A courtroom verdict on the AI lawsuit of the year lands with a big basically nothingburger — and separately, a Microsoft CEO just handed the crypto crowd their best argument of the decade without trying to. Welcome to Tech Podcast Podcast — we listened so you could skip straight to the signal. Today we’ve got Sreeram Kannan on a16z crypto, Cerebras at a reported $95 billion IPO valuation, and Liz Lopatto’s read on what the Musk-Altman lawsuit actually produced. And then Eric Ries shows up to explain why good companies go bad, which hits a little harder when IBM and SAP have been hovering over this conversation all week. Big claims hit the wall all week. Today we see which ones are still standing. a16z crypto, with Sreeram Kannan:
AI is intelligence accelerator. Crypto is coordination accelerator. The Microsoft CEO Satya is asked in a podcast, is AI kind of like its own species? And then he laughs and says, no because at the end an AI cannot own property, cannot take liabilities, cannot enter into contracts, cannot start companies. But that's exactly what a blockchain is intended for.
All week we’ve been asking which layer of the agentic stack has an actual structural case behind it — and Sreeram Kannan on a16z crypto is the first guest who came in with Satya Nadella’s own words as the setup. Nadella says AI can’t own property, can’t take liability, can’t enter contracts — Kannan’s move is to say blockchain closes exactly those three gaps. Okay, that is a genuinely sharper frame than anything we heard Monday through Wednesday. If autonomy needs ownership, contracts, and liability rails, Kannan is saying crypto gives you the rails. Fine. But then he calls it "sci-fi forward" and puts a three-year clock on it, so I want one thing: does he have a deployed self-sovereign agent that has actually signed a contract or taken on liability, or is the three-year number just vibes with a calendar attached? That’s the deterministic-execution problem we’ve been circling all week — if an agent gets "almost right" on a contract, that’s not an enterprise software inconvenience, that’s a liability event. Kannan naming the constraint is a real step up from the chapter-header AI takes we’ve been cataloging, but the gap between "blockchain enables this in theory" and "an agent has done this in production" is still wide open. And it doesn’t sound like the host pushed him on a single deployed example — which, honestly, has been the pattern all week. Somebody finally names the seam, and then the follow-up just vanishes. From Peter H. Diamandis at Peter H. Diamandis:
In this episode, the mates welcome Andrew Feldman, Co-founder and CEO of Cerebras Systems, and discuss several tech news such as Google’s I/O comeback, the jury verdict in Elon Musk’s OpenAI lawsuit, Anthropic’s accelerating enterprise momentum, and a long interview with Andrew Feldman of Cerebras after its major IPO.
Diamandis episode 256 is doing a lot of curation in one sitting — Google I/O as a "comeback," Karpathy landing at Anthropic, and Andrew Feldman in the room talking about a $95B IPO valuation for Cerebras. Three distinct inflection points, and at least one of them has a number concrete enough to actually stress-test. The Cerebras number is the one I want to pull on. Feldman is literally sitting there — co-founder, CEO, post-IPO — and the question is whether $95B is priced on real agent inference demand, or whether this is just a founder on a press tour with a friendly room. We’ve been building the Taiwan chip argument all week — five layers, from fabrication through customer qualification — and a US chip company going public at $95B is either evidence the silicon-shield disruption thesis is getting priced in, or it’s just a momentum number with nothing to do with that thesis. Did they actually name which problem Cerebras solves in that stack? And Karpathy joining Anthropic is in the same headline. That’s not decorative — if the architecture question all week has been "why Anthropic specifically for code generation," Karpathy showing up there is a real personnel signal. Diamandis calling Google I/O a "comeback" is editorial verdict, not reporting. Karpathy and the $95B number are the two things I’d actually time-stamp. Nilay Patel, writing in The Verge:
Both Elon Musk and Sam Altman are big personalities, and people have a lot of feelings about both of them and the AI industry. And in the end… nothing happened! The jury found that Elon had filed his lawsuit after the statute of limitations had run out. You’ll hear Liz explain exactly what’s going on there.
We’ve had the Musk v. Altman lawsuit as background noise all week — and now The Verge sent Liz Lopatto to sit through the whole trial, and the headline is literally "much ado about nothing." The jury didn’t even get to the OpenAI conversion question — they tossed it on statute of limitations grounds. So the most-covered AI governance dispute of the year produced zero legal verdict on whether OpenAI’s nonprofit-to-for-profit conversion was legitimate. That’s actually a meaningful data point: all that litigation pressure, and we learned nothing actionable about the governance structure. It mostly just generated podcast content. The one thing Lopatto’s reporting does confirm is that the courtroom was a "zoo" — protests outside every day, big personalities, high drama — and structurally the whole case dissolved before it touched the actual OpenAI conversion story. If you wanted a legal answer on whether Musk was wronged when OpenAI went for-profit, you didn’t get one. Honestly, the statute of limitations exit ramp is almost the most interesting part — it means nobody ever had to argue the merits. Which leaves OpenAI’s governance conversion exactly as contested as it was before the trial started, just with a lot more courthouse zoo footage. Here's Startup Hustle:
Eric Ries wrote The Lean Startup. Sold 2 million copies. Helped hundreds of people build companies from nothing. Now he's back with a harder question: why do so many of those companies eventually go bad? In this episode, Matt Watson sits down with Eric to talk about his new book Incorruptible—a deep dive into the invisible forces that corrupt organizations, why profit-maximization becomes pathological, and what it actually takes to build a company that stays great.
Eric Ries, Lean Startup, two million copies — now he’s got a new book called Incorruptible, and the core argument is that profit-maximization itself becomes the pathology. That’s a specific structural claim, not just "culture matters." It’s interesting timing — all week we’ve been asking whether IBM and SAP’s operational gravity is a moat or a trap, and Ries is basically saying the moat is the trap: the same profit engine that built the company is what eventually corrupts it. The Costco origin story is the concrete anchor here — Sol Price, a betrayal, a company built explicitly around fiduciary duty to the customer. That’s not a bumper sticker, that’s a named founding decision. The question Matt Watson should have pressed on: does the stage-gate model failure Ries describes map onto any specific company that went through the Lean Startup framework and still went bad anyway? Right — and if the answer is "yes, even lean companies go corrupt," then the book is doing something harder than Lean Startup did. If the answer is "well, they didn’t follow the framework properly," that’s just a sales pitch for Incorruptible dot co before May 26th. Got thoughts on today’s briefing, a story we should be tracking, or a correction? Send us a note anytime at techpodcastpodcast at lantern podcasts dot com. We read every message.
You’ll find links to every story we covered today in the show notes, so if something deserves a closer read, that’s the place to pick it back up.
That’s Tech Podcast Podcast for this Friday. Thanks for listening, and have a good weekend. This is a Lantern Podcast.