Practical AI, built in public
Greg Does AI
Learning the machinery while everyone else is still reading the memo.
Practical AI experiments for sensing signals, testing simulations, and making ambiguity easier to use. The agents in the background are playing right now.
- 4
- Projects
- 2
- Live builds
- 0
- Agent decisions
01 · Featured · Updated build
Tanks AiLOT
A browser simulation for watching AI agents make tactical decisions in a constrained world. The useful part is not the tanks — it is seeing reasoning turn into movement, outcomes, and repeatable experiments. Watch the trailer, or play the latest build right here.
Deploying tanks…
Taking too long? Open Tanks AiLOT on its own page
02 · Selected work
Selected AI projects
High-level notes on active work, kept intentionally concise until each project is ready for a deeper public write-up.
03 · Receipts
Build log
Figuring it out in public means showing the work, not just the launches. Dated entries: what was tried, what broke, what stuck.
04 · About
About Greg
Greg Kitchen builds practical AI systems that make complex signals easier to understand and act on. His work sits at the intersection of strategy, market sensing, product thinking, and rapid prototyping.
The through line is translation: turning ambiguity into structure, structure into tools, and tools into clearer decisions for people doing consequential work.
- Clear signal
- Find what matters without adding noise.
- Human judgment
- Use AI to sharpen decisions, not replace context.
- Useful loops
- Favor systems that can be tested, improved, and reused.