Notes

AI is fundamentally reshaping the economic premises of software security: as vulnerability discovery shifts from scarcity to abundance and the cost curves of both offense and defense collapse, traditional proxies like "finding more bugs" no longer equate to real security progress. Technical tools can establish factual truths about vulnerabilities (what is), but they cannot resolve the value judgments required to weigh real-world risks against engineering trade-offs (what ought to be). Security research should study security itself: as AI commoditizes technical execution, our core responsibility shifts from demonstrating what is achievable to justifying what is worth solving.

Starting from the meritocratic ideals shared by China and the United States, this essay examines how success, intelligence, and meaning became intertwined—and how stalled mobility and AI are pulling them apart.

A look back at four principles I followed when I began working with LLMs in 2023—no fine-tuning, no RAG, no constrained decoding, and no temperature 0—and how they grew into my approach to agents.

(Parody work)

A paper-style narrative around P versus NP that proposes an empirical proof via benchmark saturation, using a large 3-SAT benchmark, model accuracy, and scaling-law extrapolation to announce a mathematical conclusion.

A reflection on subjectivity in the age of AI agents: humans differ from agents not only by capability, but by belief, history, responsibility, and the need to stand behind their own judgments.