Remember training your webcam AI in Module 1? Now do it with ears! Teach the AI 3 different sounds — a clap, a whistle, a "shh" — from your own examples, and watch it learn to tell them apart. Same idea, new sense.
Pick 3 sounds that are really different (clap 👏, whistle 😗, "shhh" 🤫). Make the sound while you tap Capture — do each one several times so the AI gets lots of examples!
🛡️ Your microphone stays on your device — nothing is uploaded.
Once you've taught all 3, make one of your sounds and watch the bars. The AI compares what it hears now to all your examples — you trained a real audio AI!
This is AST — a real audio classifier that learned from 2 million sound clips — running on your device. Clap, whistle, tap the desk, meow, play music... it names what it hears. Record 3 seconds and see! (~90 MB one-time download; then offline.)
You trained an AI to hear — teaching it your own sounds and watching it learn to tell them apart. Ears, trained from scratch!