yes, blowing smoke is very amazing. This vid is probably one of 100 attempts to make a more or less flawless drive. This is amazing when hundreds of users can proof it works most of the time.
This looks very much like old Tesla FSD 11 that was hardcoded and pretty limited. You could do many drives without intervention. And if you just publish the good drives, it looks like done. And that's what's happening here. Blowing smoke. They are nowhere because their approach is not going to work. Neural net is very probably the only way to solve autonomous driving. And if you go the neural net approach, you must have big data. And where is the mobileye fleet that delivers this data?
@@dasstraat This isn't any more impressive than Tesla's latest FSD. And Tesla doesn't require the sort of HD mapping Mobileye relies on. Each approach has its merits, and I applaud Mobileye for what they've achieved, but they were also promising this back in 2017 :)
@@skwdenyerActually the mapping is just an extension to what Mobileye offers. Tesla is limited, because they need to save money as much as possible, on sensors and mapping. Watch Engineering Explained video about them, if you want to learn more. And if mapping is indeed needed - so what? Most likely it's not that difficult to map, if this will work, the business will explode.
Would love to see this at normal speed.
Then you can see why certain drivers lost patience ;-)
The driver's hands almost gripping the wheel says a lot about how reliable the system is XD
No it says a lot about crazy Chinese
Tesla FSD is level 2 same as Li auto. Nio, Xpeng are level 4 and BYD is level 3 along with Huawei Avatr. Waymo robot taxi is level 4
Elon Musk broke the rule of FSD level 2 by diseabling hand on wheel regulation. Only level 5 is full self driving with no hand on wheel lol
Zeekr NZP is level 3
Very nice
Amazing
yes, blowing smoke is very amazing. This vid is probably one of 100 attempts to make a more or less flawless drive. This is amazing when hundreds of users can proof it works most of the time.
Well done Mobileye🎉
Let's do it!!!
This looks very much like old Tesla FSD 11 that was hardcoded and pretty limited. You could do many drives without intervention. And if you just publish the good drives, it looks like done. And that's what's happening here. Blowing smoke. They are nowhere because their approach is not going to work. Neural net is very probably the only way to solve autonomous driving. And if you go the neural net approach, you must have big data. And where is the mobileye fleet that delivers this data?
Hmm.. this is what Tesla promised back in 2017..
Then they broke up with Mobileye and start exploding. Wish Mobileye a bright future.
@@dasstraat This isn't any more impressive than Tesla's latest FSD. And Tesla doesn't require the sort of HD mapping Mobileye relies on. Each approach has its merits, and I applaud Mobileye for what they've achieved, but they were also promising this back in 2017 :)
@@skwdenyerActually the mapping is just an extension to what Mobileye offers. Tesla is limited, because they need to save money as much as possible, on sensors and mapping. Watch Engineering Explained video about them, if you want to learn more.
And if mapping is indeed needed - so what? Most likely it's not that difficult to map, if this will work, the business will explode.
@@skwdenyer Humans use maps too.. and FSD is not even close
That could be remotely driven.