Sharp Tech with Ben Thompson

Ben and Andrew react to a week of Anthropic discussion, including Dario Amodei’s leaked memo to employees, why a compromise is still possible, and answering a variety of questions in response to Ben’s article this week. At the end: A terrible AI law for young parents, surveying the implications for Netflix after Paramount wins the Warner Brothers bidding, a dispatch from dating app hell, a question about feeds on chatbots, should Google be the model for ChatGPT ads?, and thoughts on the business of F1 and the new season.

Ben and Andrew discuss the global memory shortage and answer a listener’s question: how did this happen? Topics include: What memory chips have in common with logic chips, why Intel left the memory market in the 1980s, how the international shipping market explains today’s shortage, how major players will address the problem going forward, possibly with some help from the Chinese. From there: A look at the “thin” future of AI computing as memory and performance takes precedent, AI computing that will still be done on device, and Apple’s AI strategy. At the end: Responses to Shopify concerns in the AI era, a question about the future of brands as AI finishes what digital advertising started, Shopify preferencing its own AI tools, and a looming test of Apple’s commitment to the Vision Pro.

Ben and Andrew react to a killer round of earnings for Spotify and Daniel Ek’s final earnings call, including thoughts on Spotify’s transformation of the music industry, how a record company oligopoly helped create the definitive tech bundle, and why Spotify’s personalization requires an addendum to aggregation theory. From there: The difference between AI spending at Google and Amazon, why the AI buildout should be funded by debt, un-optimized tech companies, and the logic and risks of Amazon’s AI future. At the end: Reactions to a viral essay on AI job loss, a note about the Princeton Law Review, an ad man strikes back after Ben’s Anthropic criticism, and thoughts on Giannis’ investment in Kalshi.

Ben and Andrew react to a bloodbath for public Saas companies with thoughts on the future of software in the AI era, beginning with why companies choose to outsource solutions to Saas companies today, and why those moats may be more durable than skeptics think. Then: Why SaaS skepticism remains fair, including an analogy to the newspapers in the ’90s, the absence of anti-fragility, a variety of headwinds that will impact pricing power, start-ups with superior cost structures, and looming consolidation and layoff questions. At the end: The biggest SaaS company of them all and what Microsoft’s roadmap should look like, a response to data center skepticism, supply and demand for hyperscalers, why Ben hated the Anthropic Super Bowl ads, should AI hallucinations be good case law?, and a Vision Pro announcement.

Unpacking the latest round of Meta earnings, including Wall Street’s about-face after last year’s CapEx squeamishness, whether Zuckerberg’s astronomical CapEx plans are more evidence he yearns to be more than an app maker, why Meta owes a thank you to Apple, Apple and Meta in the AI era, and a word about Instagram messages. Then: Are we in an “AI is a Bubble” bubble? Thoughts on mass adoption among software makers, demand that looks insatiable, product managers vs. engineers, and the era of perfect competition among employees. From there: Why hyperscalers should not solve the CapEx problem by co-investing in TSMC, why Ben sympathizes with TSMC, and a note on Samsung. At the end: Andrew shares his experience with Bucks-Lakers in the Vision Pro and reviews Ben’s takes.

TSMC’s pricing power in the AI era, a brief history of TSMC’s culture and CapEx decisions, and ongoing capacity constraints that should be pushing tech companies to build up competitors. Then: Thoughts on Netflix after Ben’s interview with co-CEO Greg Peters, including Wall Street’s concerns despite enormous success, whether and how the Warner Brothers acquisition could be a counter to YouTube, and the difference between Netflix content and user generated YouTube content. At the end: Questions about the Curt Cignetti of tech, a victory lap on OpenAI, advertising in chatbots, advertising as a force for good, and Andrew’s Starbucks habit.

