The next big thing in AP 2.0

[OilSucks]

I don't even care about FSD at this point. Sure I'm a little upset I spent that much money to future proof a lease car (I know, I can hear all of you laughing lol) but I just want AP2 to be on the same level as AP1. At the very least, please please please give me the ability AP1 has to 'see' cars in adjacent lanes and classify them. It made me feel safer and more comfortable with the AP system knowing it could see what I could.

It really sold me on the AP package when I test drove an AP1 car (not knowing it was an AP1 car mind you) and excitedly added the EAP option to my car. It sucked finding out that the car I wanted really was not the car I bought though.... I'm giving them a year from June 2017 to bring it up to AP1 speeds or I'm calling and demanding some sort of financial compensation for being sold something that isn't as advertised.

 

[buttershrimp]

I don't even care about FSD at this point. Sure I'm a little upset I spent that much money to future proof a lease car (I know, I can hear all of you laughing lol) but I just want AP2 to be on the same level as AP1. At the very least, please please please give me the ability AP1 has to 'see' cars in adjacent lanes and classify them. It made me feel safer and more comfortable with the AP system knowing it could see what I could.

It really sold me on the AP package when I test drove an AP1 car (not knowing it was an AP1 car mind you) and excitedly added the EAP option to my car. It sucked finding out that the car I wanted really was not the car I bought though.... I'm giving them a year from June 2017 to bring it up to AP1 speeds or I'm calling and demanding some sort of financial compensation for being sold something that isn't as advertised.

I totally agree with that, and I think the June 2018 time frame is reasonable. I'll go so far as saying that if a big leap forward doesn't happen by the end of the year, I'd be very surprised. It won't take long for the owners to start demanding their money back on FSD and EAP rather than complaining that it's not there yet.

 

[buttershrimp]

It seems that the way it was introduced to the AP2 team is that Elon got this crazy idea, and introduced it for the team as a goal with a very steep progress curve and a great height of failure. Result is that even though the work might have been doable, several key engineers quit due to pressure.

Of course if the AP team was involved in the decisionmaking and with setting the timeline, they would feel that they have been a part of the decision and be motivated to deliver it on time. They would view the Autopilot as their baby, and not a threat from outside.

That is if the public statements made by the engineers that quit is true. If it's true, it's clearly a basic management mistake #1: Top-down management, and not involving the correct people in the decision-making.

I agree with the general principles regarding delivering complex products, but I think that Elon has a well established history of making optimistic predictions, only to take longer. I also am not a fan of hero worship, but I will say this about Elon. He's really really really smart, more Albert Einstein than Henry Ford. I don't think any of us doubt that he is a force of nature even though he is also human ( at least for now .

Everytime I think Elon is seriously underestimating the complexity of an issue... I remember an interview he gave regarding a person coming in and trying to sell him a "superior" self-driving technology.... and his response at the time told me that he is, in fact, in touch with the complexities of these issues, as he seemed to have a healthy respect for how smart his group of engineers are and how hard they work.

Perhaps the most impressive thing about him is that his behavior tells me that he understands he is not going to be the person who invents a self-driving car, but he is confident that he is going to be one of the key people forcing that achievement into a reality.

We are all fortunate that some math geek who was a little too fascinated by Ender's Game and Babylon 5, made a massive amount of money on PayPal, otherwise, I have a feeling he'd be like most other tech billionaires, who did some impressive things but didn't actually manage to do something bigger.

That being said, Elon is going to be like most people in the world, he'll continue to try to do things for reasons he's mostly unaware, and he'll continue to be his own worst enemy (and best advocate). I, for one, am really glad that dude is doing what he's doing.

 

[number12]

I don't even care about FSD at this point. Sure I'm a little upset I spent that much money to future proof a lease car (I know, I can hear all of you laughing lol) but I just want AP2 to be on the same level as AP1. At the very least, please please please give me the ability AP1 has to 'see' cars in adjacent lanes and classify them. It made me feel safer and more comfortable with the AP system knowing it could see what I could.

It really sold me on the AP package when I test drove an AP1 car (not knowing it was an AP1 car mind you) and excitedly added the EAP option to my car. It sucked finding out that the car I wanted really was not the car I bought though.... I'm giving them a year from June 2017 to bring it up to AP1 speeds or I'm calling and demanding some sort of financial compensation for being sold something that isn't as advertised.

Don't wait a year... maybe 3 months, definitely 6 months.

