Uploaded August 2026 | Updated September 2026, 2 weeks ago
Support the channel by becoming a member at Patreon.com/FutureAzA, youtube.com/@FutureAzA/join, or subscribing on X.com here: Twitter.com/FutureAzA. You can also make a one-time payment on PayPal here: paypal.com/ncp/payment/3AS47G2JNWU9A. Or you can also use my buddy's Tesla referral to get three months of free FSD. https://ts.la/ron42764
Tesla FSD processes sensor data 36 times a second, but simulation alone can't replace real-world driving miles. Phil (ex-Rivian, ex-Apple) joins to break down imitation learning, reinforcement learning, and why training humanoid robots like Optimus is an even harder problem than self-driving cars.
Huge thanks to Phil Beisel for joining the discussion.
Follow Phil on X:
https://x.com/pbeisel
0:00 Sensor Data Input
0:57 Meet Phil Rivian
2:27 Simulation Vs Reality
5:52 Reinforcement Learning Rewards
14:20 Optimus Vs FSD
23:14 Composable Robot Skills
27:34 China Compute Race
30:41 Demo Versus Product
#Tesla #TSLA #EVs
Support the channel by becoming a member at Patreon.com/FutureAzA, youtube.com/@FutureAzA/join, or subscribing on X.com here: Twitter.com/FutureAzA. You can also make a one-time payment on PayPal here: paypal.com/ncp/payment/3AS47G2JNWU9A. Or you can also use my buddy's Tesla referral to get three months of free FSD. https://ts.la/ron42764
Tesla FSD processes sensor data 36 times a second, but simulation alone can't replace real-world driving miles. Phil (ex-Rivian, ex-Apple) joins to break down imitation learning, reinforcement learning, and why training humanoid robots like Optimus is an even harder problem than self-driving cars.
Huge thanks to Phil Beisel for joining the discussion.
Follow Phil on X:
https://x.com/pbeisel
0:00 Sensor Data Input
0:57 Meet Phil Rivian
2:27 Simulation Vs Reality
5:52 Reinforcement Learning Rewards
14:20 Optimus Vs FSD
23:14 Composable Robot Skills
27:34 China Compute Race
30:41 Demo Versus Product
#Tesla #TSLA #EVs
