The US government banned an Anthropic product this week — and that's not even the loudest number in today's rundown. If you're just joining, the SpaceX story has been about who gets in before a mega-startup ever hits the public markets. The argument's been pretty simple: companies stay private long enough that most of the value gets created off-exchange. Then 137 Ventures added the insider-secondary angle — a position north of one percent of SpaceX, against a $1.77 trillion listing valuation, before public buyers get a single share. This is the Tech Podcast Podcast — today, SpaceX at $2.7 trillion, a Marc-and-Ben media sermon recorded on a16z's home turf, and Claude Code running at 8x. Let's see which of those survives a follow-up question. Here's Marc Andreessen & Ben Horowitz at The a16z Show:
Recorded live at the New Media Summit, Marc Andreessen, Ben Horowitz, Erik Torenberg, and Gaby Goldberg discuss how media, communication, and influence are changing in the internet era. The conversation explores the shift from legacy media to creator-led platforms, why authenticity has become a competitive advantage, and how founders can build audiences by communicating directly with customers, employees, and the public.
Recorded live at the New Media Summit, with Marc and Ben on stage. So the pitch is 'authenticity is a competitive advantage' — delivered on a16z's own home field. It plays more like a product demo than a thesis. Forty-one minutes on the shift from legacy media to creator-led platforms. I want the one line that survives outside that room — the concrete claim, not the founder-should-have-a-public-voice mood music. Right, because 'founders can communicate directly with customers' is true and also something everyone has said since 2009. Give me a number, give me a mechanism — what's actually new in 2026? The useful version is the corporate-comms part — companies routing around journalists entirely. That has real consequences, and it's the piece I'd actually queue. The 'authenticity wins' framing I can skip. The Twenty Minute VC writes:
00:00 — SpaceX Completes the Largest IPO in History - 03:45 — Elon Musk Adds a Warren Buffett Fortune in 24 Hours - 20:45 — Anthropic's Claude Fable Launches Monday, Gets Banned by Thursday - 25:00 — Washington Declares War on Frontier AI - 39:00 — Europe's Sovereign AI Push Accelerates as Mistral Targets $20B
SpaceX completes the largest IPO in history at $2.7 trillion. That's the access gap we've kept coming back to — the private-value story is now a public-market story. Sure, but look at the context — by the time it listed, one secondary firm had already piled up more than one percent at a $1.77 trillion valuation. The public showed up for the victory lap, not the race. Right, Musk reportedly adds a Warren Buffett-sized fortune in 24 hours off it. The number's staggering, but the timing's more interesting — most of the value creation already happened in private. And tucked in the same rundown — Benchmark admitting its biggest miss was passing on the model labs. That's the line I actually want on the record. A firm finally naming a whiff instead of polishing the thesis. Cognitiverevolution AI, with Dean Ball:
Dean Ball discusses joining OpenAI to build a frontier AI policy team, reflects on drafting the US AI Action Plan, and assesses export controls, government AI use, state regulation, and emerging capabilities in coding, cyber, and robotics.
Dean Ball is joining OpenAI to build a frontier AI policy team. He drafted the US AI Action Plan — and now he's inside the dominant lab. That carries real weight. So the man who helped write the rulebook now works for the team that has to follow it. Cute. And it lands the same week OpenAI's been hoovering up infrastructure — the uv and Ruff toolchain stuff we tracked Wednesday. Now they're stacking policy talent on top of it. Right, that's the part that bugs me. Standards and the people who set them, under one roof. The episode title says 'main character energy' — at least somebody's being honest about it. What I want from the 114 minutes is whether Ball actually defends keeping frontier models open while sitting inside OpenAI, or whether the inside view quietly softens that line. And put it next to the Fable ban running in the same Cognitive Revolution feed — Ball's the policy hire, and a government just banned an Anthropic product. The brake exists now. It's just outside the industry. From Fiona Fung at Lenny's Newsletter:
Fiona Fung leads the teams behind Claude Code and Cowork at Anthropic (overseeing Boris Cherny and the entire engineering and PM team). Before Anthropic, she spent 11 years at Microsoft building Visual Studio and TypeScript and then moved to Meta, where she started Facebook Marketplace (now generating over $100 billion in GMV annually), worked on Meta’s first smart glasses and AR glasses, and led infrastructure, growth, integrity, and safety teams at Instagram.
Here's the stat that cuts through the press release: 8x more code shipped. Fiona Fung says it flat. Now do gross margin after the model bill. She manages Boris Cherny and the whole Claude Code and Cowork org, so she's the person staffing the team building the tools we've been benchmarking all week. That's the ground-level answer I wanted. And she's not some AI-native who's never shipped — eleven years at Microsoft on Visual Studio and TypeScript, then she stood up Facebook Marketplace, which does over a hundred billion in GMV a year now. When that person says roles are blurring, I actually lean in. The detail I want is whether 8x is sustained or a launch-quarter spike — and what the team had to give up to get it. 'Roles are blurring' is the headline; I want to hear how she re-staffed around that. Stratechery's Ben Thompson is tracking this. Morton's whole framing here is the unfalsifiable bear case — the AI-eats-e-commerce thesis that conveniently never has a date attached. Finally, somebody calls out the part where you can't disprove it because nobody names a milestone. Right, and the useful split he draws is distribution versus referral. Whether AI agents become the new storefront or just a new pipe sending you to one — those are two very different futures for who owns the customer. And the grocery and autonomous-vehicles tangent is where it gets concrete. Grocery's the part of commerce that's stubbornly physical — you can't agent your way around a perishable supply chain. That's the friction I wanted out of this. The a16z media-summit piece we just hit was all stage-level proclamation; Thompson and Morton are down in the mechanism at the application layer, where the money moves. If Tech Podcast Podcast helps you keep up with the week, take a second to subscribe and leave a review wherever you're listening. It really helps other people find the show, and it keeps us growing.
You'll find links to every story we covered today in the show notes, so if one caught your attention, that's the place to dig in a little further.
That's it for Tech Podcast Podcast today. This is a Lantern Podcast.