Advertise on podcast: Sharp Tech with Ben Thompson

Andrew and Ben begin with reflections on Ben’s changing relationship with the Mac and two markets that may ultimately be impossible to serve with one product and operating system. Then: Why Apple’s plans for your living room make sense, reactions to the Duo, and the near-term outlook for humanoid robots in the home. From there: More specifics on why agents put certain types of advertising at risk, thoughts on Amazon adopting a YouTube Premium model, and a question about product pricing that spawns a lesson in advertising, generally. Then: Reactions to OpenAI’s math progress, anxiety among mathematicians, inter-disciplinary skills, joining forces on Machine Intelligence Island, and a reminder that equating raw IQ with value is a bad idea. At the end: Grading Ben Affleck’s AI knowledge, Mike Tomlin’s Minecraft habit, and a digression on IBM, Compaq, and the decompilation of video games (and why fair use is not a viable defense for hackers).

Ben and Andrew begin with a note on VIM and reactions to a confusing product launch from OpenAI at Dev Day earlier this week. Then: Will personal agents be a winner-take-all market? What can Apple and Google do in the consumer agent space? Will normal people take longer than expected to adopt these products? And did Meta’s push into the enterprise break the company’s hot streak? From there: Emails on the future of payments, “super intelligence” branding for AI, America.gov, and firing Dario Amodei. At the end: AI Show & Tell with notes on generational adoption shifts (and Tamagotchis), movie games, a personalized sports page, App Store hell, and going full KGB on messaging.

Ben and Andrew begin with the Meta Connect keynote and a simple question: is this the first time since 2021 that Meta has been cool? From there, new VR glasses, the Muse Charm, a basketball analogy for Meta Connect in 2026, and the infrastructure and talent investments that have helped Muse resonate over the past few weeks. Then: An emailer shares his Muse epiphany, a potential secondary effect of agents saving users money, why Muse is not sharing user data with Meta, and the ultimate “aggregator of aggregators” upside. Then: Extended thoughts on the Muse vs. Amazon tensions, why Meta’s Walmart deal was potentially crucial, and Amazon’s moat in logistics (and returns) that may prove durable in the agents era. At the end: A CarPlay victory lap, and a round of AI Show and Tell with a wedding planner, a SunOS coder, and a fantasy football league.

On today’s Ben and Andrew begin with Andrew’s article on AI alarmism, the primacy of the effective altruists in AI risk conversations, and why these discussions are exhausting without more evidence, even as AI risk becomes the biggest story in American politics. From there: A distinction between regulation malaise five years ago and where we are today, the grind of modern software, and questions responding to Ben’s Update on Monday, including thoughts on why OpenAI and Anthropic won’t unilaterally slow down, why Sam Altman and Elon Musk were incentivized to agree with Dario Amodei, and an emailer who asks whether doomerism hysteria is crowding out more urgent economic questions. Then: Questions on China’s incentives in AI, the strategic benefits of chip sales and distillation, military implications of American lead in AI, and China’s messy cybersecurity situation. At the end: An ode to ApplePay, Walmart, and Walton Goggins (but not Matt LaFleur).

Andrew and Ben begin with Ben’s article on Tuesday on the transformative value of LLMs writing things down. Topics include: How assistants have helped Ben solve his blindspots, the omnicompetent assistants that are now available to everyone, a brief detour into markdown file history, the arrival of OpenAI’s Astra models, a future of models that use computers, the astounding value of a pro subscription to Anthropic and OpenAI, and the pace of AI progress in the fall of 2026. From there: More on the Hugging Face incident, a listener asks for advice on getting started with agents, and Meta’s Muse arrives to help normal unlock the potential of agents. At the end: Reactions to a pair of viral p(doom) tweets from Anthropic employees, why one writer’s pandemic callback offers an unintentionally instructive parallel, and an emotional announcement about the iPhone Air.

On today’s show Ben and Andrew begin with notes on a home networking project before turning to the release of Fable 5.1, including Anthropic’s announcement that it has walked back the data retention safety initiative the company announced earlier this year. From there: Reactions to a viral post on the OpenAI-Hugging Face incident, the problem with anthropomorphizing machines, and why Ben finds this discourse unhinged and unproductive. At the end: Follow-ups to last week’s Meta discussion on the tension between libertarian and conservative impulses, Shopify is a platform and not an aggregator (and that’s great), a well-timed question about 3-D Printing, an F1 app PSA, a listener’s girlfriend hates vibe coding, and apropos of last week, a pronunciation email.

On today’s show Ben and Andrew begin with initial reactions to Meta’s settlement with 29 states, including behavioral remedies that will impose restrictions on teens using Facebook and Instagram. From there: Reactions to a report that Nvidia is set to acquire HuggingFace, including why a robust open source ecosystem looks increasingly urgent for Nvidia, and why OpenAI’s surprisingly impressive jalapeño chip underscores the point. At the end: Netflix pivots from a longstanding refusal to funnel customers to rivals, F1’s declining ratings on AppleTV, ESPN abandons the bottom line for college football, Stanley Druckenmiller uses AI for his Wall Street Journal op ed, a vibe coding question (and more adventures with Ben’s over-engineered psychosis), a coffee pot question, parameter sizes and the mysteries of model performance, and a hurling PSA yields an appreciation for various sports on TV.

