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

HW2 Autopilot - is it just me? or....

This site may earn commission on affiliate links.
Hi Everyone,

This is my first post here. I've got a 90D Model S on order, due for delivery during Feb/Early March - with HW2.

The "is it just me?" title is because I'm a bit concerned that we're in the middle of a major over-promise/under-delivery on HW2 autopilot and I was wondering what others felt about this.

The first promise was (and still is) in the design studio:

"Tesla's Enhanced Autopilot software is expected to complete validation and be rolled out to your car via an over-the-air update in December 2016, subject to regulatory approval".

Then we got a much reduced promise from Elon's tweet on 23rd Dec:

"Looks like we might be ready to rollout most of Autopilot functionality for HW2 towards the end of next week"

Notice he says most of _Autopilot_ functionality (not Enhanced Autopilot). There's a big difference.

Then, what we actually got at the end of last month was limited rollout to a small number of vehicles, with (as far as I can tell) autosteer limited to 35mph. Is this correct? If it is, that's not a useable autopilot compared with the HW1 baseline.

I was rather expecting a bit more in terms of complaining about all of this - but I haven't seen much as yet. Have I got it wrong?

Interested in your views!

Cheers

Hugh
 
Still Waiting!

11525510885_4f65044dc6_z.jpg
 
You have missed another announcement where EM said that HW2 features will come out gradually, added once a month.

Tesla did not promise Dec delivery- only 'expected' it will happen that Month.

Also, it sound like you expect EAP to be better than AP1 off the bat- that will not happen as AP1 is really good and it took months to perfect it.
 
Then, what we actually got at the end of last month was limited rollout to a small number of vehicles, with (as far as I can tell) autosteer limited to 35mph. Is this correct? If it is, that's not a useable autopilot compared with the HW1 baseline.

Let's see how it might go.
Current autopilot is entirely new software with new neural net and as it was said - it needs some millions miles to iron out.
That is just a nature of neural nets - they are very slow to train - not a feature of software as is. From software point of view capabilities for HW1 are almost certainly already here - they just need to be verified and tuned.

Given few dozen thousands of HW2 cars - the 'reduced' autopilot seems be tuned in a month or two, also overpasses will be mapped so there will be solid ground for next iteration with less speed restrictions and more features - so speed restriction is just for the purpose that neural net will be tested and tuned for feature recognitions, overpasses will be mapped etc.

thus, the point here is what mostly affects new features - is a 'speed' of learning of complex neural net. not a code itself. With that understanding you might expect that by March there will be a new iteration of software with more capabilities ( and more cars to fleet learning ),

if those calculations are correct - then HW1 baseline capabilities similar to those which current hardware allows might be available by April/ May but there will be some more features. From there HW2 must be gradually getting more features.
 
Hi Everyone,

This is my first post here. I've got a 90D Model S on order, due for delivery during Feb/Early March - with HW2.

The "is it just me?" title is because I'm a bit concerned that we're in the middle of a major over-promise/under-delivery on HW2 autopilot and I was wondering what others felt about this.
....

I was rather expecting a bit more in terms of complaining about all of this - but I haven't seen much as yet. Have I got it wrong?

Interested in your views!

Cheers

Hugh

It's the way that Tesla is. They set very aggressive expectations for themselves, and generally meet them, but never on time. I think that that's how Elon likes to drive his team. He promises crazy things, and they they work like mad to almost get there... Usually...

Other automakers miss deadlines as well - reliably and repeatedly. But they don't publicly announce what they're doing and so people don't know about it and don't get excited about it.
 
The whole roll out has me flummoxed. I have AP1 and the most recent update language that I have read seems to indicate that the AP2 cars had no AP at all from launch and are now only getting less that what I already have, like autosteer only up to 35 MPH or some such craziness. I can go 85 MPH with AP1 on with no issues at all.

I would have thought that they would have rolled out the AP2's with at least what the AP1's already have.
 
The entire intent of eAP and FSD seems to be about over promising.

Where it all comes down to regulations.

There isn't anywhere within the entire Tesla site where they promise you anything, on any timetable with eAP or with FSD.

All you know is if you have the hardware, and bough the package(s) that you'll see extra features every release. At this point we have no idea when they'll reach parity with AP1. It could be in 3 months, and it could be in 9 months.

But, if I bought eAP I wouldn't really care about being on parity with AP1 because AP1 isn't really all that good. But, to do really do better likely requires some level of regulatory approval.
 
It's the way that Tesla is. They set very aggressive expectations for themselves, and generally meet them, but never on time. I think that that's how Elon likes to drive his team. He promises crazy things, and they they work like mad to almost get there... Usually...

Other automakers miss deadlines as well - reliably and repeatedly. But they don't publicly announce what they're doing and so people don't know about it and don't get excited about it.

This is absolutely correct. I hope musk continues to set aggressive goals. Tesla coming just under their aggressive goals, means they still beat others who achieve their mediocre goals.

