24 ms·

Tesla's Autopilot Chief Keller Steps Down After Two Years

126 points by trimbo 8y ago

loser777 8y ago
Not mentioned in the bloomberg article (at the time of reading): Keller is apparently moving to _Intel_ (according to a quick search). Wow, I really hope that doesn't worsen the current imbalance of the x86 and GPU markets. Given that Keller seemed to be to Ryzen at AMD what Raja was to Vega/Polaris, it doesn't look good.
- philjohn 8y ago
  Eh, Raja wasn't all that great - he promised the earth and delivered nowhere near.
  - boojums 8y ago
    Isn't Navi really the first architecture Raja would have have much impact on? He rejoined AMD in 2013 [0] and designs for architectures take several years to go from design to consumers. Jim Keller joined AMD a year earlier and we didn't see stuff he worked on until Zen. [0] https://www.anandtech.com/show/6907/the-king-is-back-raja-koduri-leaves-apple-returns-to-amd https://www.anandtech.com/show/6907/the-king-is-back-raja-ko...
- slivym 8y ago
  That's really hilarious, because Intel's entire automotive division is run by the guy who came from the Mobile Eye acquisition. For those that don't remember - Mobile Eye was one of Tesla's tech suppliers until they split up. Then Mobile Eye basically came out and said that Tesla's self-driving technology was unsafe. It was a massive PR hit for Tesla, and a massive PR hit for Mobile Eye (who wants a technology partner that'll publicly accuse you like that). It'll be interesting to see Keller work in that environment.
  - dogma1138 8y ago
    Tesla tried to throw MobilEye under the bus and MonileEye basically said that they’ve used their platform out of spec and out of its functional domain, there were also some cheeky remarks by Amnon who was making statements like “if you have big balls like Tesla you can only use a single sensor type with no redundancy”. Currently MobilEye seems to be the only vendor who doesn’t oversells their capabilities like crazy and has a solid evolutionary roadmap on all 3 landscapes - technology, regulation, and business model.
    - sah2ed 8y ago
      Any direct links to learn more about Mobile Eye's roadmap for the curious?
      - dogma1138 8y ago
        Their CES talks are usually quite good https://www.youtube.com/watch?v=GQ15HWCw_Ic https://www.youtube.com/watch?v=GQ15HWCw_Ic This one about their safety model is also quite interesting: https://www.youtube.com/watch?v=L92hR5CSw1E https://www.youtube.com/watch?v=L92hR5CSw1E
    - Tepix 8y ago
      Do you also consider Waymo to be a vendor?
      - dogma1138 8y ago
        If they won't be on some level it's an extreme waste of effort.
      - jacksmith21006 8y ago
        Of course. Why else has Google spent billion on the tech?
  - PinguTS 8y ago
    MobilEye was long time technology provider of the same technology for BMW and later Volvo. MobilEye has basically developed their technology with those companies. Tesla was than later coming into the game as customer of MobilEye. The difference is, that Elon enabled much and sold it differently than the very same technology was used and still is used at BMW and Volvo. Then with first big accident (the death in Florida) Elon wanted to accuse MobilEye. But actually it was Elon who sold that technology in away it was still not ready for. MobilEye was very aware of that. But taking all the blame was the reason to split up. Btw. BMW and Volvo both have radar on board, which the first Tesla lacked of, because Elon doesn't wanted to spent that money.
- opencl 8y ago
  Not just Zen. He headed the development of the Athlon, Athlon64, and also Apple's early custom SoC efforts.
- observr 8y ago
  > Keller is apparently moving to _Intel_ (according to a quick search). I did find it odd when he took a hybrid hardware/software role at Tesla. Probably was mainly working on clockless low power neural net processors. Considering what he did at AMD with meagre resources, I can't wait to see what impact he makes at Intel with an unlimited budget.
tim333 8y ago
I wonder how Andrej Karpathy is getting along. He was in " Tesla hires deep learning expert Andrej Karpathy to lead Autopilot vision" headlines not long back. https://news.ycombinator.com/item?id=14599668 https://news.ycombinator.com/item?id=14599668
- jpdus 8y ago
  Did you actually read the article? >Andrej Karpathy, Tesla’s director of AI and Autopilot Vision, will have responsibility for Autopilot software.
nakedrobot2 8y ago
I think the brunt of the "Tesla Autopilot Deaths" problem is not the tech, but the marketing department (which has to include Elon Musk of course). This "kind of autonomous driving" is absolutely deadly, by definition. It's never going to work well. "The car will drive by itself, until it can't. So don't pay any attention, until you need to save your own life. We'll give you 2 seconds warning". I mean, really? I don't know why Tesla could possibly think this is a good idea. By extension, anyone who actually uses or trusts autopilot is basically forfeiting their life. I think it is cognitively harder to use autopilot and pay attention, than to simply drive safely. As far as the tech goes, I am in agreement with Elon. I think LiDAR is a short term advantage, but the companies/stacks depending on LiDAR will reach a local maximum from which they won't recover without totally starting again from scratch. LiDAR is short range, low resolution, and extremely expensive. There is no reason that visible light (+ IR) CMOS cameras can't do better by an order of magnitude, at a lower cost.
- netsharc 8y ago
