This is what Tesla Autopilot sees using neural networks that take 70,000 GPU hours to train and output 1,000 tensors (predictions) at each timestep. Source: http://www.tesla.com/autopilotAI pic.twitter.com/eBzvnLJtRn
-
-
If it can be fooled someone is there it's just a matter of time before a powerful enough projection fools it into thinking nothing is ahead of it when there is. This doesn't even require hacking. That's another issue.
-
Again humans can also be fooled.. probably radar tech can do better job.. so I believe autonomous can only get better as days go.
Kraj razgovora
Novi razgovor -
-
-
definitely a key question. perfection would be nice but we probably shouldn't expect that to be a valid goal.
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi
-
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.