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

Unveiling on Oct 17th

What's your best bet on what Tesla will reveal on Oct 17th?

  • Autopilot 2.0

    Votes: 233 36.4%
  • Model 3 - Part 2 (better-looking face, Head up display driving w/o wheel etc)

    Votes: 65 10.2%
  • Non-performance 100D (world's longest range EV)

    Votes: 100 15.6%
  • 65 kwh/ 80 kwh battery for Model S/X (60/75 will be removed)

    Votes: 11 1.7%
  • Model Y (small SUV)

    Votes: 72 11.3%
  • Tesla Semi

    Votes: 15 2.3%
  • Tesla Bus

    Votes: 12 1.9%
  • Tesla Pick-up truck

    Votes: 11 1.7%
  • Faster/better supercharger (200 kw charging, or charging snake)

    Votes: 52 8.1%
  • Iron Man metal suit

    Votes: 69 10.8%

  • Total voters
    640
This site may earn commission on affiliate links.
I don't understand you. How can more time be needed for "running more corner cases, difficult road/traffic scenarios etc." -- do you expect them to show the vehicle on the road driving during the announcement? I don't...When AP1.0 came out in Sept 2014 it took over a year, until Oct 2015, for the software. I can't see that happening again, but at the same time I don't see how they can need a couple more days for the software when no car actually needs it.

I expect them to show a demo on the road or at least a closed course, yes. In September 2014 they did in fact do live demos - I was there and rode in the cars on the closed course. The reps most definitely removed both hands from the steering wheel as the car steered itself, slowed down for signs, etc.

Given that precedent I expect them to do some kind of demo, yes.

On the other hand you could be right also - now that I've seen the exact wording of Elon's tweet which you linked to. You are right - the verbiage taken alone does refer to a refinement of an announcement.

But again on the other hand, I don't know why people can't envision last minute scrambling to finish a software product for demo when/if the schedule was already very aggressive to begin with. Again, neural networks take time to train. Nvidia has stated that its Drive PX2 takes an entire day to train (which is very fast, according to Nvidia, compared to prior, slower hardware) to drive with nothing more than following a human's steering input. That's a full day.

Tesla is not necessarily using Drive PX2 in any case - they are using CUDA architecture but likely a less powerful, slower version. Drive PX2 is a development tool but the latest reports say that the shippable product which OEM's will actually put into production cars will not even sample until at least a year from now.

Obviously Tesla is shipping AP 2.0 sooner than a year - so is probably using some slower hardware which, if it learns to drive completely end to end with no human-curated image annotation (ie Mobileye) may in fact take several days to train.

If Tesla is scrambling around last minute on sensor fusion, etc. and they are tweaking the logical weighting of different sensor inputs then perhaps they are training and re-training the network.

Remember that right before 7.0 was scheduled to finally drop, Elon tweeted that a few more days/weeks were needed while they ironed out some "corner cases" - which turned out to be I-405 in Los Angeles with poorly marked lanes. Months later Elon stated in an interview that they were scrambling to find solutions to keep the cars in lane on that very poorly marked freeway - and that initially the cars were not trained correctly because of human driver mistakes - and had to be re-trained and refined.

I could be totally wrong. Probably am.
 
The better question is how someone couldn't understand how announcements need refinement. Timeframe could be questioned, but this? High profile public communication needs constant refinement - often even during the presentation.

Maybe this lack of understanding is why my assistant frequently tells me I am not allowed to interface with people outside our company before she reads/approves my outgoing e-mails...

heh.
 
The better question is how someone couldn't understand how announcements need refinement. Timeframe could be questioned, but this? High profile public communication needs constant refinement - often even during the presentation.
I don't understand how such an announcement might need refinement.

I do understand how a new product might need refinement.

Rest-In-Pease to the APPLE guru - however he knew all about revealing something that does not work - LIVE on stage. Or was that Bill Gates?

Who was that that got a BSOD during a failed OS reveal?
 
I think it's obvious. The announcement is AP2.0 and Elon summoned a car from NYC to come pick him up in Fremont, but it got a flat tire outside of Salina, Kansas and had to stop to get it fixed. The engineers (foolishly) didn't give the Tesla a credit card to pay for the repair, so it had to stay in Salina and wash dishes for two days before it could get back on the road to Fremont.
 
I think it's obvious. The announcement is AP2.0 and Elon summoned a car from NYC to come pick him up in Fremont, but it got a flat tire outside of Salina, Kansas and had to stop to get it fixed. The engineers (foolishly) didn't give the Tesla a credit card to pay for the repair, so it had to stay in Salina and wash dishes for two days before it could get back on the road to Fremont.

Finally we have an explanation that makes sense. You win.
 
I think it's obvious. The announcement is AP2.0 and Elon summoned a car from NYC to come pick him up in Fremont, but it got a flat tire outside of Salina, Kansas and had to stop to get it fixed. The engineers (foolishly) didn't give the Tesla a credit card to pay for the repair, so it had to stay in Salina and wash dishes for two days before it could get back on the road to Fremont.
There ya go. The start of summon-a-fleet Tesla cars.
 
calisnow, not sure why you're going all in on AP 2.0 and training neural networks when they haven't even sent any invitations for these presumable demos. The two-day delay could literally mean anything from polishing Powerpoint slides to getting legal clearance for some language for literally any product imaginable, including a Tesla bank/financing product that enables buyers to lease a car + solar + storage for a single monthly fee (which is my personal bet at this point).

To me, the last minute nature of the announcement and the lack of invitations are the biggest strikes against this being AP 2.0 considering how big of a news event that would/should be.

Of course, I could be wrong as well. )
 
calisnow, not sure why you're going all in on AP 2.0 and training neural networks when they haven't even sent any invitations for these presumable demos.

Probably because I've been hoping/waiting for AP 2.0 for a long time and I want it to get released into the wild very, very badly so I can order another Tesla. So, as humans tend to do, I interpret patterns to fit the outcome I hope to see.

Also a large deal I'm working on is going sideways (well, a large deal for my tiny company anyway) and my brain gets much more enjoyment hoping that Christmas hits on Wednesday than it does thinking about my work headaches.

Speaking of which - I should get back to a spreadsheet...
 
I vote that it might be an A-B / B-A Driverless Tesla (uber) type car fleet.

I like your vote but I fear it's too aggressive even for Tesla Vision. Driverless Uber would mean Level 5 and even I don't dream we are that close to full autonomy.

...including a Tesla bank/financing product that enables buyers to lease a car + solar + storage for a single monthly fee (which is my personal bet at this point).

You make sense. I don't like it, but you make sense. You have the dark horse bet but it would gel with the Solar City stuff coming up in a few weeks.
 
  • Like
Reactions: DIL
Probably because I've been hoping/waiting for AP 2.0 for a long time and I want it to get released into the wild very, very badly so I can order another Tesla. So, as humans tend to do, I interpret patterns to fit the outcome I hope to see.
Good scientific explanation. :)

If I were holding off buying an S/X for AP 2.0 (which I would be at this point if I were going to buy one), I'd be rooting for the same.

In your favor, I do think the post-blowout quarter timing would be right to introduce a new product such as AP 2.0 hardware that may, temporarily slow the production line as they switched over. And, the timing somewhat matches Elon's "end of the year" mention. But, I would just think that it'd be bigger news than a last minute announcement. In fact, the biggest impact would be for it to be announced simultaneously with the Model 3 part 2 reveal which would make it an iPhone "are you getting it yet?" moment, though I don't think the two timelines will align for that to happen.