  I fully agree. I wish the "autopilot" would just be used passively, e.g. auto-braking when it spots something bad ahead, or some sort of warning if it notices the driver deviating from the lane (and maybe only active avoidance if that deviation means avoiding crashing into something in the other lane).
  - radiorental 8y ago
    Essentially Volvo's approach: "Here is an extraordinary statistic: since the Volvo XC90 went on sale in the UK in 2002 it has sold over 50,000 vehicles, yet not a single person has been killed while driving it, or as a passenger." http://www.bbc.com/news/business-43752226 http://www.bbc.com/news/business-43752226
    - lmm 8y ago
      That's a curious turn of phrase, and immediately makes me ask: how many other people have been killed by Volvo XC90s? The XC90 is a big, tall, heavy car that makes me nervous as a cyclist - does it protect its occupants at the expense of other road users?
code4tee 8y ago
The “autopilot” concept isn’t going to work at scale with cars. It’s either autonomous or it’s not. As these accidents demonstrate drivers can’t be relied upon to ‘monitor’ the autopilot and react if something goes wrong. It’s a deeply flawed design concept from an otherwise very innovative company.
- crististm 8y ago
  You touched a nerve on some people. And rightly so. Perception and context awareness is not instantaneous. Here's a quick improvement on testing for the learning agent: watch the driver while he is driving, watch the road, AI predicts what do in the same situation, diff that with the driver; diff it offline if necessary. Bang! massive improvement in learning data, time and pinpointing blunders. If they are not already doing this, they should.
  - martin_bech 8y ago
    Tesla is doing this. Comparing driver vs autopilot, all information from the fleet, is sent back to Tesla and used as learning data. Thats one of the reasons why all the cars have free internet access.
    - Someone1234 8y ago
      So Tesla is using its drivers as guinea pigs in a high steaks game that could result in death? A lot of people want to treat Tesla like they're just a tech company, while ignoring that the "A/B test" conducted here could literally get people killed (and has). For example, in the most recent death some owners are claiming that an over-the-air software update caused the issue (lane centering), it made their cars drive differently. In tech regular updates are common, but normally the biggest consequence for getting one wrong is angry customers, not dead customers.
      - JoeAltmaier 8y ago
        This is done everywhere. Testing tires or medicines or road surface treatments. We all risk our lives every day to some degree. Testing how best to preserve them with A/B testing is a long respected approach. And is it fair to put the blame solely on a cherrypicked action by Tesla engineering? A tragic coincidence is a compilation of events.
deng 8y ago
Title is misleading. Keller was at Tesla for two years, but was head of "Autopilot" for less than a year (he took that part over from Lattner, who was head of software for less than 6 months).
- andygates 8y ago
  That seems like an unfortunate amount of churn at the top of a long-term development department. Trouble in Teslaville, or is this just regular valley bed-hopping?
  - davedx 8y ago
    > Trouble in Teslaville, or is this just regular valley bed-hopping? A bit of both I'd say. From what I've been following, Tesla's autopilot efforts have been a particularly rocky road, comparable with the Model 3 ramp-up troubles.
  - diiaann 8y ago
    Lattner was fired.
JudasGoat 8y ago
Here is someone that should be paid like a sports superstar. There seems to be so few people with his skill set.
- btian 8y ago
  you're free to pay him $50 million a year. I don't believe there are laws that would prohibit it
jacksmith21006 8y ago
Been a software engineer majority of my life and if I worked on technology that could kill people and my company was not approaching how I thought made sense it would be difficult for me to stay. I do not know if that is what is happening here but might be. IMO, you can NOT be kind of driving. SDC technology should NOT go directly to consumers at this point. Should be done with a commercial offering as we work through testing. SDC - Self Driving Cars
ggg9990 8y ago
What they really need to do is have the decision on when to ship be put in the hands of someone other than the person who is managing the development of the technology.
Proven 8y ago
Rats jumping ship
loceng 8y ago
I would love to see the chain of command, along with specifically who, for decision making relating to the investigation of and up until the press releases relating to that last crash where someone died (where the roadway safety buffer structure wasn't replaced properly), where based on information and analysis in HN comments, it seemed like Tesla/someone at Tesla was disingenuous or had a thinking error in information they translated up the chain of command. The PR definitely felt like the marketing department - who we can assume doesn't have the same technical understanding or abilities - had sway in perhaps a situation they shouldn't have. The only way for Elon to survive the hole that's getting dug deeper is to own up to analyze their process, own up the mistake, and be very public/open about the process and improvements to their processes they will make - and highlighting if it was brought to light by information/analyze from the tech community; the liability also shouldn't be any higher than 50% for Tesla IMHO, because if that safety infrastructure was properly replaced then there's a very high probability the individual would have lived, and they would become a spokesperson crying foul.

