All-In

Debt Spiral or NEW Golden Age? Super Bowl Insider Trading, Booming Token Budgets, Ferrari's New EV

with David Sacks, David Friedberg, Chamath Palihapitiya, Jason Calacanis
14 Feb 2026 22 min read 1h 58m
TL;DR

AI tools are intensifying work rather than reducing it, spurring massive demand for AI-native employees who can structure work for agents—while prediction markets hit $2B in volume at the Super Bowl, exposing the inevitability of insider trading unless heavily regulated. The hosts argue on-premises infrastructure is staging a comeback as enterprises realize they can't afford to leak confidential data to public LLMs.

Key Moments
David Sacks
“I think we're kind of moving from what some people I think maybe Jensen has called um task-based jobs to purpose-based jobs. And I think a key skill of employees is going to be the ability to structure work for themselves and their AI agents.”
Sacks explains why the UC Berkeley study showing AI increases work intensity actually validates his prediction that AI will increase knowledge worker demand.
▶ 1:55
Chamath Palihapitiya
“once you use these tools, it is is difficult for a company to be able to control how their data is used subsequently thereafter. Meaning, if I give you Jason a PDF of some really important strategy document or a PowerPoint deck or a really critical model, and you're interrogating it with one of these models, if you're just using Chat GPT, the main line instance of it, you're leaking all of that prompt and response metadata back to Chat GPT.”
Chamath raises the critical security risk that drives enterprises back to on-premises infrastructure despite cloud's cost advantages.
▶ 7:38
Jason Calacanis
“we now have three or four of these. Uh we give them a notion account, a Slack account, and we give them a Google Docs account. They have their own email. And I think all of this technology was here all along. It was really or maybe for the last 6 months, let's say. Really good models out there, but no company would give the keys to the kingdom to allow these agents to actually act on your behalf.”
Calacanis describes his firm's deployment of autonomous AI agents ('replicants') running on Open Claw, which now handle 20% of employee work.
▶ 11:31
David Friedberg
“I think the question is is it really insider trading if you and I were making a side bet and I knew something about you and I had some edge or some advantage and I made a bet with you. Is that fair? Should the government have a role in regulating that?”
Friedberg questions whether prediction markets should be regulated like securities when they're fundamentally about information asymmetry.
▶ 22:16
Chamath Palihapitiya
“There are a certain percentage of these prediction markets that are about the well-functioning of society and the use of inside information gets to the truth faster. And I think that has value, especially if it uncovers corruption or misdeeds.”
Chamath argues prediction markets can serve as an incentive mechanism for whistleblowing and revealing truth faster, despite enabling insider trading.
▶ 27:38
About the show
›
The All-In podcast features four prominent investors and entrepreneurs discussing the week's biggest tech, business, and market developments. Known for contrarian takes and deep dives into emerging trends, the hosts analyze AI acceleration, market dynamics, and capital allocation opportunities shaping 2026.
Takeaways
1
AI agents demand token budgets as operational costs At current cloud API pricing, deploying AI agents costs $300/day per agent ($100K+ annually), forcing companies to implement 'token budgets' per employee. Unless inference costs drop 10x, token spending will exceed salaries for top developers, fundamentally changing how companies think about AI labor economics.
2
Data security reverting enterprises back to on-prem Using public LLM endpoints for proprietary work leaks all prompt/response metadata, forcing enterprises to choose between cloud convenience and data control. This is reversing the 15-year cloud migration trend—companies now need local infrastructure despite higher OpEx, similar to VAX terminal architectures.
3
Prediction markets expose insider trading as inevitable With $2B wagered at the Super Bowl, prediction markets are replicating pre-Reg FD stock markets where information asymmetry is the only edge. Unless heavily regulated, they'll inevitably become dominated by insiders (Israeli soldiers betting on strikes, halftime show leakers), benefiting 'sharps' over retail 'squares.'