TBPN Diet: The AI Lab Market Map, Robinhood Brings Startups to Retail, GLPs & Hedge Funds

Hosts: John Coogan & Jordi Hays
Guests & Contributors: Tyler Cosgrove, Quin, Gwen
Date: February 19, 2026

Episode Overview

This episode of Diet TBPN dives deep into the ever-evolving world of AI labs, thanks to Tyler Cosgrove’s comprehensive "market map" visualizing the entire AI ecosystem. The hosts, joined by contributors, break down the nuances distinguishing “trad labs” and “neolabs,” discuss the democratization of startup investing through Robinhood, muse on competition in the private space race, explore the impact of weight loss drugs on high-pressure finance jobs, and more. The fast-paced, insight-rich conversation is peppered with humor, techno-insider observations, and actionable takeaways for anyone tracking AI, startups, or the future of work.

1. Mapping the AI Lab Ecosystem

[00:02 – 15:00]

Tyler Cosgrove’s Market Map:

• Every company visualized: Tyler’s latest map plots every known company using embeddings of all 7.5M English Wikipedia articles and reduces them to two dimensions for interactive browsing.
• Methodology: Quin describes using custom embeddings to identify “company-ness” from Wikipedia ([00:33–01:44]).
  • “I ran them through an embedding model...every article has a vector. It’s like 2500 [dimensions].” – Quin ([00:44])
  • “You can basically find things that are companies and then you have an embedding for every single one… If you map it down to 2D, you can have this cool 2D map.” – Quin ([01:38])

Categories in the AI Lab Market Map:

• Trad Labs: OpenAI, DeepMind, Anthropic, Xai, Mistral ([04:10–04:57])
  • “You’re in the trad lab world when you’re thinking about a big pre-train run.” – Jordy ([04:57])
• Sovereign Labs: Non-US labs like Mistral (EU), Cohere (Canada), and a range of Chinese open-source labs ([05:16–05:54])
• Legacy Labs: Entrenched labs at major enterprises (Microsoft Research, AT&T Bell Labs, FAIR at Meta) ([06:11–06:54])
• Neolabs: The newer entrants, often spun out of existing labs, focused on novel research (e.g., Prime Intellect, Thinking Machines, SSI) ([07:16–08:39])
  • “Prime Intellect is kind of the prototypical, like, quintessential neolab.” – Quin ([07:16])
• SaaS Labs (Trad & Neo):
  • Trad SaaS Labs: AI for enterprise data (e.g., Thing Machines)
  • Neo SaaS Labs: Startup-focused SaaS AI products (e.g., Ramp Labs, Cursor) ([08:39–10:01])
• Post Labs: Labs working on evaluation (“evals”) or building on top of other trained models (e.g., Meter, Epoch, Pangram) ([10:01–10:40])
• Safety Labs: Focused on AI safety, mechanistic interpretability (Anthropic, Goodfire, Eleuther AI) ([10:43–11:00])
• Consumer Labs: AI for direct consumer products (Eureka Labs, Humans) ([11:05–11:34])
• Other Categories:
  • Visual Labs, Neo-auditory Labs (e.g., Eleven Labs, Suno, Gemini for music/voice)
  • Wet Labs: AI for biological sciences
  • Kinetic Labs: Robotics/neural robots (old and new generations)
  • Dark Labs: Government or defense-related (SHIELD AI, DARPA) ([11:46–14:26])

Notable Quotes:

• “Thinking Machines and SSI are two of like the first case studies that set the tempo for…it’s possible to do some research outside the big trad labs.” – Jordy ([07:59])
• On the ecosystem’s blurry boundaries:
  • "These lines are pretty blurry." – Quin ([10:01])
• On the need for competition:
  • "[The] US needs two SpaceX's, it needs two launch companies competing vigorously...because our adversaries aren't standing still." – Gwen ([20:40])

2. Robinhood Brings Private Startups to Retail Investors

[15:00 – 16:50]

• Robinhood Ventures now lets retail investors access a basket of top private startups (Databricks, Ramp, Stripe, etc.).
  • “Historically investing in private markets was limited to institutions and the elite, but not anymore.” – Tyler ([15:05])
• Pricing Moves: Early buys of Databricks and Ramp already up; Air Wallix and Mercore prices fluctuate ([15:32])
• Risks: These closed-end funds can diverge from actual asset values, leading to speculative bubbles and retail investors getting “their face ripped off.” ([16:12])

Notable Quotes:

• “With FOMO from Access, this could easily trade at a very high multiple to nav, leading to a lot of retail investors getting their face ripped off.” – Tyler ([16:12])

3. XAI, Benchmarks, and the Future of LLMs

[16:50 – 19:22]

• XAI focus shift: Elon Musk announces XAI will focus on “real-world utility” over academic benchmarks (like Humanities Last Exam).
  • “Actually, I don’t think HLE is a great measure of usefulness.” – Musk, paraphrased by Jordy ([16:50])
• Grok 4’s minor revision: Four-agent system with distinct system prompts, aiming for qualitative reasoning ([17:25–18:07])
• Tech Speculation: Can Tesla’s existing expertise in silicon and inference become an XAI advantage? ([18:07–18:54])
• Big moves: News of massive investments and partnerships involving XAI, SpaceX, and Saudi infrastructure ([18:54–19:38])

4. The Private Space Race: Blue Origin vs. SpaceX

[19:42 – 22:32]

