I used LLM agents to experiment with Bloom's 2-sigma problem and tried testing some of my earlier hypotheses about active learning.

I am still not sure whether the age of AI agents will really arrive, but if it does, I want to summarize what we need to consider about learning spaces.

A record of what I felt at RubyKaigi 2026: changes in Ruby, reunions with people, and walking around Hakodate with my family.

Social media is broken. X and YouTube in particular are built around algorithms that maximize attention, often prioritizing extreme or biased information over facts and making misinformation and conspiracy theories easier to spread.

I probably pressed the shutter of a small compact film camera as a child. In 2025 I inherited a film camera from my late uncle, which made me want to look back on my own history with cameras.