I use experiments to answer questions I can't answer by thinking about them.
Build something. Test it. See what happens.
Testing whether realtime AI can keep a conversation natural, maintain a valid restaurant order and know when to involve a person.
Explore the experiment →
Can Wi-Fi count people in a room? I'm investigating RuView's claims, starting with an empty room and one person. No test results yet.
Explore the experiment →
Testing editorial judgement, not just audio processing. Source selection and all production tests are still ahead.
Explore the experiment →Make enough to answer the question.
See what happens when people use it.
Continue, change it, or stop.