↓
01
Agents
Chat Thread
What are neural networks, in simple terms?
Thinking
Think of them as pattern-matching machines. You feed them examples, and they learn to spot the underlying structure.
Can you give me a real example?
Thinking
Sure — image recognition. You show it thousands of cat photos, and it learns what "cat" looks like.
Thinking
Thinking ▸
The user wants a simple explanation...
I should avoid jargon like "backpropagation"
A metaphor would work — pattern matching is intuitive
Composer
Tell me more
⏎ Send⇧⏎ Newline/ Commands
Tool Use
⚡ search_web("agents")✓
Found 3 results from arxiv.org, wikipedia.org
📄 read_file("model.py")✓
Agents
researchrunning
code-reviewthinking
deployidle
monitoridle
Analyze
Tests
Refactor
Deploy
Verify
02
Writing
🌱
View all writing →How LLMs learn to play games
seed · March 2026
🌿
Structured outputs for game state
budding · in progress
🌳
Building AI agents that actually work
evergreen · coming soon
03
Lab
🎮
Game AI Arena
Watch language models compete head-to-head in classic games
Coming soon🤖
Agent Playground
Interactive demos of AI agents solving real tasks
Coming soon🧊
Voxel World
Real-time voxel rendering and physics in the browser
Coming soon04
Work
Ramp
Building AI agents for finance automation
CurrentSpeck
AI frontend engineer, enterprise agents
YC W24Sweep
Founding engineer, first hire
First HireFocus
AI agents, game development, interactive media, and creative engineering