TBPN
TBPN

AI Viruses, OpenAI's First Device, WSJ Mansion Section | Samir Kaul, Patrick Wendell, Grant LaFontaine

August 7, 2026

AI Summary

5 min read

AI Viruses, OpenAI's First Device, and the DeepMind Exodus

The episode opens with a striking scientific milestone: researchers at Stanford and the Arc Institute trained an AI model to recognize patterns in naturally occurring viral DNA, then used it to generate genetic sequences for entirely new viruses. After synthesizing those sequences and inserting them into bacteria, the team found the bacteria produced viable viruses capable of infecting other bacteria. The study, published Thursday in Science, represents the first time AI has been used to design entirely new viruses that function in the real world.

The hosts immediately flag the obvious concern, but also the critical caveat: the AI was trained only on bacteriophages—viruses that infect bacteria—and specifically excluded viruses that infect humans, plants, animals, and fungi. As a result, the model cannot generate viruses capable of infecting people. The debate among the hosts centers on whether this is a material breakthrough or just a new tool for something scientists have been doing for years. The answer comes down to acceleration and cost: if generating viruses becomes a thousand times cheaper, that could reshape biotech in positive ways but also carries biosecurity implications. The research fellow quoted in the piece noted that governments and scientific organizations have been slow to develop guardrails that

Continue reading the full summary in the app — free to try.

Read Full Summary →

Free • No credit card required

Listen to Audio Summary Open in App

What you'll learn

  • 1 (03:06) **AI-Designed Viruses** - Researchers at Stanford and the Arc Institute used AI to generate entirely new viruses that can infect bacteria, marking the first time AI has designed functional real-world viruses.
  • 2 (08:20) **OpenAI's First Consumer Device** - Bloomberg reports OpenAI is developing a hockey-puck-sized, battery-powered smart speaker without a screen, expected to cost $300-$400 and launch in 2027.
  • 3 (11:12) **NotebookLM and the Future of Voice** - The hosts discuss the evolution of NotebookLM and how live voice mode might replace pre-generated podcasts for learning.
  • 4 (16:42) **New Mexico Judge Orders Meta to Pay $900 Million** - A judge has ordered Meta to pay over $900 million in penalties and establish a fund for harms linked to its platforms, continuing a wave of social media addiction lawsuits.
  • 5 (19:36) **The Tobacco Master Settlement as a Template** - The hosts draw parallels between the current social media lawsuits and the 1998 Tobacco Master Settlement Agreement.
  • 6 (25:29) **SemiAnalysis: DeepMind is No Longer a Frontier Lab** - A new report claims DeepMind is no longer a frontier AI lab due to mass departures from its RL teams and poor compute allocation, with Google choosing to sell compute to competitors like Anthropic.
  • 7 (33:42) **Demis Hassabis's All-In Bet on AI** - The hosts revisit a story about Demis Hassabis testing Mark Zuckerberg's commitment to AI before choosing to sell DeepMind to Google.

+ Full timestamped outline available in the app

Guests on this episode

Grant LaFontaine

Show Notes

  • (01:31) - AI Viruses
  • (08:18) - OpenAI's Device Takes Shape
  • (16:39) - Meta Faces Major Child Safety Case
  • (25:25) - 𝕏 Timeline Reactions
  • (36:10) - WSJ Mansion Section
  • (43:41) - Samir Kaul, a general partner at Khosla Ventures, discusses the firm’s investment in Jeff Dean’s Discovery Loop and its potential to apply AI to scientific research with tangible outcomes. He also shares his views on AI regulation, inflated private-market valuations, venture capital strategy, founder support, and the importance of taking bold risks to generate exceptional returns.
  • (01:09:45) - Patrick Wendell discusses his role as Databricks co-founder and VP of Engineering, where he leads AI products and internal AI adoption. He explains how AI coding tools can nearly double engineering capacity while creating rapidly escalating consumption costs, and highlights model switching, intelligent routing, and optimization as key ways to control spending without sacrificing productivity.
  • (01:27:30) - Grant Lafontaine discusses Whatnot’s $500 million Series E funding round at a $20 billion valuation and the live-shopping platform’s rapid growth. He explains how authentic, knowledgeable sellers can build substantial businesses with small audiences, while outlining Whatnot’s focus on customer experience, new markets, product categories, and future streaming formats.
  • (01:44:17) - 𝕏 Timeline Reactions


TBPN is made possible by:

Ramp - https://ramp.com

Public - https://public.com

Cisco - https://www.cisco.com

Console - https://www.console.com

CrowdStrike - https://www.crowdstrike.com

Figma - https://www.figma.com

MongoDB - https://www.mongodb.com

NYSE - https://www.nyse.com

Railway - https://railway.com

Shopify - https://www.shopify.com

Codex - http://openAI.com/codex


Follow TBPN: 

https://TBPN.com

https://x.com/tbpn

https://open.spotify.com/show/2L6WMqY3GUPCGBD0dX6p00?si=674252d53acf4231

https://podcasts.apple.com/us/podcast/tbpn/id1772360235

https://www.youtube.com/@TBPNLive

TBPN

More from this podcast

TBPN →