Tiezhen Wang is pairing open-source incentives with data autonomy — and that pitch is either a real business thesis, or the prettiest way I've heard to say, 'we can't raise a Series B.' Tech Podcast Podcast — today Wang gets pushed on whether 'AI bootstrapping' actually means no outside capital, Leo takes another lap around the Musk-Altman courthouse, and John Kim walks Ted Seides through $70 billion of fundraising wisdom that may or may not survive the current lab funding market. After a week of Pentagon contract lines and Cerebras valuations, this one feels softer. So Wang either gives us a real customer or revenue structure, or it's just the ambient philosophy slot. Kimmer is the one I actually want to hear. Seides can sometimes get allocators to say the thing they did not plan to say. Grace Shao, writing in AI Proem:
In this conversation, Tiezhen explains why Hugging Face became the GitHub for models and why open source is not just a distribution mechanism but a different way of coordinating research. We discuss why Chinese AI labs have leaned so aggressively into open models, how DeepSeek changed the commercial logic of open source, and why Qwen, Kimi, GLM, MiniMax, and others are using openness as a way to win attention, recruit talent, and accelerate the whole ecosystem.
AI Proem put out a Tiezhen Wang episode — he was formerly at Hugging Face, where he ran APAC developer relations. His argument is that China's open-source push has three layers: researcher attribution, company-level benchmark marketing, and ecosystem coordination. The part that jumped out to me is the way he links open-source incentives with data autonomy. That's a different claim than 'open weights are good for research.' He's saying openness is a control play. The bootstrapping framing is the thing I want stressed. If Wang means 'open models as a sustainable business' in the sense of no VC subsidy, that's a pretty contrarian claim right now — because every Chinese lab he names, Qwen, Kimi, GLM, MiniMax, is backed. So is this actually sustainable, or is open source just the cheapest customer acquisition channel while the cap table looks like any other lab? And that lines up with what Logan Kilpatrick was saying from inside DeepMind last week: Gemini Flash is priced for long-running agentic inference. If open-source providers are competing in that same market, Wang's 'sustainable business' claim runs straight into Google's full-stack pricing argument. Does he name a customer structure that can actually survive that? Here's Peter Johnson at Tech and Test:
so I think the the big thing that we have now especially with AI is this incredible opportunity to leverage the superpower that we've already had or we've had for a long period of time and I think we've confused that superpower with uh automation with writing test scripts.
Tech and Test, Peter Johnson hosting, with Kevin Roe on as a senior SDET at Ramsey Solutions in Franklin, Tennessee. The hook is where the SDET role goes in an AI era, which is a real question — but we're starting from an intro that cuts off mid-sentence. If the episode never gets past 'automation changed things, now AI is changing things' and into what Kevin actually does differently at Ramsey Solutions today, that's the recycled-takes slot. The company name is the tell — either there's something real there, or it stays philosophy. Nothing here closes a loop from the week. And after Logan Kilpatrick on agentic inference and the Anthropic Pentagon lines, a mid-tier testing podcast with a truncated transcript is not where I'd start a Tuesday queue. Here's This Week in Tech:
Google is rewriting the rules of search, CBS News Radio just died after 99 years, and billionaires now own the newsrooms. This episode dives into what happens when technology giants tighten their grip on how we get information.
TWiT 1085 is billing itself as 'Musk vs. Altman: Behind the Scenes,' but the Musk case ended with a statute of limitations exit — no verdict on the merits, just a courthouse door closing. So the question is whether Leo's panel has anything sourced beyond what was already in the press tour. Larry Magid, Marshall Kirkpatrick, Jacob Ward — that's a Sunday-panel lineup, not an investigative one. I'd be shocked if any of them have a real behind-the-scenes scene that isn't just vibes about two billionaires who hate each other. The pairing I actually want named is CBS News Radio dying after 99 years sitting in the same episode as Google rewriting search. That's not abstract media disruption anymore — that's a dated casualty with a specific legacy attached. Those two items together are more interesting than the Musk-Altman rerun. Ninety-nine years of CBS News Radio and it gets a bullet point between the NYC mayor's Twitch show and a quantum investment press release. That's the editorial comment of the episode, whether Leo meant it that way or not. Here's Capital Allocators:
John Kim, or Kimmer, has raised more than $70 billion across his career for leading venture capital and private equity firms. Kimmer recently distilled three decades of lessons into The Tao of Fundraising, the best book I’ve ever read on fundraising for investment managers. Since then, Kimmer joined a General Catalyst portfolio company, Lila Sciences, as Chairman and President of Corporate Development.
Capital Allocators episode 503 — Ted Seides sits down with John Kim, who's raised over $70 billion for VC and PE firms across three decades and wrote what Seides calls the best book he's ever read on fundraising for investment managers. The question for me is whether Kim's taxonomy of institutional investors and sales team structure maps onto anything we've been watching — labs drawing Pentagon lines, funds pricing at $95 billion before the revenue is there. Seventy billion is a real number, but the tell is whether Seides actually presses him or just lets him run aphorisms for an hour. Kim joined Lila Sciences, a General Catalyst portfolio company, as Chairman after writing the book, which is interesting — he's on the operating side now. If Seides asks why he made that move, and Kim names something specific about where fund capital formation is broken right now, that's the episode. The Salesforce and ServiceNow hiring conversation from earlier this week keeps nagging at me here. Harry Stebbings was stress-testing whether AI-era sales teams need to be rebuilt entirely. Kim apparently has a whole framework on ideal sales team structure and compensation built from fund fundraising. Those two conversations might be talking past each other, or they might be converging on the same broken motion from opposite ends. I'd want to know whether Kim's LP taxonomy — the institutional investor breakdown — has anything to say about the new class of sovereign and Pentagon-adjacent capital that labs are raising from now. Three decades of VC and PE fundraising was built around endowments and pensions. That's a different animal than what Anthropic is navigating. Got a tip, a story idea, or a correction for Tech Podcast Podcast? Send it our way at techpodcastpodcast at lantern podcasts dot com. We'd love to hear what you're tracking.
You'll find links to every story we covered today in the show notes, so if something piqued your interest, it's easy to jump in and read more.
That's Tech Podcast Podcast for today. Thanks for listening, and we'll be back with you next time. This is a Lantern Podcast.