Tesla should continue to set aggressive goals, even if they still miss them repeatedly. If they start always achieving their goals, then we know they have become complacent and are aiming too low.
 
I am a big Tesla Fan owning now 2 Tesla's the latest one with HW2 even would like to update our P85 to the Hw2 and AP2 once it is available . Know many people including myself are focused on the auto drive capabilities but the safety aspects of all the sensors and the ability to prevent and warn is even more important and a great reason to buy. Wish that we could get the update soon and perhaps prioritize the safety aspects of the suite to be implemented first .
 
I was rather expecting a bit more in terms of complaining about all of this - but I haven't seen much as yet. Have I got it wrong?

Just keep reading some of the recent threads in this forum, you will find a whole new world of complaining (me included) about this issue.

In fact, I'm surprised you started another new thread about this same subject that has been hashed and rehashed without reading what has already been posted on multiple threads and several pages of posts.
No disrespect intended, welcome to Tesla!
 
  • Like
Reactions: kavyboy
The whole roll out has me flummoxed...I would have thought that they would have rolled out the AP2's with at least what the AP1's already have.

Well, let's recall two things:

1 - AP1 took a full 12 months between hardware shipping and 7.0 being turned on in October 2015. That mean AP2 is rolling out on a much faster schedule than AP1 did.

2 - Remember, AP1 used a curated vision system trained by images annotated by human beings at Mobileye. As far as we know, Nvidia's solution is much less supervised learning powered by much more powerful neural nets. But it is a different system entirely. Amnon Shashua has publicly derided Nvidia's non-annotated learning approach several times at conferences. With Mobileye Tesla had a "vision in a box" kit ready to go on some level, but with Nvidia they are starting from zero. And despite this fact Tesla still seems to be rolling out AP2 on a faster schedule than AP1 was.

The entire intent of eAP and FSD seems to be about over promising. But, if I bought eAP I wouldn't really care about being on parity with AP1 because AP1 isn't really all that good. But, to do really do better likely requires some level of regulatory approval.

I think AP1 is amazing. I guess I'm easily amused and impressed.
 
  • Like
Reactions: EinSV and CHG-ON
In a nutshell, Tesla does a lot more over-promising and under-delivering than they do under-promising and over-delivering. This is especially true with deadlines. There are countless examples... and I'm really not sure why they keep doing this to themselves. The worst example of this (IMO) is the Kwh rating on some of the batteries... but there are many others. With a car this awesome and very little to no competitors that can offer any products even close to equivalent, there is no need to over-promise and under deliver... and yet, they continue to do so time after time. It's gotta be an EM culture thing. I truly don't get it.
 
  • Like
Reactions: u00mem9
In a nutshell, Tesla does a lot more over-promising and under-delivering than they do under-promising and over-delivering...There are countless examples... and I'm really not sure why they keep doing this...I truly don't get it.

Elon has publicly stated multiple times why he operates this way. He sets nearly impossible goals - shoots for the stars - lands on the moon. His "underdelivering" is only in relation to his own impossible goals - he way OVERdelivers compared to anyone else. I think it's funny that it annoys so many people who feel the need to type about it on message boards - meanwhile conveniently forgetting that Tesla's sales and revenue growth says everything you need to know about whether his "Publicly commit to the impossible in order to motivate ourselves" strategy is working. It's a strange kind of psychological fanboyism - like they've been wounded by a lover and had their pride hurt or something. If you want to be happy then ignore what Elon says and just focus on the product he keeps delivering to the world.
 
  • Like
Reactions: oktane and OBX John
Based on my own experience with AP1 and what was promised and where it is at today.....I'll wait to buy my next Tesla bases on what the technology does at the point of purchase and not what has been oversold and over promised just from he underwhelmed. I AM a fanboy but after what was promised by Elon at the "D" event and what we actually have....I'll wait to buy my next one once the included hardware is software functional and does what I want to pay for.
 
Elon has publicly stated multiple times why he operates this way. He sets nearly impossible goals - shoots for the stars - lands on the moon. His "underdelivering" is only in relation to his own impossible goals - he way OVERdelivers compared to anyone else. I think it's funny that it annoys so many people who feel the need to type about it on message boards - meanwhile conveniently forgetting that Tesla's sales and revenue growth says everything you need to know about whether his "Publicly commit to the impossible in order to motivate ourselves" strategy is working. It's a strange kind of psychological fanboyism - like they've been wounded by a lover and had their pride hurt or something. If you want to be happy then ignore what Elon says and just focus on the product he keeps delivering to the world.

I've come to realize this is Elon's strategy - his promises have more to do with motivating his employees than misleading his customers.

However, if we didn't complain on public message boards when deadlines are missed, then we wouldn't be doing our part to motivate his employees! If everyone acts like the promises don't mean anything, then where's the pressure to perform?

So I say we should complain away, knowing that we all must play our parts in keeping Tesla #1! ;)

Now, where's my AP2??? :mad::rolleyes:o_O:)