It feels odd to write a blog post about vibe coding in late 2026. All those posts were written back in early 2025. But here we are.
I’ve spent the last few years working for a defense contractor, largely on projects involving LLMs, agents, and neurosymbolic systems. Perhaps ironically, for government and legal reasons, I’ve been pretty limited in how I could use those same AI tools during development.
A few years ago, CJ McFate and I wrote a paper1 about System 1 vs System 2 in theory of mind reasoning. For the unfamiliar, System 1 and System 2 come from a long tradition of dual-process theories of cognition, but the terminology was popularized by Daniel Kahneman in his 2011 book Thinking, Fast and Slow. As the title implies, the basic idea is that we have two modes of thinking: System 1 is fast, automatic, and efficient, while System 2 is slower and more deliberate. If this all sounds familiar due to the recent release of System 1 models in AI, you’re onto something. We’ll come back to those.
I’ve been meaning to start writing more about AI outside of papers and work projects, so… here goes.
This blog is a place for things I’m thinking about, working on, or find interesting—AI research, things I learn while building with AI, questions about how we evaluate these systems, and probably the occasional opinion about where the field is going.
Some posts will be technical. Some won’t be. Some may grow out of projects I’m working on; others may just be things I found interesting enough to write down.