126 points by trimbo 8y ago

Not mentioned in the bloomberg article (at the time of reading): Keller is apparently moving to _Intel_ (according to a quick search). Wow, I really hope that doesn't worsen the current imbalance of the x86 and GPU markets. Given that Keller seemed to be to Ryzen at AMD what Raja was to Vega/Polaris, it doesn't look good.

philjohn 8y ago
Eh, Raja wasn't all that great - he promised the earth and delivered nowhere near.
- boojums 8y ago
  Isn't Navi really the first architecture Raja would have have much impact on? He rejoined AMD in 2013 [0] and designs for architectures take several years to go from design to consumers. Jim Keller joined AMD a year earlier and we didn't see stuff he worked on until Zen. [0] https://www.anandtech.com/show/6907/the-king-is-back-raja-koduri-leaves-apple-returns-to-amd https://www.anandtech.com/show/6907/the-king-is-back-raja-ko...
slivym 8y ago
That's really hilarious, because Intel's entire automotive division is run by the guy who came from the Mobile Eye acquisition. For those that don't remember - Mobile Eye was one of Tesla's tech suppliers until they split up. Then Mobile Eye basically came out and said that Tesla's self-driving technology was unsafe. It was a massive PR hit for Tesla, and a massive PR hit for Mobile Eye (who wants a technology partner that'll publicly accuse you like that). It'll be interesting to see Keller work in that environment.
- dogma1138 8y ago
  Tesla tried to throw MobilEye under the bus and MonileEye basically said that they’ve used their platform out of spec and out of its functional domain, there were also some cheeky remarks by Amnon who was making statements like “if you have big balls like Tesla you can only use a single sensor type with no redundancy”. Currently MobilEye seems to be the only vendor who doesn’t oversells their capabilities like crazy and has a solid evolutionary roadmap on all 3 landscapes - technology, regulation, and business model.
  - sah2ed 8y ago
    Any direct links to learn more about Mobile Eye's roadmap for the curious?
    - dogma1138 8y ago
      Their CES talks are usually quite good https://www.youtube.com/watch?v=GQ15HWCw_Ic https://www.youtube.com/watch?v=GQ15HWCw_Ic This one about their safety model is also quite interesting: https://www.youtube.com/watch?v=L92hR5CSw1E https://www.youtube.com/watch?v=L92hR5CSw1E
  - Tepix 8y ago
    Do you also consider Waymo to be a vendor?
    - dogma1138 8y ago
      If they won't be on some level it's an extreme waste of effort.
    - jacksmith21006 8y ago
      Of course. Why else has Google spent billion on the tech?
- PinguTS 8y ago
  MobilEye was long time technology provider of the same technology for BMW and later Volvo. MobilEye has basically developed their technology with those companies. Tesla was than later coming into the game as customer of MobilEye. The difference is, that Elon enabled much and sold it differently than the very same technology was used and still is used at BMW and Volvo. Then with first big accident (the death in Florida) Elon wanted to accuse MobilEye. But actually it was Elon who sold that technology in away it was still not ready for. MobilEye was very aware of that. But taking all the blame was the reason to split up. Btw. BMW and Volvo both have radar on board, which the first Tesla lacked of, because Elon doesn't wanted to spent that money.
opencl 8y ago
Not just Zen. He headed the development of the Athlon, Athlon64, and also Apple's early custom SoC efforts.
observr 8y ago
> Keller is apparently moving to _Intel_ (according to a quick search). I did find it odd when he took a hybrid hardware/software role at Tesla. Probably was mainly working on clockless low power neural net processors. Considering what he did at AMD with meagre resources, I can't wait to see what impact he makes at Intel with an unlimited budget.