 

[ChrML]

I agree with the general principles regarding delivering complex products, but I think that Elon has a well established history of making optimistic predictions, only to take longer. I also am not a fan of hero worship, but I will say this about Elon. He's really really really smart, more Albert Einstein than Henry Ford. I don't think any of us doubt that he is a force of nature even though he is also human ( at least for now .

Everytime I think Elon is seriously underestimating the complexity of an issue... I remember an interview he gave regarding a person coming in and trying to sell him a "superior" self-driving technology.... and his response at the time told me that he is, in fact, in touch with the complexities of these issues, as he seemed to have a healthy respect for how smart his group of engineers are and how hard they work.

Perhaps the most impressive thing about him is that his behavior tells me that he understands he is not going to be the person who invents a self-driving car, but he is confident that he is going to be one of the key people forcing that achievement into a reality.

We are all fortunate that some math geek who was a little too fascinated by Ender's Game and Babylon 5, made a massive amount of money on PayPal, otherwise, I have a feeling he'd be like most other tech billionaires, who did some impressive things but didn't actually manage to do something bigger.

That being said, Elon is going to be like most people in the world, he'll continue to try to do things for reasons he's mostly unaware, and he'll continue to be his own worst enemy (and best advocate). I, for one, am really glad that dude is doing what he's doing.

I completely agree. I love Elon despite of his shortcomings.

I just hope that even when Elon is smart, he still involves his time in every decision (assuming he is not going to develop it all himself )

 

[AnxietyRanger]

I agree with the general principles regarding delivering complex products, but I think that Elon has a well established history of making optimistic predictions, only to take longer. I also am not a fan of hero worship, but I will say this about Elon. He's really really really smart, more Albert Einstein than Henry Ford. I don't think any of us doubt that he is a force of nature even though he is also human ( at least for now .

Everytime I think Elon is seriously underestimating the complexity of an issue... I remember an interview he gave regarding a person coming in and trying to sell him a "superior" self-driving technology.... and his response at the time told me that he is, in fact, in touch with the complexities of these issues, as he seemed to have a healthy respect for how smart his group of engineers are and how hard they work.

Perhaps the most impressive thing about him is that his behavior tells me that he understands he is not going to be the person who invents a self-driving car, but he is confident that he is going to be one of the key people forcing that achievement into a reality.

We are all fortunate that some math geek who was a little too fascinated by Ender's Game and Babylon 5, made a massive amount of money on PayPal, otherwise, I have a feeling he'd be like most other tech billionaires, who did some impressive things but didn't actually manage to do something bigger.

That being said, Elon is going to be like most people in the world, he'll continue to try to do things for reasons he's mostly unaware, and he'll continue to be his own worst enemy (and best advocate). I, for one, am really glad that dude is doing what he's doing.

A wise post overall, thank you for it.

I would say one thing, though, unfortunately when it comes to Tesla, Elon also has a history of making predictions that didn't come true at all for the current product - or maybe not even next one - but did perhaps eventually come true. This has happened with Pxx models - repeatedly - and AP1, for example.

This means there is precedent that what he said would will come in AP2, might only appear in AP4. Let's hope that is not the case.

 

[buttershrimp]

A wise post overall, thank you for it.

I would say one thing, though, unfortunately when it comes to Tesla, Elon also has a history of making predictions that didn't come true at all for the current product - or maybe not even next one - but did perhaps eventually come true. This has happened with Pxx models - repeatedly - and AP1, for example.

This means there is precedent that what he said would will come in AP2, might only appear in AP4. Let's hope that is not the case.

Interesting, I'm curious what were some of the statements regarding Pxx models and AP1 you were referring to? They definitely went "all in" when they offered AP2 with future FSD, but I thought the most recent article where they say its "highly unlikely" that they will need to retrofit AP2.0 cars with AP2.5 hardware. I just wonder about that. The extra processor in 2.5 seems to be a bigger deal. I really wonder how the computing power is going to be upgraded. I sometimes think my car's inconsistencies in autopilot might be explained by sluggish processing.... especially if they are actively recording video, etc. I would think an upgrade will be likely on those FSD cars if needed.

 

[AnxietyRanger]

Interesting, I'm curious what were some of the statements regarding Pxx models and AP1 you were referring to? They definitely went "all in" when they offered AP2 with future FSD,

With Pxx I am referring to the entire sada from P85D HP, through P85DL and all the P90DL V1, V2, V3, chasing numbers (0-60, 1/4 mile...) that Elon presented early on - basically only P100D, in the end and perhaps P90DL V3 now that the software limiters have been lifted, started fulfilling what had been said... that wasn't much help to the owners of the early models.

