QFM101: Machine Intelligence Reading List - February 2026

Source: Photo by Cash Macanaya on Unsplash

This month's Machine Intelligence Reading List covers solo AI-augmented startups, autonomous research, and the shifting nature of work. Linas Beliunas makes the case for The One-Person Unicorn, arguing that AI agents now handle enough operational work to make billion-dollar solo ventures realistic, while agent-slack provides the Slack automation plumbing to make that kind of agentic workflow concrete. Nick Bostrom's Optimal Timing for Superintelligence offers a formal framework for when to deploy superintelligent AI, and the Data Engineering for Large Models book argues that data quality sets the upper bound on model performance.

On the creative and research frontier, Google's Aletheia agent tackles autonomous mathematics research — solving previously open problems from the Erdos Conjectures database — while one developer gave Claude access to a pen plotter to create an iterative human-AI-machine art loop. The 5 Levels of AI Coding charts a maturity framework where the real bottleneck shifts from implementation speed to specification quality, a former CTO's test catches AI cheaters in interviews, and Andrew Yang reflects on The End of the Office as remote work, AI, and cultural shifts reshape professional life.

As always, the Quantum Fax Machine Propellor Hat Key will guide your browsing. Enjoy!