Jun 14, 2025 · 2h 56m · tbpn
Weekly Recap: Apple WWDC, Andrew Huberman, YC Demo Day, Intern Phone Assembly, $14B Meta x Scale
John Coogan · 1h 2m spoken Andrew Huberman · 36m spoken Garry Tan · 9m spoken Tyler Cosgrove · 1m spoken Tech Investor / Commentator · 39s spoken
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this expansive weekly recap, co-hosts John Coogan and Jordy analyze Apple's WWDC announcements and AI strategy, interview Dr. Andrew Huberman on NIH research funding reform, report live from Y Combinator's Demo Day, and break down Meta's landmark $15 billion investment in Scale AI alongside live in-studio intern stunts.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. The hosts hold 39.2% of the talking time here. How this is scored →
The hosts as informed peer 4.3 Guest teaching 3.4 Guest disagreement 0.5 The hosts pushing back 0.2
speaking balance: gold is the hosts, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 1:21:10 Huberman dismisses gray market peptide experimentation
Huberman forcefully dismisses fitness biohackers and gray market peptide clinics, insisting that without controlled clinical human data, self-experimentation lacks scientific rigor.
Hardest push from the hosts ▶ 2:51:40 Host commentary on FTC regulatory evasionThe hosts and commentator challenge the pretense of recent hyperscaler deals, pointing out that paying $15 billion while absorbing the founder is a blatant backdoor acquisition to bypass FTC scrutiny.
Biggest teaching moment ▶ 1:04:00 Huberman details the NIH funding and indirect cost apparatusHuberman educates the hosts on how university indirect cost rates (IDC) work, why 40% cuts endanger non-endowed public institutions, and how basic science yields downstream therapies.
The host holds their own ▶ 15:35 Coogan dissects Siri's architectural failureCoogan demonstrates deep technical knowledge by explaining how Siri's legacy keyword router architecture prevents it from operating as a fluid, generalized generative agent without a complete architectural overhaul.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | The hosts as informed peer | Guest teaching | Guest disagreement | The hosts pushing back | Why |
|---|---|---|---|---|---|---|
| Weekly Recap Overview and Episode Topics | 4 | 0 | 1 | 1 | Coogan and Jordy open the recap by breaking down public reaction to Apple's WWDC Liquid Glass announcement, framing it around Ben Thompson's analysis and staging an intern bit. | |
| Breaking News: Snap Announces Next-Gen AR Specs | 4 | 0 | 0 | 0 | Coogan and Jordy review breaking news about Snap's sixth-generation Specs and contextualize hardware constraints against Meta's Reality Labs budget. | |
| Handcrafted Finality vs Generative AI and the Siri Architecture | 6 | 0 | 0 | 0 | Coogan outlines the architectural difference between Siri's deterministic API keyword routing and modern transformer-based response generation, explaining why Apple struggles with cloud AI. | |
| Gruber's Liquid Glass Review and Tyler's Storage Constraints | 4 | 0 | 0 | 0 | The hosts read John Gruber's Daring Fireball review and a critical commentary from Signal characterizing Apple's intelligence efforts as spiritually lost. | |
| WWDC AI Disappointments and Future Search-First Interfaces | 5 | 0 | 0 | 0 | Coogan discusses the paradigm shift from scrolling to search-first interfaces in photo indexing and reviews historical design reactions to major iOS overhauls. | |
| The Skeuomorphic Era: Tactile Metaphors in Early iPhone OS | 5 | 0 | 0 | 0 | Coogan cites Sebastian De With's essay on skeuomorphic design in early iPhone OS, noting how tactile visual metaphors eased consumers into touchscreen interactions. | |
| The Flat Design Evolution to Modern Physicality | 5 | 0 | 0 | 0 | The hosts trace the evolution from iOS 7 flat design and dynamic parallax layers to contemporary computational UI rendering in VisionOS. | |
| Sponsor Message: Vanta | 5 | 0 | 0 | 0 | Coogan analyzes Apple's Foundation Models framework, emphasizing the advantage of zero-marginal-cost on-device inference for small developers over cloud APIs. | |
| Sponsor Message: Numeral | 5 | 0 | 0 | 0 | Coogan explores why Apple's native voice dictation lags behind OpenAI Whisper, citing corporate copyright liabilities and PR constraints with training data. | |
| DHH Interview on Apple's Organizational Stagnation | 3 | 0 | 0 | 0 | The hosts reference DHH's critique of Apple organizational paralysis before cutting to their intern unboxing components for a manual iPhone assembly bit. | |
| Checking In: Home Buttons, Motherboards, and Reshoring | 3 | 0 | 0 | 0 | Coogan and Jordy check in on the intern's iPhone 4S assembly, using the stunt to satirize the cultural and economic realities of manufacturing reshoring. | |
