Trump's War on "Artificial Intelligence," Elon v. Scott, and Google's Wall Street Meltdown

The episode opens with Kara Swisher and Scott Galloway joking about Halloween, Scott’s apparent hangover, and their upcoming Pivot tour before turning to President Trump’s effort to discourage government use of the term “artificial intelligence.” Kara says Trump called AI a “very bad and inaccurate name” and prefers “superintelligence,” which she notes has the unfortunate abbreviation “SI,” also associated with Sports Illustrated. They mock the move as a bizarre political priority while AI companies are filing enormous IPOs, committing hundreds of billions of dollars to computing, and openly discussing the possibility that advanced systems could destroy humanity. Their running joke is that renaming the iceberg will not stop it from hitting the ship. Kara says they will deliberately use “AI” throughout the episode, partly because the letters appear in many ordinary words. They also announce the Pivot Post newsletter at pivotshow.com, which will include answers to listener questions, essays, tour photos, and behind-the-scenes material. The hosts discuss writing as an unusually difficult form of “mental fitness,” comparing finishing a book to childbirth because the pain somehow fades enough for people to do it again. They say they will experiment with the newsletter rather than turn it into another long weekly column.

Kara introduces Elon Musk’s attack on Scott after Scott described Tesla as the “patient” and SpaceX shareholders as the “organ donors” in a discussion of a possible Tesla-SpaceX merger. Musk responded on X that Scott “never had any clue.” Scott says the attack validated his position, while Kara points out that Musk has responded to even relatively minor criticism with threats and insults. Scott says he has called Musk a “menace,” an “addict,” and a “traitor to the West.” When Kara asks why he used “traitor to the West,” Scott explains that he believes Musk is undermining the military capability and the wishes of Americans who want the United States to help defend Europe against Russian aggression. He offers a historical analogy: if a Boeing CEO during World War II had secretly disabled B-17 bombers or their guidance systems while they were being used against the Nazis, that person would be viewed as a traitor. Kara notes that Vladimir Putin praised Musk, which Scott treats as an especially poor character reference. They briefly joke about whether Scott’s Deadpool costume could protect him from Musk’s alleged sword threats, but the substantive conclusion is that Musk’s attacks do not answer Scott’s concern about related-party transactions and shareholder interests.

Kara and Scott turn to the approaching midterms, beginning with Alaska. Kara describes an investigation ordered by Republican Governor Mike Dunleavy into reports that Democratic operatives recruited a candidate named Dan Sullivan in an effort to split the incumbent Republican senator’s vote and help Democrat Mary Peltola. Trump has called for Peltola to be prosecuted, while the Justice Department has sued to block certification of results in another Sullivan contest. Kara also criticizes Trump’s proposed financial rewards for reports of voter fraud, including a possible $1 million reward leading to an arrest. The hosts see momentum for Democrats. They cite polling showing Democrats ahead 53% to 44%, a RealClearPolitics generic-ballot lead of 8.3 points, Decision Desk HQ giving Democrats a 79% chance of winning the House with about 238 seats, and Kalshi putting the probability of a “blue tsunami” at 57%. Scott defines a tsunami as at least 235 House seats and 51 Senate seats, a blue wave as 218 House seats with Republicans retaining the Senate, and a “mega tsunami” as 240 House seats and 54 Senate seats. Kara praises Democratic candidates such as Amy Acton and says Alexandria Ocasio-Cortez’s phrase “common sense, not communism” is an effective response to Republican attacks. They remain cautious because polling can miss Republican voters, citing Susan Collins races as an example. Scott says Trump is offering voters a $5,000 “Trump dividend,” payments to farmers, rewards for fraud tips, and a promise to settle Camp Lejeune water-contamination claims if Republicans win. Kara is more alarmed by Trump’s apparent cognitive and behavioral decline, new rape allegations, and the possibility of chaos between Election Day and the seating of new lawmakers. Scott says Trump seems to believe he has two years to “molest the country for as much money as possible,” with his lawyers assuring him he will not go to prison.

Microsoft beat quarterly expectations, with revenue up 19% and Azure cloud revenue up 48%. Capital spending was $45.2 billion, below expectations, and the shares rose nearly 4%. At the same time, Microsoft offered voluntary buyouts to roughly 60,000 employees while saying it would continue investing in AI. Alphabet’s revenue grew 25%, but it missed quarterly expectations for the first time in almost four years. Free cash flow was negative by $9 billion, long-term debt increased almost sixfold to $125 billion, and cloud revenue more than doubled to $30.7 billion. Search advertising revenue grew 12%, below expectations, and the stock was headed toward its largest post-earnings decline. Scott says Microsoft succeeded because it offered Wall Street growth without an even larger spending shock: Azure was growing quickly while capital expenditures came in lower than expected. Alphabet, by contrast, spent $51.4 billion on capital expenditures in one quarter—more than the entire quarterly revenue of Google Cloud—while its negative cash flow and debt increase raised questions about whether AI returns will catch up with investment. He describes the market as dividing workers and computing infrastructure into a K-shaped economy: humans receive buyouts while GPUs receive enormous budgets. The hosts also discuss the scale and mechanics of Microsoft’s buyout. Scott says companies calibrate offers based on how many people they want to leave, often using a generous package alongside the implicit threat that employees who decline could later be laid off with worse terms. Kara says remote workers and younger employees can be overrepresented in layoffs because they are easier for managers to treat as abstractions. Both favor making difficult personnel decisions promptly and treating departing employees generously rather than repeatedly inventing excuses for poor performance.

