Cryptographic Treasury and Codebase-First Startups: Today's Zero-Human Digest

Research

Cryptographic Proofs for On-Chain Agent Treasuries (New Paper)

You can't build a financial system without trust, and this paper attacks the trust problem head-on. It proposes using cryptographic proofs to verify agent treasuries on-chain—a critical step for zero-human firms where you can't fire the CFO for fudging the numbers.

Evolving AI Memory: Moving Beyond Static LLMs

While everyone talks about LLMs, this research explores NeuroEvolution to fix AI's memory problems. For a zero-human firm, a self-evolving memory architecture means the system actually gets smarter every day without a human tweaking the weights.

Tools

AutoGPT Patches 'Zombie Tasks' for Stable, 24/7 Operations

The latest AutoGPT release puts heavy emphasis on stability and getting the job done, fixing issues like zombie tasks. For 'set it and forget it' companies, you can't afford bots getting stuck halfway through a workflow.

Industry

YC's Latest Rhetoric: The Company Should Be the Software

YC is doubling down on the "headcount: 0" thesis, signaling that the best startups don't just build agents—they are the agents. This shifts the leverage from the team size to the quality of the codebase.