As for AP1, the "sins" are less IMO, but still point to a pattern: ramp-to-ramp is not yet here, though I guess people still expect that might come. However, I don't think anyone expects traffic light detection or "meeting you at the curb" summon being enabled. The pattern being that Elon likes to make bold, outlandish statements and sell a product, but it might take the next version for this to come out...

I agree the AP2 claims are in their own stratosphere... Listen to Elon present AP2 to press on October 19th, 2016 and consider the street reality on September 19th, 2017...

 

[buttershrimp]

I see, you are saying his statements about the performance limits of these cars didn't end up coming to fruition. It's interesting, he really is his own worst enemy on this front and doesn't realize how casual comments regarding the best case scenario are taken as facts. Almost like some engineer trying to look good in front of the big boss, is saying, hey Elon, this version is capable of 0-60 of x... Elon tweets. Then the other engineers say, dammit who gave Elon the numbers from the unrestricted battery pack?

I hear you, but I suspect he is more aware of this now... at least his comments seem to reflect that... let's hope

 

[Pale_Rider]

He was keynote at UVWVA AI robotics conference last weekend and mentioned FSD in a Q&A toward the 46 minute mark of the video: Rick Astley - Never Gonna Give You Up

@calisnow would you mind posting the link?

Thanks!

 

[Sawyer8888]

Wait a minute.

The next big thing in AP 2.0?
(glass-half-full perspective / implies there was already a big positive thing)

Doesn't this make more sense?...

The first big thing in AP 2.0
(glass-half-empty perspective / reality)

 

[Kant.Ing]

Of course I'd like to see FSD first, since FSD will statistically cover Auto Wiper, TACC, AEB, and LOL!

 

[Electroman]

ramp-to-ramp is not yet here

What do you mean by "ramp-to-ramp is not yet here" ? Ramp-to-ramp IMO means, highway driving - end of the entry ramp to beginning of the exit ramp. And AP1 does it amazingly well. I use it every day.

 

[calisnow]

What do you mean by "ramp-to-ramp is not yet here" ? Ramp-to-ramp IMO means, highway driving - end of the entry ramp to beginning of the exit ramp. And AP1 does it amazingly well. I use it every day.

I believe the phrase typically means the car changes lanes in its own, finds its way back to the offramp automatically when the GPS wants it to exit etc - a state we still do not have.

 

[dark]

I believe the phrase typically means the car changes lanes in its own, finds its way back to the offramp automatically when the GPS wants it to exit etc - a state we still do not have.

I am losing my faith with this statement for AP2.0.

 

[Electroman]

I believe the phrase typically means the car changes lanes in its ow

How will the car know which lane is a better lane to be in? What you are describing is ramp to ramp autonomous driving. I am not sure if that is what Musk promised.

 

[RDoc]

How will the car know which lane is a better lane to be in? What you are describing is ramp to ramp autonomous driving. I am not sure if that is what Musk promised.

My understanding of ramp to ramp was that the car would change highways to get the the final off ramp. There's no chance of that happening, but I never believed it anyway.

What I did believe was that the car would park itself in my garage and move itself out to the curb on summon. There were even illustrations of that feature. That won't happen either.

 

[Electroman]

car would change highways

The connector between two highways - like in a clover leaf architecture - is a ramp by definition.

 

[daktari]

Hm, the next big thing seems to be to reduce the loss of accuracy. How else to interpret Karpathy?

Quote from twitter:
Andrej Karpathy @karpathy·8t
To get neural nets to work one must be super-OCD about details. With bugs nets will train (they "want" to work), but work silently worse.
Andrej Karpathy @karpathy·8t
The "move fast & fix stuff until it compiles then it's probably fine" approach is inadequate when each bug silently subtracts 2% accuracy.
Andrej Karpathy @karpathy
When my loss goes down it's not "cool, it's working!", it is "hmm it should be going down faster, something must be wrong". Okay, </rant>.

 

[AnxietyRanger]

What do you mean by "ramp-to-ramp is not yet here" ? Ramp-to-ramp IMO means, highway driving - end of the entry ramp to beginning of the exit ramp. And AP1 does it amazingly well. I use it every day.

What I meant was - and sorry if I was unclear - the AP1 feature promised for "8.1" that never materialized so far, i.e. automatic ramp-to-ramp.