7 Extra Tesla Full Self Driving (FSD) Views & Observations

For years, the CleanTechnica reader neighborhood has amazed me. The standard of thought, variety of thought, and depth of knowledge is astounding. After I printed a narrative final evening on Tesla Full Self Driving (FSD) progress, I didn’t actually know what to anticipate. It’s a real gray space that shows great progress in know-how and AI but in addition continues to have evident flaws and has progressed much more slowly than folks initially bullish on the know-how anticipated (like me). A number of years in the past, I used to be selling Tesla has having the very best strategy towards creating robotaxis, and I did for years. I’ve been dissatisfied with FSD and don’t see it as anyplace near the place I anticipated it to be at this time in 2020, however I’m additionally clearly open to and desperate to see it enhance. It lastly did make a notable enchancment (to my eyes) in model 11, however I really spent a lot of the article explaining its shortcomings — so, I didn’t know the way readers would reply.

Effectively, after a number of feedback, I noticed an amazing lengthy one which I assumed needs to be a standalone article, after which I noticed one other that deserved to be featured, and one other. … The remark thread is stuffed with nice views, and I gained’t add all of these right here, however I’m going to repost a number of that add a little bit of perspective to my preliminary piece.

From Mark H:

I’m actually amazed by all of it.

I’m amazed that our human brains are so profitable at processing what our optic nerves ship it, in order that we will drive as efficiently as we’ve got.

I’m amazed {that a} not prepared for prime-time beta software program which might actually hurt the driving force and others at any time is safer than the human and his hubris. I actually couldn’t have imagined this at this early stage of growth.

I’m amazed that it’s nonetheless that secure after releasing to a non-professional take a look at fleet of over 400,000.

Whether or not it forces folks driving to pay extra consideration, I don’t know and don’t care. I’m merely amazed that it’s. I do know NHTSA is dying to close it down because of public fears of the pc killing machines. They haven’t up to now because of the information.

These of us testing are conscious of the numerous many flaws that are documented so properly. Add to that’s the video era trying to collect likes by blurring the road between “be careful” and “watch this”. I’m amazed that it’s doing higher than people and our hubris. We’d not be allowed to observe the instruments on the trail to autonomous driving if it weren’t for this. It’s so essential to do not forget that getting us from level a to level b is simply one of many objectives. And it’s secondary to the first aim summed up by Musk in two phrases, “Don’t crash.” Truthfully, I’ve to thank a few of my fellow people for not crashing into me from a false slowdown. Whereas I attempt to disengage when folks strategy, I’ve on a handful of event received caught immediately with somebody tailgating me adopted by a false slowdown.

From ctromley:

What the FSD boosters fail to acknowledge is that any total security enchancment whereas in growth is totally irrelevant. As a result of the one factor FSD enforces higher than something is intense vigilance over what the automobile is doing. As a result of it retains proving time and again that it could actually do one thing sudden, silly and harmful at any time.

It forces the driver, who’s finally accountable, to make sure driving extra security. After all security stats shall be spectacular.

The nice take a look at will come when FSD is sweet sufficient that folks will begin letting their guard down. THAT is when it must make *zero* errors, as a result of the press could have a area day with studies of some defenseless passenger shedding their life as a result of the autonomous automobile made an sudden, silly and harmful transfer that no human with a functioning mind would ever do.

You realize that day is coming.

From Paul Fosse:

My emotions precisely. This newest beta is an actual enchancment, however I’m nonetheless left with the sensation that it may very well be 1 12 months or 10 years to complete the software program. Why? As a result of the progress has been so inconsistent. Everyone knows that freeway autopilot/FSD has been wonderful and just about unchanged (besides it simply modified to the brand new stack not too long ago) for about 5 years. Why didn’t Tesla make that palms free, take legal responsibility for the driving and let folks loosen up on lengthy journeys and do different issues? It might simply have to offer us 5 seconds discover when it was getting confused and it wanted consideration. That might have been an enormous step ahead and Tesla may have executed that 4 or 5 years in the past. Then work on town driving.

From Pitounet:

Individuals will count on much more from a robotaxi, simply take into account what folks count on from regular taxis and you’ll understand that Tesla is way off.

After I take a taxi or a Uber I count on him to do no error ant alol, anny error will make me extraordinarily uncomfortable and Want I didn’t take a taxi. The overwhelming majority of instances meet my expectation.

As soon as I had a Uber make two errors on me on the second error I requested him to drop me within the spot and I requested Uber my trip to be refunded.

It has occurred to me a number of instances that I wasn’t snug in a taxi as a result of the driving force wouldn’t drive like I would like. Both being too blunt or too sluggish at purple strains, it decreased my consolation and I would definitely not take this driver one other time if I had the selection. So not solely does a taxi driver must make no error, but in addition does he must drive “usually”, easily sufficient and assertively sufficient.

It’s techno idolatry to count on folks to be extra tolerant with robotic automobiles simply because they’re managed by software program. The bar for accepting to be pushed by an AI is far larger than you appear to comprehend.

From Daniel Wilder:

Completely agree with the article. Love my Tesla, no regrets about shopping for FSD in 2019, and find it irresistible on the freeway. However utilizing it round city is nerve-racking and it annoys different drivers with it’s non-human-like actions. However I’m having fun with watching it enhance and sit up for each new launch.

From Baird Edmonds:

That is an trustworthy description of my expertise with FSD over the previous 2 1/2 years with my M3LR. Nonetheless many errors in fairly easy circumstances, actually too many to listing right here. I like the automobile and have nice admiration for what Tesla has achieved however FSD is way from prepared for autonomous driving.

From bjrosen:

I’ve had FSD for 4 years, I don’t suppose it’s ever going to work. The most recent beta continues to be horrible. Tesla painted themselves in a nook by claiming that the {hardware} that was out there in 2016 was adequate to carry out stage 5 self driving and promoting it with the promise that they might get there with simply software program enhancements. If that they had simply bought it as enhanced autopilot and left the FSD claims for the longer term when higher sensors turned out there they might have made extra progress. Cameras have been getting higher, radar has been getting higher, FLIR cameras can be found and LIDAR is getting fairly low-cost. Tesla could have higher cameras and a brand new radar in HW4, and possibly that can assist, however it could actually’t be retrofitted into the million automobiles that have already got FSD so they’re caught attempting to do the not possible which is to make it work in automobiles which have insufficient sensor suites. At what level will they admit that they will’t get there with the {hardware} that they’ve.

That captures just about each helpful perspective I can suppose up — and, in fact, many I didn’t. However preserve the dialog going. There’s absolutely extra to say.