Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Steve Wozniak doubts fully self-driving vehicles are 'quite possible yet’

This site may earn commission on affiliate links.
Well no one expects them "quite yet". The most optimistic predictions from specialists like MobilEye for the first lvl 4-5 systems are around 2021, in tech-time that`s 2+ generations.

Mobileye knew how to do good PR job but it did not remotely have any self driving technology. Back in the days Tesla was using Mobileye's vision system but was developing its own Autopilot system. Mobiieye did not even believe AI machine learning should be used. It was only using manual input for image recognition (see the link below). The two eventually got a fall out when Mobileye learned Tesla was developing the machine learning technology on its own and did not want to share it. BTW I think Intel should have a rude awakening now that it has paid a hefty price for nothing.

Mobileye Bullish on Full Automation, but Pooh-Poohs Deep-Learning AI for Robocars
 
Hopefully he is wrong. (My FSD order went through today, lol.)

I am sure we are at least a year off from the first FSD feature release, but hopefully not much more than that?

Full out FSD definitely does seem many moons out though.

Depends on how you define "self driving". You'll note that Tesla does not quantify it in terms of actual autonomous driving levels which have defined characteristics.

IMO real FSD is probably at least 3-5 years away, maybe more, but that doesn't mean Tesla can't deliver some partial level of autonomy on highways and call it FSD.
 
  • Like
Reactions: super20g
Here's my take. I have a P3 with EAP and FSD. I've experienced the software to date. It works well in some situations and not very well in other situations. In my mind, I see the pathway to FSD vehicles with three components. One component would be the cars themselves and all of the features (hw and sw) they bring to the table, and the second component has to do with standards about how the roads are striped and signed in a standard way for all FSD cars to use them. The third component is the regulatory treatment making it all possible, which is starting to move forward.

Yes, some of the features work now, and they are interesting and fun to use. But I see a lot of work ahead (and time) before we can say that a car has full self driving features in any scenario. There are just too many variables that need to be worked on and overcome. I will be there with my car and my support of Tesla. But I am the first to acknowledge that we have a long way to go to make all of this technology safe and usable in all the various driving scenarios there are out there....
 
I think FSD is "right around the corner". Tesla test cars are usually six months ahead of what's in the fleet and we hear the computer with the new chip is testing. I think the EAP and FSD are progressing fast. I have a local residential street that has about a 150' yellow center line at an intersection. It is unmarked after that. If I accelerate fast enough to 18 mph driving from that intersection, I can briefly get the gray steering wheel and turn EAP on while it "sees" the line. It continues to drive down the street perfectly with no lines. The street is macadam with no curbs and only grass on the edges. When driving on these roads without EAP on you can see faint grey lines but it currently needs a road marking to enable EAP. It only needs one line. Have faith, this is going to work, IMHO.
 
It seems pretty obvious that FSD is going nowhere.

Indeed I strongly suspect that Tesla was initially planning to release “drive on autopilot” as part of FSD. Remember those talks about FSD features “coming out this summer” announced a few months ago by Elon?
 
It seems pretty obvious that FSD is going nowhere.

Demonstrably untrue.

Why would Tesla go through the time, trouble, and cost, of developing HW3, which is only needed for FSD, if it was "going nowhere"?


I
Indeed I strongly suspect that Tesla was initially planning to release “drive on autopilot” as part of FSD.

Also demonstrably untrue.

The functionality of drive on autopilot is explicitly part of EAPs promised features (on ramp to off ramp), and have been since EAP was announced in 2016.
 
  • Like
Reactions: super20g
The interview felt like Woz was trying to prop up GM every chance he got. While I don't think he was paid by GM at all, someone definitely called in a favor. "Hey Woz, we're getting beat by Tesla, could you say something good about the Bolt kkthxbye"

Yeah sorry Woz, Bolt is not comparable to a Tesla in any shape or form. Plus I highly doubt someone of Woz's wealth, age, and shall we say...size, would drive a Bolt much. An econobox with cheap interiors, he's just not going to comfortably fit in there. When people age, the last thing they want is to sit in crammed cars. This is why I highly doubt Woz propped up the Bolt because he just loooves to drive that car.
 
Funny - I passed what seemed like two Apple self driving cars the other day on the highway in TN that were in a caravan. One looked like it had no driver at all in the front two seats.

And Tesla and others have provided amazing videos of self-driving cars doing the completely autonomous thing while navigating suburban streets, stopping at stop signs and lights, stopping for pedestrians.

Of course it's in perfect weather.

And during the day.

And the area where it is operating has been mapped out beforehand.

And there is a $5,000 LIDAR sensor being used to scan things out to several hundred meters.

FSD is going to be very gradual and it is highly unlikely that even next year's 3rd generation AP system will be able to operate the car completely autonomously in the majority of situations.

A use case that I'm almost positive Tesla is working for as a 1st step of FSD is navigate on autopilot where the driver is freed from having to interact with the vehicle at all from exit to exit. Tesla will introduce something like this and bill it as FSD, which it is, but only for limited use case.

You would activate it as the car is approaching the highway and the car would then drive itself to your exit ramp and you'd be free to do other things.

It will still fail and require the driver to take over in MANY situations that can and do happen to drivers on a weekly or monthly basis such as inclement weather, traffic detours, multiple blocked lanes, construction situations, etc.

Just my prediction.

The fabled FSD car that takes you to the shopping mall completely on its own, parks itself, comes and gets you when you are ready to leave, and then drives you home while you play Pokemon and recover from your two martini lunch, I'm guessing that is 5 years away at a minimum. Probably more like 7-10 years.
 
The only people I would believe when hearing a claim that full self driving is not possible would be lawyers and politicians; because that is where the burden really lies... regulations.

I’d rate them even over insurers because insurers will insure anything they can make money off of, even if it doesn’t even exist.

But politicians will block anything if they can make money doing so, even if it does exist... :p
 
The only people I would believe when hearing a claim that full self driving is not possible would be lawyers and politicians; because that is where the burden really lies... regulations.

I’d rate them even over insurers because insurers will insure anything they can make money off of, even if it doesn’t even exist.

But politicians will block anything if they can make money doing so, even if it does exist... :p


I keep seeing this- but many states already explicitly allow FSD cars to one degree or another.