Daniel in SD
(supervised)
I'll one up you.
I'd bet against L3 in any domain this year.
It would be interesting if a forum had a betting pool that ranked people based on the accuracy of their predictions.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Autonomy Investor Day - April 22 at 2pm ET
- Thread starter diplomat33
- Start date
I'd bet against L3 in any domain this year.
It would be interesting if a forum had a betting pool that ranked people based on the accuracy of their predictions.
Ludalicious
Active Member
Daniel in SD
(supervised)
I'm pretty sure it was L4 in Q2 next year. He did mention provision for remote operation so I think that would make it L4.
See this is the fly in the ointment. Tesla has always impressed on the hardware front, but I can't honestly say I've been impressed on the software front... I keep remaining hopeful but it gets harder with every passing year.
I like the idea of a betting pool just to put money where a persons mouth is.
Maybe for charity.
But, we do need to avoid things that actually require regulatory since those are way outside of Tesla's control.
It really should be able the capability itself. It's pretty easy to tell if something is even ready for regulatory evaluation.
It was Level 3+ Q2 next year early in the presentation. But later Musk was asked about what does FSD feature complete mean at end of 2019 — and he said yes to ”Level 5 no geofence”.
That’s great about this question that Elon answered, this was just about feature complete inside Tesla. Level 5 no geofence end of 2019, this was definitely outside of regulatory approval.
Elon just acknowledged that Tesla would be liable for an accident.
So at least we got that one.
So at least we got that one.
Yeah, and I'd totally bet against that.
All bets would be short term 1-3 year things.
We'd have to have a neutral 3rd party voting/testing committee.
Wow, this presentation definitely looks different depending on one thing: do you believe Musk. 
Daniel in SD
(supervised)
A robo taxi can't be L3 unless you want a drunk guy with a revoked license responsible for taking over. haha.
Watched the whole thing, as people said it was going so well until Elon started talking about Robotaxis.
A lot of the questions were very poor. TBF if you have really good insightful questions to ask as an investor you don't want to ask them in public as that everyone hears the answer, hence you tend to get a lot of pointless questions.
The best question for me was the one re can the person in the robotaxi start driving it and might they have to take over. Elon's answer suggested that the robotaxi thing is ill though out for the moment. This ties in with @Daniel in SD point. All that matters is does self driving work. If it does you can make a lot of money, but does it? If people can get in a car and take over driving it then it isn't a proper robotaxi. It is something closer to a hire car that brings itself to you. Would you need a driving licence to call up a robotaxi if you could take over driving it? Would you be responsible for it? The very vague answer suggests to me that the whole robotaxi thing was just tacked on and not well thought out. This makes me worry that FSD is not as fully functional as suggested.Once they are confident enough to lose the steering wheel then we are talking a real robotaxi.
If the robotaxi numbers are correct, Tesla should stop selling cars pretty soon. Why would you sell a device that you believe is capacity constrained and can generate $30,000 a year in gross profits at a price where you get closer to $5,000 in one off gross profit. For the moment they need the financing and volume, but soon the cars would generate so much cash they would be better using all of the cars themselves.
I have to add, it was very impressive up to this point and I want to believe that FSD is that close. Why doesn't the video continue onto the demos?
A lot of the questions were very poor. TBF if you have really good insightful questions to ask as an investor you don't want to ask them in public as that everyone hears the answer, hence you tend to get a lot of pointless questions.
The best question for me was the one re can the person in the robotaxi start driving it and might they have to take over. Elon's answer suggested that the robotaxi thing is ill though out for the moment. This ties in with @Daniel in SD point. All that matters is does self driving work. If it does you can make a lot of money, but does it? If people can get in a car and take over driving it then it isn't a proper robotaxi. It is something closer to a hire car that brings itself to you. Would you need a driving licence to call up a robotaxi if you could take over driving it? Would you be responsible for it? The very vague answer suggests to me that the whole robotaxi thing was just tacked on and not well thought out. This makes me worry that FSD is not as fully functional as suggested.Once they are confident enough to lose the steering wheel then we are talking a real robotaxi.
If the robotaxi numbers are correct, Tesla should stop selling cars pretty soon. Why would you sell a device that you believe is capacity constrained and can generate $30,000 a year in gross profits at a price where you get closer to $5,000 in one off gross profit. For the moment they need the financing and volume, but soon the cars would generate so much cash they would be better using all of the cars themselves.
I have to add, it was very impressive up to this point and I want to believe that FSD is that close. Why doesn't the video continue onto the demos?
It was an interesting presentation. I liked it.
It also doubled down on everything Musk and Tesla have failed so far with AP2 in a massive way. The gamble just got really real.
Not that it was acceptable ever before (as lots was promised at end of 2016 too) but now we can no longer accept incremental improvements on driver’s aids from Tesla. They must now show Level 5 level robotaxi progress by end of 2019 to meet these expectations.
It also doubled down on everything Musk and Tesla have failed so far with AP2 in a massive way. The gamble just got really real.
Not that it was acceptable ever before (as lots was promised at end of 2016 too) but now we can no longer accept incremental improvements on driver’s aids from Tesla. They must now show Level 5 level robotaxi progress by end of 2019 to meet these expectations.
Don’t normally respond to my own posts, but WOW was it ever worthwhile!
Well done Elon!
Not only that, but he also claimed his FSD would work in downtown Manhattan end of 2019, fully autonomous. Does he even know what is planned? There will be much confusion because of this...
Eno Deb
Active Member
Well, it seems this event didn't really tell us much about where they really are. The technical presentations were good (especially Karpathy's), but mostly on an "engineering for laymen" level (which was probably appropriate given the audience at an investor event). But how close are they really to solving the current shortcomings of Autopilot? From remarks by Karpathy and Musk it sounded like the AI team is still pretty much exclusively focused on vision and perception, while driving policy is still driven by hand-coded heuristics. But the latter is what is currently lacking. The non-answers by Musk and Karpathy when someone asked when the lane merging behavior might improve seemed telling.
Let's hope we get some good reports from the driving demos.
Let's hope we get some good reports from the driving demos.
I will give Elon this: he is consistent on Tesla Network and FSD. This was not a Level 2 cop-out.
This was Level 5 no geofence end of 2019 commitment.
Good for him, if it is the truth.
This was Level 5 no geofence end of 2019 commitment.
Good for him, if it is the truth.
No, but it had been for the AP2/3 narrative in recent times.
We were slipping into Level 2 driver’s aid domain and Tesla’s comms seemed to concur.
Now it is back to Level 5 no geofence this year.
Similar threads