| Andrew Huberman on NIH Budget Cuts and Public Trust | 3 | 8 | 2 | 1 | Andrew Huberman explains the proposed 40% NIH budget reduction, detailing the breakdown between basic science and clinical trials, the indirect costs dispute, and public trust erosion post-COVID. | |
| Medical Research Priorities: GLP-1s, Peptides, and MAHA | 4 | 8 | 2 | 1 | Huberman details how Gila monster basic research produced GLP-1s, criticizes the gray-market testing of unpatented peptides like BPC-157 and Pinealine, and warns about unregulated stem cell treatments. | |
| Media Responsibility, Psychedelics, and the Grant Deletion Test | 3 | 8 | 2 | 0 | Huberman contrasts high-dose clinical trials for psilocybin and MDMA against ineffective microdosing trends and outlines his proposed deletion test for evaluating academic research grants. | |
| Reforming Academic Science and Huberman Lab Milestone | 4 | 7 | 1 | 0 | Huberman advocates replacing individual PI lab branding with mission-based labs to empower early-career researchers before sharing that Huberman Lab reaches 20 million people. | |
| Live from YC Demo Day: Garry Tan on AI and Founder Volition | 4 | 6 | 0 | 0 | Garry Tan outlines Spring batch trends at YC Demo Day, noting that 90% of companies are AI-focused, founder applications among 18-22 year olds jumped 110%, and emphasizes founder agency. | |
| YC Founder Spotlight: Agentic People Search with Tom and Eric | 4 | 6 | 1 | 0 | Eighteen-year-old founders Tom and Eric explain their agentic people search platform, demonstrating parallel LLM processing across hundreds of millions of public profiles. | |
| Outset Founder Discusses AI-Moderated Market Research and Series A | 4 | 6 | 0 | 0 | Outset founder Aaron discusses raising a $17M Series A for AI-moderated user research, explaining why voice and video without visual avatars generate superior qualitative data. | |
| Godella Unveils Fast Physics Simulation Models for Engineering | 4 | 7 | 0 | 0 | Godella founders explain their physics-informed machine learning model, which achieves a 4,500x speedup over traditional CFD and mechanical simulation software. | |
| Morpho AI on Automating Industrial Robot Hardware Design | 4 | 6 | 1 | 0 | Morpho AI founders discuss automating custom industrial robotics hardware design in days rather than months and express near-term skepticism regarding humanoid robots. | |
| Vibrate Founders on AI-Assisted Grading and Classroom Workflow | 4 | 5 | 0 | 0 | Vibrate founders describe their Chrome extension embedded directly inside Google Classroom and Canvas to give teachers automated essay grading and formative feedback. | |
| Bloom Demonstrates Instant Native App Generation via App Clips | 4 | 6 | 0 | 0 | Bloom founder demonstrates instant mobile app creation and hot-reloading through Apple App Clips, bypassing App Store review cycles for ephemeral apps. | |
| Analysis of Meta's $15 Billion Investment in Scale AI | 6 | 3 | 1 | 1 | Coogan and Jordy analyze Meta's $15B 49% stake in Scale AI, highlighting how hyperscalers use quasi-acquisition deal structures to circumvent FTC review and acquire technical leadership. | |
| Insider Reporting on Alexandr Wang, Jason Droege, and Meta AI | 5 | 6 | 0 | 0 | A tech reporter breaks insider details regarding Jason Droege taking over Scale AI as CEO while Alexandr Wang transitions into leading Meta's Superintelligence Lab. |
Statements from this episode (31)
Snap's 2024 Spectacles required a $99 monthly annual developer lease
“The most recent edition of spectacles were released in September, 20, 24 to developers only that edition of the glasses was available under a leasing model that record that required users to commit to paying 99 a month for a full year if you wanted to develop …”
Snap glasses developers can integrate Google's Gemini AI models
“Snap said that developers will be able to incorporate Google's Gemini AI models into programs that they developed for smart glasses.”
Snap's market cap equals Meta's $14B Scale AI deal
“Snaps market cap, exactly what meta just paid for a 50% stake in scale AI, fourteen billion.”
Apple will likely defer to other companies for generative AI
“They're not getting over their skis and they will eventually, I mean, if they're to win, probably defer to the other companies and play to their strengths and let, let the other companies win in the places where they are.”
Every major hyperscaler except Apple has acqui-hired an AI lab
“Every major hyperscaler except for apple now,
Has basically done one of these like interesting aqua hires of a foundation model lab to kind of juice up their AI efforts.”
