← All episodes

TBPN Gets Addicted to Social Media, Japan Twitter, Warren Buffett's Protégé, Deals Deals Deals

| 18 products mentioned
TBPN TBPN host
Tae Kim guest
Brett Adcock guest
Andre guest
Watch on YouTube social media litigation nvidia strategy ai inference demand semiconductor supply chain addictive design features geopolitical tech competition parental controls and regulation

TBPN covers landmark social media litigation, Nvidia's strategic pivot toward AI inference, and emerging tech opportunities across semiconductors, robotics, and geopolitical supply chains. Host discussion centers on attorney Mark Lanier's $6M verdict against Meta and YouTube for addictive platform design, contrasted against a nuanced argument that content—not features—drives engagement, plus analysis of why Nvidia remains positioned to dominate AI compute despite recent stock drawdowns and competitive threats.

Key takeaways
  • Mark Lanier won a landmark $6M verdict (compensatory + punitive damages) by using folksy courtroom tactics and props to convince jurors that Meta and YouTube's features—infinite scroll, algorithmic feeds, autoplay, notifications, like buttons—were deliberately engineered to exploit vulnerable teens; thousands of similar cases are pending, creating existential risk for social media business models.
  • The court's focus on addictive platform features over content mirrors cigarette litigation strategy, but the analogy breaks down because content quality and creator authenticity—not UI design—determines whether users return; Sora's failure despite identical addictive mechanics (infinite scroll, algorithmic feeds, likes) proves content is the real driver.
  • Inference demand from AI agents and coding assistants is exploding, not training demand; Nvidia is acquiring Grock and pairing it with Vera Rubin architecture to serve 75% traditional inference and 25% low-latency inference—a strategy that mirrors Jensen Huang's historic playbook of identifying macro shifts (like his 2019 Mellanox acquisition for large-scale GPU clustering).
  • CPU shortages are the next supply constraint even more acute than GPU shortages; AI agents require 4x more CPU cores than prior infrastructure models because orchestration, tool calls, database queries, and web searches run on CPUs not GPUs—a trend most investors haven't consensus-priced yet.
  • Taiwan fab capacity is the ultimate bottleneck; TSMC has preemptive power over all AI capex and can grant wafer allocations based on relationships; Intel and Samsung cannot quickly replicate TSMC's decades of trial-and-error expertise, making semiconductor geopolitics (including potential US-China negotiations) a leverage point for companies like Nvidia.
  • Parental controls, algorithmic feed transparency, and user opt-outs are more pragmatic regulatory solutions than banning addictive features outright; Instagram already lets creators hide like counts, and similar in-app toggles for infinite scroll and algorithmic recommendations could reduce liability risk while preserving platform functionality.

Recommendations (6)

Key Context recommends

"Go subscribe to Key Context on Substack. Follow Tae Kim on social media"

TBPN · ▶ 1:33:46

ChatGPT
ChatGPT uses

"I put in Gemini, I put in ChatGPT and just to make sure they're right and they're right"

Tae Kim · ▶ 1:31:02

Gemini
Gemini uses

"I put in Gemini, I put in ChatGPT and just to make sure they're right and they're right"

Tae Kim · ▶ 1:31:02

Slack
Slack uses

"We do that directly in, you know, Slack or Teams"

Andre · ▶ 2:50:16

Teams
Teams uses

"We do that directly in, you know, Slack or Teams"

Andre · ▶ 2:50:16

CrowdStrike

"Our head of security actually he rolled out CrowdStrike on our devices um and he did in like 40 minutes instead of what he said would have taken him like a day or two"

Andre · ▶ 2:58:18

Mentioned (12)

Substack
Substack "Go subscribe to Key Context on Substack" ▶ 1:33:46
Granola
Granola "We use granola internally and that should you should kind of just assume that's happening" ▶ 2:24:44
Nvidia H200
Nvidia H200 "We're going to see billions of dollars of H200 orders" ▶ 1:15:08
TSMC
TSMC "It feels like TSMC is not ramping capex nearly fast enough over the next few years" ▶ 1:14:26
Google TPU
Google TPU "Google is dying to get more TPU wafer capacity" ▶ 1:16:03
ARM CPU
ARM CPU "I wanted to know how you're processing the ARM CPU announcement" ▶ 1:18:04
xAI
xAI "Would you expect XAI/SpaceX at any point to get to basically just open up a shop as like a neocloud" ▶ 1:23:01
Grok
Grok "Is it going to come from training Tesla models, Optimus or Grok" ▶ 1:23:28
Colossus 2 "The biggest win has been Colossus 2, which was built very fast" ▶ 1:23:47
Starlink
Starlink "Think what Elon did with Starlink. I mean that's a telecom infrastructure play and this will be a..." ▶ 1:24:26
Sora
Sora "You saw what's happened to Sora right like you know everyone's all excited about Sora and that got" ▶ 1:32:34
AWS
AWS "I don't need to go set up an AWS instance" ▶ 2:30:47