A call for Apple to finally be confident in its Vision Pro hardware, a brief history of broadcasting sports on TV, and yet another reminder that immersive live experiences should be the killer use case that brings users to the Vision Pro platform. Then: Q&A on the AVP, a question on the Siri and Gemini partnership, and thoughts Ben’s interview with United Airlines CEO Scott Kirby and on how tech spearheaded the company’s revival. At the end: Questions on Meta Compute and the end of Aggregation Theory, Grok and its offensive bikini problems, and follow-ups on Baumol’s Cost Disease, humans wanting humans, streaming economics, and venting about the Green Bay Packers.

Andrew and Ben return from the holidays to discuss Ben’s Article AI and the Human Condition, and various responses to the preponderance of pessimistic forecasts for what AI will mean for the future, including thoughts on employment, sex, and the problem with trying to regulate human nature. Then: An email about OpenAI spawns discussion of cultural assumptions, market incentives vs. social incentives, and tech as an amoral force. At the end: Unpacking the logic of Nvidia’s deal with Groq, a regulator’s own-goal, questions on streaming TV vs. music, sperm racing, and advice for a listener debating whether to embrace suburban living.

Andrew and Ben begin with reactions to ChatGPT’s new image capabilities, a reminder of OpenAI’s strategic advantages vs. Google, Disney’s deal with Sora, and Gemini 3 Flash. From there: Netflix and its competition for attention, Netflix continues its foray into podcasting, and a question about movie theaters highlights costs that Netflix will have to internalize going forward. Then: Extended thoughts on SpaceX and the possibility of data centers in space, while a listener does some field reporting on AWS usage. At the end: Strategies for a successful remote work life, Tesla and Rivian’s aversion to CarPlay, the new United app and developer trade-offs, oenophile preferences, Taco Bell, Christmas traditions, and an attack on Andrew for hypocrisy.

Andrew and Ben talk through Netflix’s proposed $72 billion deal to buy Warner Brothers, including the logic for Netflix, the frictionless nature of competition on the internet, and the threat that Netflix sees from YouTube. Then: David Zaslav’s windfall, and an argument about the regulatory questions that may scuttle this deal. At the end: The better business model between YouTube and Netflix, an emailer wonders why everyone is mourning the end of a Hollywood business that can’t succeed in the modern marketplace, groupchats and the college experience, and a word about Flighty.

Thoughts on OpenAI as Sam Altman declares a “Code Red” in response to Gemini 3, including real concerns about ChatGPT’s market position, why the missed ads opportunity is becoming more acute, and ominous Google history. From there: Context on Alan Dye’s departure from Apple, Meta’s emphasis on a new design language, and the Meta fundamentals regardless of AR/VR and its AI efforts. At the end: Amazon’s tranium chips and its AI efforts, Substack and the challenge of customer acquisition, defending tech at Thanksgiving dinner, and various F1 takes before this weekend’s finale in Abu Dhabi.

The ways tech has improved suburban life, why urbanism may have peaked, and the first and second order effects of Tesla's full self-driving technology. Then: A Thanksgiving mailbag! Topics include: A note from the Gemini team, a correction on Llama 4, a hater’s question on the utility of ChatGPT group chats, Ben and Andrew share their daily use cases for AI and Ben shares his prompt, while an emailer wonders about a boss using Copilot to send holiday gifts. From there: A subject matter expert debates whether to train his AI replacement, a listener seeks advice about working for his father-in-law, a question about balancing work and young children, Unidentified Anomalous Phenomena, robot vacuums, Ben's ancestry, and a word about the Giannis era.

Ben and Andrew begin with Gemini 3, what to make of its terrific benchmark results, and why TPUs provide Google a sustainable cost advantage in AI. From there: Google’s opportunity in the enterprise space, Apple’s white label deal, and questions about both OpenAI’s future growth and challenges that loom if ChatGPT can’t incorporate advertising. At the end: Thoughts on what Gemini means for Nvidia, Anthropic’s market in AI, why Amazon and OpenAI are losers of the Anthropic-Nvidia-Microsoft announcement, and a correction regarding Charter cellular coverage.

