ME: "Guys...I'm an aerospace engineer and scientist with experience in multi-sensor fusion and have owned a Tesla Model Y for over two years. I'm telling you this car is getting worse and worse without radar being actively used and that "vision only" is a bad idea."
Look up "appeal to accomplishment" and please refrain from listing your credentials as a way to convince people you're right. There are likely lots of accomplished people on these forums. Not saying you're wrong; you're just not convincing people the way you want to.
A THOUSAND MUSK FANBOYS ON HERE: "You just don't understand Elon! You didn't watch 'AI Day' and don't understand his brilliance! Our Teslas try to kill us only...er...rarely! People only have eyes, so...!"
Agree there are people who drink the koolaid and toe the party line. But I also see "A THOUSAND" opposing viewpoints saying that their car is worse without radar, and plenty of people suspect the true reason for radar removal was to adapt to a parts/supply chain shortage. I would agree with that, and to help make the decision easier, fusing garbage data from the low-res radar was counterproductive in some situations.
ME: *continues to watch my radar-equipped Model Y autosteer becoming unusable at night and in the rain with "Tesla Vision" updates...*
I haven't noticed any weather-related performance issues post-radar removal. It was always hit or miss. If weather is so bad that vision is continually impaired, there's no other sensor out there that can take over as a redundant sensor. In other words, if vision is down, you can't drive on radar, even though it has the advantage of seeing through rain. It unfortunately can't see lanes, stop light colors, or read signs.
For me, highway overpass and signage AEBs were completely eliminated when radar was disabled. Great! But the car is now slower to recognize slowdowns ahead, and it brakes late. I speculate that was due to loss of radar. And the following-speed logic feels all over the place. I also miss detection of the car in front of the lead car, again leading to late braking.
AUTO INDUSTRY: *continues to improve multi-sensor fusion and drive radar and lidar costs down*
TESLA (Dec 2022, see below): "Umm...we're putting radar back in...."
ME: HAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHA
Part of Elon's argument all along is that extra sensors cost a lot of money, and they wouldn't be needed IF the vision and "brain" were good enough. He cites humans as an example. I don't necessarily agree with all that, but I didn't have an issue with his approach.... except now it's clear the vision and "brain" are improving fairly slowly, and people get impatient based on his FSD promises. There could come a time when the costs of these extra sensors do come down enough that Elon's willing to use them. It seems like this HD radar might be an example of this. I personally think it will be limited to just the new robotaxi model, and it won't be coming to the S3XY fleet for a while, due to the cost.
As much as I'd love a vision-only solution (you have to admit it would be a very elegant solution IF it worked well), I am a big fan of adding HD radar back into the mix. But, if FSD still kinda sucks with the new radar after a year or two in development, they have no more excuses. Perhaps their software prowess is not as good as they claim.