Apple holds a patent for a transparent glass iPhone
“They have a patent for a glass iPhone that, that you can see through or something like that, or something that actually would appear more like glass or appear transparent.”
Future photo apps will replace scrolling entirely with AI search
“In the photos app, they are understanding that in the future, you won't want to scroll at all. You will just search and search will be so powerful because every photo and video will be tagged so thoroughly from AI and it will understand the context so much tha…”
Apple's on-device AI enables apps with zero ongoing cloud costs
“It is very cool in terms of just, you can now vibe code an iOS app that has some sort of AI functionality. It's not gonna be incredible. It's not gonna be frontier. But you can get something that's served into the store that only uses Apple's APIs and creates …”
Apple voice transcription lags far behind market competitors
“So Sarah Guo at conviction, former former guest on the show, friend of the show said that apple transcription is still so bad. So, so bad compared to what's available on the market. They should be embarrassed. Okay. So I agree.”
Apple transcription does not use modern transformer architecture
“And it's clearly not the modern transformer based architecture.”
Much AI reinforcement learning research focuses on difficult math problems
“A lot of the reinforcement learning is now done on like extremely hard math problems. And so you might need to go to PhD programs to recruit.”
Congress is considering a 40 percent budget cut to the NIH
“There is an upcoming vote in Congress in September, I believe, and it's on the table to cut the overall budget for research for NIH by 40%, four zero.”
Most Nobel Prizes in medicine stem from NIH-funded work
“The majority of Nobel prizes has in physiology and medicine and chemistry and related fields have come from work that was either initially seeded by, or certainly supported by the NIH.”
Indirect research cost cuts will disproportionately harm public universities
“A severe cut to the indirects, while on the face of it might sound good, that's going to disproportionately hurt the public schools without large endowments, ok, because they don't have money to dip into.”
Huberman: GLP-1 drugs ended the debate on the cause of obesity
“This peptide acts at the level of the brain and the gut to make people feel more full. It also, by the way, shut down the debate as to why people are fat. It's because they eat too much. Right? They eat more than they burn. Remember people used to argue about …”
Huberman: Most NIH-funded research is completed before grant approval
“Much of the work that's funded by NIH. Here's the dirty secret is already completed because they tend to fund things that are very certain to be completed. And so people put forward grants based on work and preliminary data showing I can do this. And it's alre…”
Stem cell treatments are not yet ready for clinical use
“Is this to say that stem cells are bad? No, they're not ready for clinical use yet. And the United States has been very careful about things like this”
Huberman: Microdosing psilocybin has no major therapeutic effect on depression
“The data on micro dosing psilocybin are basically clear that it's not known to have any major effect on major depression. Or some of the other things that psilocybin is being tested for clinically including major depression, PTSD, et cetera.”
MDMA achieves 60% to 70% remission rates for PTSD
“The remission rates on PTSD, by the way, from properly dosed and spaced MDMA is remarkable up to somewhere between 60 and 70% remission rates on PTSD.”
Huberman: Neuralink will solve paralysis
“I fully believe that Neuralink is going to solve paralysis. I really do. I'm not saying that cause we're on exits. I know the head neurosurgeon there, they're well on their way.”
Academic science should replace PI labs with mission-focused labs
“I think that the independent investigator model of science in this country, where it's the Huberman lab, or it's the whatever lab named after the PI, I think that's a flawed model. This is gonna really upset some people. There should be labs named after a part…”
Huberman: Joe Rogan's reach is 10x all major news hubs combined
“Rogan's reach is, like, beyond. I know the numbers. I, I'm not gonna talk about his numbers, but it's like, it's an order of magnitude greater than all the other major news hubs combined.”
Andrew Huberman says his podcast reaches 20 million people
“But, no, we are very blessed to reach you know, twenty million people.”
Garry Tan: YC Spring 2025 batch is 90% AI, 11% hard tech
“Dude, it's like 90% AI. It's about 10% 11% hard tech, which is awesome.”
Garry Tan: Accepted 18-to-22-year-old YC applicants are up 110% year-over-year
“The wild stat that we're seeing is actually the number of 18 to 20 two-year-olds applying to YC and getting it is up a 110%. Year on year, right?”
Platform neutrality could enable hundreds of thousands of 10-person $1B companies
“If we do that, we can actually have tens of thousands or hundreds of thousands of companies, each of which can get to a billion dollars net revenue, you know and like you could do it probably with 10 people, right?”
Roughly 30 percent of YC startups pivot during the batch
“Something like 30% of the companies change their idea during the batch, and that's actually great.”
Garry Tan: Y Combinator acceptance rate is down to 0.8 percent
“We're accepting companies at a 0.8% rate right now.”