No autonomy for the first M3s?

[TravelSD80]

Mobileye N.V. - The Road to Full Autonomous Driving: Mobileye and STMicroelectronics to Develop EyeQ®5 System-on-Chip, Targeting Sensor Fusion Central Computer for Autonomous Vehicles

This is an interesting read. Looks like autonomy hardware won't be ready (even engineering samples) before 2018. That clearly puts full autonomy out of the reach of the first batch of M3s. This reinforces my belief of leasing my first M3, then in early 2021 probably buy a M3 that's fully autonomous and keep it for a while. I think we are at such a technology inflection point the next 4-5 years that buying makes less sense if you are a geek. The timing of having a full SDK and hardware available in late 2018, would bode well for improved AP/autonomy in M3s in the 2020/2021 timeframe.

Note that I'm not discounting the fact that M3s should have at least the same AP features as current MSs, and possibly even more features.

 

[S'toon]

I've always considered that full autonomy for Model 3 as being fanciful wishful thinking.

 

[inspron]

I hope hope hope that there will be a retrofit option.

 

[TravelSD80]

I don't think it's fanciful at all if you look down the 2020-beyond road map. I do think if you expect it on the first batch of M3s, you are delusional.

 

[S'toon]

I don't think it's fanciful at all if you look down the 2020-beyond road map. I do think if you expect it on the first batch of M3s, you are delusional.

Probably a better way of stating it.

 

[flamingoezz]

does tesla use this company as a vendor?

isn't it possible they are further ahead or using a different method to attain the next level of autonomy?

im not concerned with having my car pick me up, but would love to be able to rest or at the very least, not have to constantly watch the road within a year or two of getting my m3. with tesla already capable of keeping lanes and sensing vehicles around it, that didn't seem an impossible transition for a 2-3 year span.

 

[Drivin]

Wait until:
The Second Reveal

It
Will
Be
Worth
It
!!!

 

[AngelArm1110]

I don't really care if it can drive itself or not, I'm buying it for me to drive

 

[Drivin]

Looks like autonomy hardware won't be ready (even engineering samples) before 2018.

This is when the SoC will be ready.
The associated hardware can be ready long before that and the SoC can be factory installed/swapped out later.

 

[igotzzoom]

Non-issue for me. As long as it will have adaptive cruise, lane keeping and AEB, that's enough for me. Full autonomous capability would be cool, but it's not a must-have for me.

 

[S'toon]

Right. Like the others above me said. I wouldn't want fully autonomous anyway.

 

[FloridaJohn]

I hope hope hope that there will be a retrofit option.

Highly unlikely. If all the hardware they would need is available today, you might see a software update to enable it. Most likely, though, there will need to be additional sensors, so you will be forced to buy a new car if you want the new features. On th Model S there is no retrofit option for Autopilot, for example.

 

[182RG]

I don't really care if it can drive itself or not, I'm buying it for me to drive

Precisely! It's like the autopilot in my airplane. Useful when I need it for high workload management situations. (IFR)

Otherwise, I prefer to hand fly it for the pure enjoyment.

 

[deonb]

does tesla use this company as a vendor?

Yes.

isn't it possible they are further ahead or using a different method to attain the next level of autonomy?

No.

 

[plankeye]

I don't really care if it can drive itself or not, I'm buying it for me to drive

Exactly! If I wanted full autonomy, I'd call an Uber.

 

[TravelSD80]

I want full autonomy so I DON'T have to pay for an Uber/Lyft. Parking in downtown San Diego is terrible, and costly. I'd love for my car to act as an uber to drop me off, go home and park, then summon my car when I'm ready. Ditto for going to the beach...parking is terrible. I don't want to pay per-ride for an Uber. In fact, I used Lyft last night to go to downtown San Diego because parking is so bad. The first Uber attempt resulted in an idiot driver that couldn't find my condo (a first). I then switched to Lyft 20 minutes later, and got to my destination 20 minutes late because of the idiot Uber driver. So I'd much rather have my (hopefully) more reliable Tesla M3.

In addition, not infrequently I have super early flights (think 6am). Uber/Lyft isn't always available at 4.30am, so if I rely on them for a ride I could get in a pickle and miss a flight should I run into more driver stupidity or unavailability. All in all, a "Uber" Tesla would be a far superior solution.

 

[JeffK]

does tesla use this company as a vendor?

haha you've been living under a rock my friend. Mobileye is THE company the majority of automakers use including Tesla.

The OP seems to forget that Tesla has a habit of using multiple EyeQ chips in it's vehicles. If I remember correctly the Model S uses a couple EyeQ3 chips. These chips are usually under $50 a piece.

The EyeQ4 chip is 10 times faster than the EyeQ3 and will be released in 2018 so it's extremely likely these will find their way into the Tesla Model 3.

You don't have to wait for the performance of the EyeQ5 of the future when you can spread the work over multiple EyeQ4 chips late next year.

 

[JohnSnowNW]

I don't think full autonomy comes without LiDAR, and I don't think LiDAR comes without a steep cost.

 

[JeffK]

I don't think full autonomy comes without LiDAR,

Supposedly Mobileye can get cm level accuracy using just optical cameras.

 

[Siciliano]

Mobileye N.V. - The Road to Full Autonomous Driving: Mobileye and STMicroelectronics to Develop EyeQ®5 System-on-Chip, Targeting Sensor Fusion Central Computer for Autonomous Vehicles

This is an interesting read. Looks like autonomy hardware won't be ready (even engineering samples) before 2018. That clearly puts full autonomy out of the reach of the first batch of M3s. This reinforces my belief of leasing my first M3, then in early 2021 probably buy a M3 that's fully autonomous and keep it for a while. I think we are at such a technology inflection point the next 4-5 years that buying makes less sense if you are a geek. The timing of having a full SDK and hardware available in late 2018, would bode well for improved AP/autonomy in M3s in the 2020/2021 timeframe.

Note that I'm not discounting the fact that M3s should have at least the same AP features as current MSs, and possibly even more features.

We are indeed fast approaching the very steep curve of exponential progression of many forms of tech (robotics/a.i./biomedical/etc). The average person is well unaware of this. Just do some reading up on Kurzweil and fill yourself in.