Kara says she has been using Claude more than Google for search and describes a particularly damaging problem with Gemini: when she asked who won the 2020 election, it would not give a straight answer, even though other AI systems answered the question. Google reportedly attributed the behavior to a technical gap and said it was fixing the issue. Kara calls the timing especially bad because Google needs its AI search products to justify huge capital expenditures and reassure users that the system can state settled facts. Scott says the problem leaves Google with two bad interpretations: either the product is not competent enough to answer an obvious factual question, or it is “sus.” He argues that a search product refusing to state the winner of the 2020 election undermines the company’s broader AI narrative. Kara says Alphabet’s financial results show that the search business may be shifting even as the company’s AI costs rise. Gemini’s app reportedly reached 1 billion users, but that scale does not resolve questions about accuracy, trust, or whether users are actually moving away from traditional Google search. The hosts connect the product failure to the financial story. Microsoft appeared disciplined enough to convince investors that AI spending could coexist with earnings growth; Alphabet looked more exposed because it paired massive spending and rising debt with weaker search-advertising performance. Their conclusion is not that Google’s AI effort is doomed, but that Wall Street is demanding visible returns and users are increasingly able to compare Google’s answers with competitors such as Claude.

Kara and Scott examine Nvidia’s $3 billion investment in Meta as part of an expanded AI partnership, Meta’s plan to raise nearly $100 billion through stock offerings, and AMD’s separate $1 billion investment in Meta. Kara says Nvidia is also taking stakes in companies including SpaceX, OpenAI, and Anthropic, making the arrangements feel like Cisco in the dot-com era. She also mentions a report that U.S. intelligence agencies believed a Chinese spy who did business with Nvidia CEO Jensen Huang targeted him, and that Trump was briefed but did not act. Scott describes the structure as a textbook circular-financing arrangement: Nvidia invests in Meta, Meta raises money partly to buy Nvidia chips, and AMD invests in Meta’s equity raise. He compares Nvidia to a shovel seller that invests in the miners and owns part of the bank financing the miners’ purchases. The central question is whether Nvidia’s demand is organic or partly self-funded. Kara recalls similar “virtuous circles” among 1990s startups that invested in one another and bought one another’s software; she called those arrangements “circle jerks” at the time. Scott says the arrangement could work if companies such as Anthropic continue scaling revenue and producing returns. He notes that Anthropic’s revenue per employee reportedly rose from $1.4 million to $6 million, putting it behind only Nvidia, and says the technology could still be transformative even if the market corrects sharply. But if Nvidia’s demand signal begins to look artificial, or if its stock falls and it can no longer finance the ecosystem, the interconnected deals could unwind quickly. His investment advice is to remain in the market while diversifying, especially for older investors, because railroads and the internet transformed the economy despite severe speculative busts.

Former Yahoo CEO Marissa Mayer faced backlash after saying she used her AI assistant, Dazzle, to identify three things that made her 10-year-old daughter special. Dazzle uses a person’s camera roll to learn about their life and preferences. Mayer later said she was testing the application rather than relying on AI to understand her daughter, but Kara says the original quote was a communications disaster because people expect parents—not an app—to articulate what is special about their children. Kara says Mayer has an awkward communication style and that, although she is not a fan, this particular episode may have been a stupid statement rather than evidence of malicious intent. She recalls her own critical coverage of Mayer’s time as Yahoo CEO, arguing that Mayer was in over her head, that Yahoo’s financial performance deteriorated, and that she nevertheless left with roughly a quarter-billion dollars. Kara also revisits the earlier episode involving a Yahoo executive who falsely claimed a computer-science minor and received substantial severance after being fired. Scott agrees that Mayer’s wording was tone-deaf, especially when many parents are already worried about AI’s effect on children. They say technology executives should be especially cautious about presenting AI as a substitute for parental attention. Kara jokes that an AI can help generate a bedtime story in a parent’s voice, but both hosts reject the idea that a chatbot should be the source of a parent’s observations about a child.

For his prediction, Scott discusses Meta’s new Instagram Features program, which will let creators upload videos up to 90 minutes long. He says Meta is trying to capture the value YouTube generates from long-form video while leveraging Instagram’s enormous user base. Kara says Instagram Reels are unusually addictive because the next recommendation is so often “on point.” Scott argues that Meta’s AI-driven recommendation system measures enormous numbers of behavioral and psychological signals to keep users watching for another second. Scott predicts that Nielsen will report streaming data for Meta by the end of the following year. He says Instagram has begun moving onto smart TVs, where it can show both short-form Reels and longer Features, bringing Meta into direct competition with Netflix, YouTube, and other streaming services. He cites viewing-share figures in which streaming accounts for roughly half of television use, with cable at 19% and broadcast at 18%; within streaming, YouTube has reached about 16% and roughly twice Netflix’s viewing time. He also notes that Tubi has around 2% of streaming, more than Peacock’s 1.6%. Kara questions how Meta will attract creators because YouTube shares substantial revenue with them, while Meta is known for taking control of uploaded material and has been fighting with publishers over rights to scrape content. Scott nevertheless predicts Meta will target YouTube first. Kara says YouTube remains less organized and less addictive than Instagram, while Scott believes the flow of attention is moving from computers and televisions toward phones. They close by directing listeners to submit questions, subscribe to Pivot Post, and attend the upcoming live tour.