can I jump in? My train leaves in ten minutes." That line is my advice, not a tested script. Small samples, one task, decades old. #psychology #neuroscience

That "any normal person would do the same" feeling is lying to you. Stanford psychologists Lee Ross, David Greene and Pamela House caught it in action and gave it a name in 1977: "the false consensus effect." The setup was awkward on purpose. They asked 80 Stanford students to walk around campus for 30 minutes wearing a sandwich board. Some boards said "Eat at Joe's." Others said "Repent." Saying no cost nothing. Students kept their course credit either way. After deciding, each student guessed how many of their peers would agree to wear it. The ones who said yes guessed 62%. The ones who refused guessed 33%. Same request. Same campus. Both sides were sure they were the normal ones. It gets worse. The students who refused also made more extreme calls about the personality of someone who agreed. Your own choice feels like plain common sense. The other choice starts to look like it says something about the person. And this one didn't quietly die in the replication crisis. In 2018, Many Labs 2 reran two other scenarios from the same paper, in a project spanning 36 countries and territories, with more than 7,000 people each time. Both held up, and both gaps came out bigger than in the original. You've seen this at work. Example: you skip the optional Friday drinks and assume most people would too. A coworker goes every week. You quietly file them as trying too hard. They've filed you as not a team player. Neither of you ever checked. Next time you think "obviously anyone would," treat it as a guess. Then ask one person likely to disagree: "Would you have made the same call? Be honest, I'm checking my own bias." And when someone's choice strikes you as weird, notice what share of people you think would make it. That number is the one this research says is probably bent toward you. One honest limit. Using your own choice as a clue about others isn't crazy, since you are one data point. The mistake is leaning on it too hard. And these were students guessing about students. #psychology #Neuroscience

Reading the thank-you letter changed nothing. Meeting the student who wrote it more than doubled their work. University of Michigan researchers led by Adam Grant ran this in a real university call center. The paper is "Impact and the art of motivation maintenance" (2007). The callers phoned alumni for donations, and part of that money paid for student scholarships. Rejection all day, and they never saw who it helped. Some callers read a one-page letter from a scholarship student. Others read the letter and then met him for about 5 minutes. A month later, the letter group hadn't moved. The 17 who met him went from about 108 to 261 minutes a week on the phone, and from $186 to $503 raised. Small study, 39 callers. A later field experiment with nurses at an Italian hospital pointed the same way. If you lead a team, quit forwarding the praise email. Bring in one person your work actually helped and ask them, "What changed for you after we fixed it?" #psychology #neuroscience

Who's harsher on your mistakes, you or everyone else? Researchers at the University of Mannheim keep getting the same answer. It's you. They call it the "beautiful mess effect." The first paper, in the Journal of Personality and Social Psychology in 2018, ran seven studies. Then Anna Bruk, Sabine Scholl and Herbert Bless tested it again in "You and I Both" (Personality and Social Psychology Bulletin, 2022). In one study, 101 German university students imagined confessing a mistake to their boss. They also imagined a coworker confessing the exact same mistake. Then they rated the confession on items like "showing courage" versus "showing weakness," and whether other people would find it "repellent" or "desirable." Same mistake. Same boss. Their own confession scored 4.65 out of 7. The coworker's scored 4.92. That gap isn't huge. But the same lopsided verdict showed up for confessing love and for revealing a physical flaw, across all four studies. People were quietly kinder to someone else's mess than to their own. And it wasn't spread evenly. The gap was biggest in people low in self-compassion, the ones who beat themselves up. In people who treat themselves kindly, it mostly vanished. Neuroticism didn't explain that. Picture it (an example, not from the study). You sent a client the wrong price list. You burn the afternoon drafting a cover story, because owning it feels like proof you're bad at your job. If a colleague owned the same slip, you'd probably just think they were honest. Try this before you hide something. Picture a coworker you respect making the same mistake and telling the boss. Rate them. Give yourself that same score, then say it plainly: "I got the client figures wrong. I've already sent the correct ones, and I'm adding a check so it doesn't happen again." That line is my practical suggestion. The studies didn't test it. The honest limit: these were imagined scenarios rated by students, not real bosses reacting to real confessions. Nobody tracked anyone's career afterward. Some workplaces really do punish mistakes, so read yours. #psychology #neuroscience

You're holding back thank-yous because they feel awkward. The people you'd thank disagree. Chicago Booth researchers Amit Kumar and Nicholas Epley tested it in "Undervaluing Gratitude" (Psychological Science, 2018). MBA students wrote a short thank-you letter to someone who'd helped them, then guessed how that person would react. Then the researchers asked the people who got the letters. The writers lowballed it. Across three experiments, recipients felt better than predicted and less awkward. The mood gap was big (d = 0.74 in the first study). And that fear quietly decides who gets thanked. Of 98 people picking who to write to, exactly 1 chose the person they expected to feel awkward. You obsess over the perfect words. They mostly notice that you bothered. Pick one coworker whose help you never mentioned and send two lines: "I never told you this, but when you [specific thing], it saved me. Thank you." Small samples, mostly students, and letters rather than hallway chats. Epley's lab reports 17 later replications, though. #psychology #neuroscience