• Jeff Bezos’ “Black Tortoise” Tweet: Signaling Blue Origin’s intention to compete fiercely for the Moon ([19:42])
• Insider Banter:
  • “What the US needs is...two SpaceX’s...two launch companies competing vigorously…” – Gwen ([20:40])
  • “I like our architecture, I like our odds...I don’t have a crystal ball into what SpaceX is doing...but I love the fact that the US would compete us against each other.” – Gwen ([21:25])
• Hosts riffing on branding:
  • “Elon needs to wear tortoiseshell glasses...be like, I turned your tortoise into my glasses and Bezos needs to start carrying a rabbit’s foot for good luck.” – Jordy ([22:32])

5. Claude & OpenAI: Platform Access Tensions

[22:46 – 23:23]

• Anthropic restricts 3rd party use: Claude’s OAuth tokens no longer permitted for external apps, signaling stricter platform lockdowns ([22:47])

6. Energy Access & Rockefeller Philanthropy

[23:23 – 25:20]

• EQT and Rockefeller foundations: New non-profit “Energy Corps” to build infrastructure in developing nations, not limited to renewables ([23:23])
  • “Energy Corp. sees a role for a broader spectrum...from fossil fuels to solar panels and nuclear plants.” – Tyler ([24:25])
• Macron cited for leadership on French energy policy ([24:48])

7. The Exponential Age: Robotic Labor & Abundance

[25:20 – 26:54]

• David Holz’s vision: 5 million humanoid robots could build Manhattan in 6 months; 10B robots by 2045.
  • “Now just imagine what the world looks like when we have 10 billion of them by 2045. Now imagine the year 2100.” – Tyler ([25:20])
  • “Dyson Sphere, Dyson Sphere by 2100 is the correct debate.” – Jordy ([25:36])
• Land’s scarcity post-robots questioned:
  • “With 10 billion of them, I don’t even think land will be scarce anymore. It’s like, hey, we’re gonna build an island.” – Tyler ([25:42])

8. Ozempic, GLP-1s, and the 'Hedge Fund Grindset'

[27:16 – 28:28]

• Rumor: Some hedge funds ban traders from taking Ozempic and other GLP-1 weight loss drugs, fearing they kill “gut instinct.” ([27:26])
  • “Traders need to make quick decisions based on gut instinct and GLP1s mess with your gut instincts, you’re not hungry for snacks, you’re not hungry for profits, you lose your edge.” – Tyler ([27:26])
• "Stacking" for success:
  • “Solution: ban GLP1s only if you’re taking them solo...microdose of tirzepatide to cut down on physical appetite, macrodose of testosterone to amplify psychological appetite.” – Dr. Cameron, relayed by Tyler ([28:03])

9. Lighter Moments: Hot Dog Market Maps, Haircuts & Overheard Advice

[26:54 – 29:41]

• SF memes: Blueberry vs. hot dog discourse ([26:54])
• "Unclankable" businesses: Robots, hot dogs, and future-proofing against tech disruption ([26:07, 26:54])
• Overheard VC advice:
  • “OpenAI and Anthropic are like Godzilla. You need to find an alleyway to hide in.” – Anonymous VC, relayed by Jordy ([29:05])
• Miscellaneous banter: Haircut alerts, jokes about product markets, and rapid-fire tech predictions.

Notable Quotes

• “Market maps are basically done. But a lot of the neolabs are not on this market map.” – Jordy ([02:49])
• “There was this question about should Brazil be allowed to clear cut the Amazon rainforest to pull forward industrial civilization...they would certainly benefit in the short term. So there’s a hot debate here and he is engaging in it.” – Jordy ([24:54])
• “You see, I’m the CEO of a hot dog company…your life is about to change. So what can you do? Buy as many hot dogs as you can. Buy stock in hot dog companies.” – Tyler ([26:54])

Important Timestamps

• 00:02 – Show intro and the start of the AI market map discussion
• 01:44 – Embeddings & methodology for mapping companies
• 03:44 – What is a Neolab? Differentiating lab strata
• 10:43 – Safety Labs, Consumer Labs, and beyond
• 15:05 – Robinhood brings private startups to retail
• 16:50 – Elon Musk on XAI & benchmarks
• 18:54 – XAI/SpaceX/AI infrastructure news
• 19:52 – Blue Origin vs. SpaceX moon race and Bezos’ “Black Tortoise” tweet
• 22:47 – Anthropic restricts Claude OAuth API use
• 23:23 – Rockefeller/EQT energy access initiative
• 25:20 – David Holz’s vision for humanoid robots
• 27:26 – Ozempic/GLP-1s and Wall Street hedge fund culture
• 29:05 – VC advice: Avoiding the Godzillas of AI

Tone & Style

The episode balances deep tech analysis with Silicon Valley insider humor, rapidly shifting from technical break-downs to memes, jokes, and quick takes on cultural and business trends. The banter is insightful but never too serious.

Summary Takeaway

This whirlwind TBPN Diet episode explores the shifting landscape of AI labs, the boundary-blurring rush to map and categorize the booming ecosystem, democratizes access to high-growth private companies, and reflects on the physical and cultural impact of technology — from space races to diet drugs and robot-driven futures. The show’s playful tone and rapid-fire format make complex subjects approachable and the latest startup gossip irresistible, even for listeners outside the Silicon Valley bubble.