Frankly I’d be even more peeved about the lack of a rain sensor with Model 3. With the Model X at least the manual operation is OK for the wipers.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Coast to coast drive happening this year for all FSD Teslas!
- Thread starter Daniel in SD
- Start date
Frankly I’d be even more peeved about the lack of a rain sensor with Model 3. With the Model X at least the manual operation is OK for the wipers.
You may want to listen to this: Mobileye is not as far ahead as you believe them to be:
Autonocast: #144: David Fritz and Jim McGregor on Chips For Autonomous Vehicles on Apple Podcasts
Listen at 20:25
To be fair, no FSD features have shipped, either. If there isn't a larger NN on the way, then Tesla wasted a whole lot of money for no reason when they decided to send HW3 to mass manufacturing. So I think the very fact that HW3 is shipping in vehicles is a fairly solid indication that larger NNs exist in the lab that aren't released to the public yet.
EVNow
Well-Known Member
He has talked about unprotected left turns and how they are getting better.
I'm seriously astonished people even think Musk makes promises ! He just does wishful thinking aloud.
But then, as Musk has said many times, if he wasn't eternally optimistic, he would have never started an EV company or for that matter SpaceX.
EVNow
Well-Known Member
Its obvious larger NN exists - we just don't know how much better it is
BTW, they say they are saving cost using HW3 compared to current h/w.
Dr. J
Active Member
Speeding up the demo video masks how slow the car is moving and "thinking," which is the condition that triggered the following driver to make the impatient and nearly deadly pass on the right. Some might consider masking actual speed as a "faked" video. Just sayin'.Tioga pass is closed in the winter so I don't think any vehicle short of a snowcat could make it
I have no idea how well Cruise's vehicles work in the snow or rain. They do most of their testing in San Francisco so I bet they work pretty well in the rain. I've seen video of them driving at night.
Tesla claims that their system will work everywhere and in all weather conditions by the end of next year.
Speaking of driving at night, I found this pretty hilarious: (at 31:27 it seems to strip time links)
Its obvious larger NN exists - we just don't know how much better it is
BTW, they say they are saving cost using HW3 compared to current h/w.
Maybe even low triple-digit dollars. But they're also taking on a lot of risk if they screwed something up, and given how critical that part of the car is, I'd bet every update gives their risk analysis folks absolute fits.
EVNow
Well-Known Member
Tesla has risk analysis folks ?Maybe even low triple-digit dollars. But they're also taking on a lot of risk if they screwed something up, and given how critical that part of the car is, I'd bet every update gives their risk analysis folks absolute fits.
Oh I don’t disagree stuff exists in the labs including better NNs.
I expressed doubt whether the stuff in the labs would be dramatically different though in its features — let alone approaching the level needed for Level 3-5 self-driving.
But thanks for the reminder about Tesla’s promise on FSD features shipping by summer 2017.
That is not exactly what I meant but not really important to dwell on.
My point was: Tesla would be talking Waymo level problems if they reall were on that level. Instead they talk about quite basic stuff.I don’t think Musks intent is to promise. His intent is to market — and in doing so he ends up misleading the customers all the time about features, their availability and timelines.
The guy (David or Jim) is clearly clueless. Which isn't surprising. Alot of people ain't aware of the actual development of neural network and deployment.
Last edited:
EVNow
Well-Known Member
Actually even this is not entirely correct.
He has this "I'm a futurist. I can predict stuff in future no one can think about." attitude. He just wants to be seen as clever and funny. This recent tweet captures what EM is all about.
Again I believe — well, I guess — Musk’s intent is to market. Sometimes he is marketing Tesla, other times SpaceX and sometimes just himself.
diplomat33
Average guy who loves autonomous vehicles
Marketing is certainly part of it. Obviously, Musk is going to want to present his companies like Tesla and SpaceX in a positive light. But I do think Musk is a pathological optimist. What people say about themselves is often revealing. And Musk has admitted many times that he has a "problem with the concept of time" and just recently in the shareholder meeting said that he would not be in this business if he would not an optimist. So Musk has told us that he is a super optimist that struggles with realistic timelines. I think that is the primary driving force why Musk makes bold claims about FSD for example. Musk seems to be a "if you build, they will come" kinda guy. He has bold ideas that seem impossible and then pushes hard to make it happen. It's what makes him a visionary to many while also pissing off and frustrating others when he misses his target.
@diplomat33
I just don’t believe that. To me it is too likely an excuse that allows him to continue marketing with statements about impending events that create hype and pull demand levers. When it doesn’t pan out he can just wave the sorry-an-optimist card and keep doing the same... too big of a chance that this is what happens in my view to ignore.
Musk is obviously smart. He probably wouldn’t keep making the same mistake on the same thing so many times over if there wasn’t a benefit for him to keep making it.
I just don’t believe that. To me it is too likely an excuse that allows him to continue marketing with statements about impending events that create hype and pull demand levers. When it doesn’t pan out he can just wave the sorry-an-optimist card and keep doing the same... too big of a chance that this is what happens in my view to ignore.
Musk is obviously smart. He probably wouldn’t keep making the same mistake on the same thing so many times over if there wasn’t a benefit for him to keep making it.
diplomat33
Average guy who loves autonomous vehicles
We might be getting closer to the coast to coast demo. Back in Sept 2018, Musk said that a coast to coast demo would probably require V10 alpha build. Tesla's fully self-driving cross-country road trip to 'probably' happen with version 10 'alpha build', says Elon Musk - Electrek
Well, yesterday, Musk tweeted that V10 is going to early access in August and if all goes well, is expected to go wide release as early as "end of august". Elon Musk hopes to release Tesla’s V10 firmware by end of August, will include Smart Summon
So looks like V10 that Musk said would be the version that could do the coast to coast is coming soon. V10 is expected to bring improvements to NOA on highway and traffic light and stop sign response.
Well, yesterday, Musk tweeted that V10 is going to early access in August and if all goes well, is expected to go wide release as early as "end of august". Elon Musk hopes to release Tesla’s V10 firmware by end of August, will include Smart Summon
So looks like V10 that Musk said would be the version that could do the coast to coast is coming soon. V10 is expected to bring improvements to NOA on highway and traffic light and stop sign response.
I've driven cross country countless times and have trouble even considering 'supervising' a drive of that length. Possibly with two or more supervisors but then they have to multitask to supervise each other also. No, its just not possible imho (the mind boggles).
--
--
CarlK
Active Member
I dig your optimism, let's hope V10 brings some significant improvements.
If they just fix the phantom braking I will be happy. Also the map is off from my cursor.
Similar threads
