Chamath Palihapitiya on Why AI's Social Contract Is Broken

Speaking from a leather armchair on a panel stage -- his name and title projected on a screen behind him -- Chamath Palihapitiya lays out a blunt comparison between the internet era and the AI era. His argument: the internet gave people a fair deal, and AI companies are blowing it.

The Internet's Original Bargain

Chamath starts with what he calls the "compact of the internet." Users contributed data, photos, and attention to platforms like Facebook and Instagram. In return, they got free tools they genuinely valued. The platforms used those contributions to build network effects and trillion-dollar businesses. Both sides got something. The deal felt proportional, even if the scale was lopsided.

AI Breaks the Deal

The AI model, as Chamath describes it, is extractive in a way the internet never was. AI companies learn from people's work, tokenize that knowledge, and sell subscriptions back to them. His summary of the pitch: "Thank you very much. See you later." Worse, the second version of this messaging -- that AI will learn your job and then replace you -- is what he calls "deeply and fundamentally irresponsible." Neither framing offers people a stake in the outcome.

What Positive-Sum AI Looks Like

Chamath gestures through his alternative with visible intensity, leaning forward in his chair and pointing at the audience. The positive-sum path means working alongside people, not extracting from them and moving on. He lists specifics: helping workers navigate regulators and government, showing that companies are hiring more people and paying them more, and taking the time to methodically work through complexity together. He frames this as a collaborative finish line, not a race to automate.

The Leadership Failure

The final thirty seconds of the clip hit hardest. Chamath argues that humans are resilient and fundamentally cooperative. Most people want to play an infinite game -- take care of their families, do meaningful work, and avoid getting exploited. The reason this is going wrong, in his view, comes down to one thing: "Terrible leadership. Terrible fucking leadership." The on-screen captions capture it in bold yellow text as he delivers the line with open hands, palms out, to the audience.

Key Takeaways

• The internet era operated on an implicit bargain: users contributed data and got free, valuable tools in return. AI companies are breaking that compact by extracting knowledge and selling it back as subscriptions.
• Framing AI as a replacement for workers is irresponsible. It turns a potentially positive-sum technology into a zero-sum threat.
• A constructive approach to AI means collaborating with workers, helping them navigate regulatory complexity, and demonstrating that the technology creates more and better-paying jobs.
• Chamath places the blame squarely on leadership. The technology is not the problem. The people directing it are.