li.
writing · notes from the work
blog
essays, notes, and thinking out loud
May 27, 2026 Project Glasswing: Claude Mythos, 10,000+ Zero-Days, and the New Cybersecurity Bottleneck.

Anthropic's Project Glasswing uses Claude Mythos Preview to find high-severity vulnerabilities across critical software. The result is a warning about AI security, zero-day discovery, and the human bottleneck in patching.

May 27, 2026 Foothills of the Singularity: What Demis Hassabis Said at Google I/O 2026.

Demis Hassabis ended Google I/O 2026 by saying we are at the foothills of the singularity. Here is what he meant, what Google announced, and why the AGI framing matters more than the product demos.

May 27, 2026 OpenAI's AI Math Breakthrough: An 80-Year-Old Erdős Problem Falls.

OpenAI says a general-purpose reasoning model disproved the planar unit distance conjecture, an 80-year-old Erdős problem in discrete geometry. Here is what happened, why the proof matters, and what it says about AI research.

May 22, 2026 The AI IPO Supercycle: OpenAI, SpaceX/xAI, and Cerebras Are Racing to Wall Street — What It Means for the Industry.

OpenAI is filing confidentially. Cerebras already crossed the line. Anthropic is closing in on a $900B valuation. The AI IPO supercycle is here — and it's about to test whether safety-first values survive contact with public markets.

May 22, 2026 Anthropic Is Now the Most Disruptive Company in Tech — And It's Just Getting Started.

Anthropic just hit #1 on CNBC's Disruptor 50, overtook OpenAI on annualized revenue, and signed chip deals across every major hyperscaler in the same week. How the safety-first lab became the most dominant force in enterprise AI.