[loandbehold]

It doesn't need to be superior, it just needs to be good enough. It's something that has been proven in many technology battles. FSD has made an amazing progress over the last two years. If they solve last 1% of cases they will be in good position to beat Waymo on price as they scale. With ride sharing price is everything. People just go for cheapest rides. That's why cybercab is two-seater, they are trying to minimize cost per mile. Smaller car means cheaper car and less energy per mile.

[monetus]

I agree on your point about good enough. I don't see how they defeat foggy weather and the like with cameras alone however, so I'm wondering if they will ever get that last 1% as things stand. I think they need more sensors to stand a chance at that.

[tbcj]

My model Y has trouble with sunny weather today, so even that isn’t quite solved. Driving east into the morning sun constantly gives warnings that the camera needs to be cleaned.

[_DeadFred_]

I don't think people think through to the end enough. Lidar can't be tricked as easily. If Tesla were to attract a group of malicious actors that hated Tesla I wouldn't put it past those actors to figure out ways to confuse Tesla's non-lidar self driving to make it frustrating to Tesla owners but more specifically to cause financial impact on a self driving taxi fleet that wouldn't be possible with Lidar.

Has Tesla accounted for malicious actors trying to sabotage it's vision choice and not just plain jane driving scenarios?

I think at scale and once rolled out malicious actors are going to be one of the number one economic impacts of self driving taxi economics.

[acomjean]

I’d argue “Good enough” is different when you’re selling to someone a life critical system.

[JumpCrisscross]

> doesn't need to be superior, it just needs to be good enough

And not hated. I’m personally pressing for heightened street legality requirements for cameras-only self driving in California, New York, Canada and the EU.