Andrew and Ben analyze SpaceX's nearly $20 billion in purchases by first touching on cell carrier history and the power dynamics that iPhones upended 20 years ago. Then: Understanding the SpaceX business and Musk's approach to strategy, what Starlink is trying to do with satellite internet on airlines, a power play with cell carriers that appears to have failed earlier this year, and now, a Plan B that may involve an acquisition and a bid to partner with Apple. At the end: Why Yann LeCun leaving Meta is the right outcome for both sides, a question about big companies and innovation spawns regulation cautionary tales and a cigar anecdote, and wondering about the impact of big tech on AI's future.

Ben and Andrew begin with reactions to the OpenAI CFO discussing a federal "backstop" for prospective financing, as well as Sam Altman's recent comments about OpenAI's spending. Then: An emailer objects to the discussion of Bubble benefits, and questions about Meta's AI spending and a looming the AI backlash as hiring contracts and electricity prices rise. From there: Unpacking the announcement that Apple will use Gemini to power Siri, and two follow-ups to last week's discussion of "Too Big to Fail" in tech. At the end: Thoughts on Amazon's grocery ambitions, Walmart's continued success, the YouTube TV-ESPN dispute, and a listener's mental model of Ticketmaster.

Discussing Ben's interview with Substrate CEO James Proud, including the "insane" challenge he's undertaken as Substrate attempts to compete with TSMC and ASML, and the ways in which a bubbly environment benefits innovation by incentivizing exactly that sort of moonshot. From there: Ben's thoughts on the "Too Big to Fail" era in tech that may be averted, reactions to the latest humanoid robot, and thoughts on both sides of the OpenAI-MIcrosoft announcement earlier this week. At the end: Taylor Sheridan leaves Paramount for greener pastures, Sharp Text and lessons from Grantland, a temperature check on Sora one month later, lessons from the Nexperia mess in Europe, a question about Magic: The Gathering, and a few corrections to last week's show.

Celebrating the return of the NBA with reactions to Inside the NBA on ESPN before turning to an extended explanation of the technology underlying the AWS outage this week and the history of US-East-1 in Northern Virginia. Then: Grappling with the trade-offs inherent to investing in resiliency to preserve the status quo, the risks that preservation comes at the expense of innovation, and Twitter as an object lesson. At the end: Questions on F1's deal with Apple, an F1 hater checks in, a sales pitch for YouTubeTV, and Ben chooses his favorite karaoke songs and explains how he learned Chinese.

Andrew and Ben begin by examining the various structural forces and business decisions that led the U.S. and the West to cede rare earth mining and refining to China, including reduced friction at the expense of resilience n a variety of areas, predatory pricing that pays extra dividends in the commodities business, and why Ben is keeping an eye on the Middle East in the months to come. From there: Reactions to the news that Spotify and Netflix are partnering and pulling podcasts off of YouTube, with thoughts on the limits of that strategy and the difference between YouTube's cut of creator ad revenue and Apple's developer fees. At the end: Uber, Tesla, and Waymo and the shape of the future autonomous vehicle marketplace, Ben's thoughts on Tesla's Full Self-Driving, OpenAI's corporate structure, and ChatGPT embraces AI erotica.

Thoughts on OpenAI after a month of infrastructure and partnership announcements, including the differences between OpenAI and a Mag7 company, Ben’s interview with Sam Altman this week, Apps in ChatGPT, and drawing on Windows, Apple, and WeChat to better understand the company’s strategy. At the end: Questions on Sora, the 90-9-1 law, and OpenAI’s approach to copyright, as well as some brief reactions to China's move to escalate curbs on rare earth exports.

Andrew and Ben begin with reactions to OpenAI's Sora 2, a new Sora app, and more thoughts on last week’s ‘Vibes’ release from MetaAI. Topics include: Parallels between Sora 2 and the GPT 3.5 release in 2022, responding to a sample of disgusted MetaAI 'Vibes' reactions, why OpenAI is investing in short form video, why the threat to Meta is clearer than ever, and fair questions about Mark Zuckerberg's leadership after the last several years. At the end: TikTok’s business prospects and security concerns, solar power possibilities for AI infrastructure, Ben's shocking embrace of the iPhone Air, and a Sharp Tech x Oreo crossover.