On today’s show Andrew and Ben begin with reflections on big Apple news that feels relatively small alongside a world of growth and dramatic questions across technology in the AI era. Then: They turn to the news that OpenAI is pausing its Internet model training, the incentive structure that makes cybersecurity challenging and how AI compounds the challenge, two digressions from Ben on misalignment and a new AI use case, and why black hat and white hat cybersecurity actors are both doing the same thing in the AI era. At the end: More thoughts on bubble benefits that may survive, an emailer considers a financial planning career but is concerned about AI, two follow-ups on AI and the writing process, and European emailer thinks Americans who love sports love everything but the actual sport.

Ben and Andrew begin with Nvidia’s announcements of a new funding model for AI infrastructure, including the differences and similarities with railroad expansion 150 years ago, why LLMs were a gift and curse to Nvidia’s business, the pressure on Nvidia coming from Google and Amazon, and the expanded blast radius as Nvidia works to mobilize third party funding. From there: Why the turnover at Google may actually be a good sign for Google’s frontier efforts, and extended thoughts the future of AI-generated output, ideas and substantiation, and Anthropic’s plan to watermark outputs. At the end: A question about the Vision Pro, the obstacles for Starlink Mobile, and a cranky emailer yields a clarifying answer on why Ben is excited about his vibe coded app.

On today’s show Ben and Andrew begin with Microsoft, Meta and Google in the wake of their latest earnings reports. First: Microsoft’s push to be the middle layer for enterprise AI, threats to the company’s competitive position in the long run, and the subtext of the open weights open letter frenzy two weeks ago. Then: Parsing Google and Meta strategies as Google hedges its frontier bets, while Meta continues its frontier spending and announces questionable plans for an enterprise business. At the end: The future of the lending environment for Meta and the other hyperscalers, a bubble check-in, a rundown of what happened between Citadel and the Situational Awareness, close encounters with Ben’s vibe coded app, and an emailer offers a compelling theory on the push for Permanent Daylight Savings Time.

On today’s show Ben and Andrew begin with takeaways from OpenAI’s sandboxing snafu, including a reminder that the big bad wolf is real, OpenAI’s failures, and the timing challenge with warnings from AI labs. From there: Talking through Ben’s article on Kimi K3 and the future of open weights AI models in the U.S., including why the U.S. open source market has stagnated, the motives of various tech constituencies to drive competition with OpenAI and Anthropic, the difference between the training market and the inference market, intelligence as a commodity, and why OpenAI and Anthropic may have a more defensible lead than it seems. At the end: The government’s disagreements about how to handle Chinese models, the argument for encouraging U.S. distillation, the security risks of building atop Chinese infrastructure, an email about Chinese innovation, concerns over American credentialism, and a potential curveball from the CCP after a week of Kimi anxiety.

On today’s show Ben and Andrew begin with a raft of OpenAI news, including reactions to the first reports of what OpenAI’s first hardware product will be, the logic of OpenAI’s hardware business generally, the Codex push in the mac app, and a request for a Codex computer. From there: The Apple-OpenAI trade secrets lawsuit, including Apple history, a reckless OpenAI employee, and the funniest part of the complaint. At the end: IBM’s stock and the specific way AI may have affected mainframe spending, how AI will manifest in real-world software, an email about Meta’s messaging as a recruiting pitch, notification management, monetizing expertise, and a word about daylight savings time.

On today’s show Andrew and Ben begin with a look at the state of Meta. Topics include: Mark Zuckerberg’s sins of commission vs. omission as a messenger, Meta’s AI opportunity, directionally correct investments, the problems with Meta as a cloud provider, and the absence of religion in Menlo Park. From there: Why Microsoft should move on from the XBOX era, and the shift in gaming habits that doomed Game Pass from the outset. At the end: OpenAI introduces GPT-Live, a question about the cost of Ben’s vibe coding adventure spawns a digression on future token costs, the cost of youth sports, American soccer and learning Chinese, tech weirdos and the future of normie app building, and Ben gets castigated for bringing a Starlink on vacation.

Ben and Andrew begin with an extended look at the market for memory chips, including market history, the supply crunch driving up today’s valuations, the myopia from SK Hynix, Samsung and Micron, new entrants from China, and why the timing of Apple’s price increases are a bearish signal. From there: Both hosts share their daily caffeine regimen, garage takes, and lots of thoughts on vibe coding after Ben’s 10-day foray into AI-assisted engineering. At the end: Codex marketing and an imminent OpenAI super app, Sam Altman’s appearance in a game of mafia, could the EU push Apple to better business strategy, stock buyback clarification, ocean data centers, metricmaxxing, and an update on last week’s hydration break take along with a better fix for international soccer.

On today’s show Ben and Andrew begin with a hydration break take as the World Cup continues before turning to the latest on Anthropic’s export control saga, what the Trump administration may not understand about AI cybersecurity, and what Anthropic still doesn’t understand about dealing with the government. From there: Fables differentiation, the dangers of ringfencing frontier AI within U.S. borders, and how open source Chinese AI may affect the marketplace. From there: Why Ben gives Fox the benefit of the doubt after its Roku purchase this week, whether Siri AI will nuke ChatGPT’s market, Noam Shazeer joins OpenAI, what e-commerce will and won’t be automated, and whether autonomous driving can scale on ICE vehicles.