tim333 8y ago

I wonder how Andrej Karpathy is getting along. He was in " Tesla hires deep learning expert Andrej Karpathy to lead Autopilot vision" headlines not long back. https://news.ycombinator.com/item?id=14599668 https://news.ycombinator.com/item?id=14599668

jpdus 8y ago
Did you actually read the article? >Andrej Karpathy, Tesla’s director of AI and Autopilot Vision, will have responsibility for Autopilot software.

nakedrobot2 8y ago

I think the brunt of the "Tesla Autopilot Deaths" problem is not the tech, but the marketing department (which has to include Elon Musk of course). This "kind of autonomous driving" is absolutely deadly, by definition. It's never going to work well. "The car will drive by itself, until it can't. So don't pay any attention, until you need to save your own life. We'll give you 2 seconds warning". I mean, really? I don't know why Tesla could possibly think this is a good idea. By extension, anyone who actually uses or trusts autopilot is basically forfeiting their life. I think it is cognitively harder to use autopilot and pay attention, than to simply drive safely. As far as the tech goes, I am in agreement with Elon. I think LiDAR is a short term advantage, but the companies/stacks depending on LiDAR will reach a local maximum from which they won't recover without totally starting again from scratch. LiDAR is short range, low resolution, and extremely expensive. There is no reason that visible light (+ IR) CMOS cameras can't do better by an order of magnitude, at a lower cost.

netsharc 8y ago
I fully agree. I wish the "autopilot" would just be used passively, e.g. auto-braking when it spots something bad ahead, or some sort of warning if it notices the driver deviating from the lane (and maybe only active avoidance if that deviation means avoiding crashing into something in the other lane).
- radiorental 8y ago
  Essentially Volvo's approach: "Here is an extraordinary statistic: since the Volvo XC90 went on sale in the UK in 2002 it has sold over 50,000 vehicles, yet not a single person has been killed while driving it, or as a passenger." http://www.bbc.com/news/business-43752226 http://www.bbc.com/news/business-43752226
  - lmm 8y ago
    That's a curious turn of phrase, and immediately makes me ask: how many other people have been killed by Volvo XC90s? The XC90 is a big, tall, heavy car that makes me nervous as a cyclist - does it protect its occupants at the expense of other road users?

code4tee 8y ago

The “autopilot” concept isn’t going to work at scale with cars. It’s either autonomous or it’s not. As these accidents demonstrate drivers can’t be relied upon to ‘monitor’ the autopilot and react if something goes wrong. It’s a deeply flawed design concept from an otherwise very innovative company.

crististm 8y ago
You touched a nerve on some people. And rightly so. Perception and context awareness is not instantaneous. Here's a quick improvement on testing for the learning agent: watch the driver while he is driving, watch the road, AI predicts what do in the same situation, diff that with the driver; diff it offline if necessary. Bang! massive improvement in learning data, time and pinpointing blunders. If they are not already doing this, they should.
- martin_bech 8y ago
  Tesla is doing this. Comparing driver vs autopilot, all information from the fleet, is sent back to Tesla and used as learning data. Thats one of the reasons why all the cars have free internet access.
  - Someone1234 8y ago
    So Tesla is using its drivers as guinea pigs in a high steaks game that could result in death? A lot of people want to treat Tesla like they're just a tech company, while ignoring that the "A/B test" conducted here could literally get people killed (and has). For example, in the most recent death some owners are claiming that an over-the-air software update caused the issue (lane centering), it made their cars drive differently. In tech regular updates are common, but normally the biggest consequence for getting one wrong is angry customers, not dead customers.
    - JoeAltmaier 8y ago
      This is done everywhere. Testing tires or medicines or road surface treatments. We all risk our lives every day to some degree. Testing how best to preserve them with A/B testing is a long respected approach. And is it fair to put the blame solely on a cherrypicked action by Tesla engineering? A tragic coincidence is a compilation of events.

deng 8y ago

Title is misleading. Keller was at Tesla for two years, but was head of "Autopilot" for less than a year (he took that part over from Lattner, who was head of software for less than 6 months).

andygates 8y ago
That seems like an unfortunate amount of churn at the top of a long-term development department. Trouble in Teslaville, or is this just regular valley bed-hopping?
- davedx 8y ago
  > Trouble in Teslaville, or is this just regular valley bed-hopping? A bit of both I'd say. From what I've been following, Tesla's autopilot efforts have been a particularly rocky road, comparable with the Model 3 ramp-up troubles.
- diiaann 8y ago
  Lattner was fired.

JudasGoat 8y ago

Here is someone that should be paid like a sports superstar. There seems to be so few people with his skill set.

btian 8y ago
you're free to pay him $50 million a year. I don't believe there are laws that would prohibit it

jacksmith21006 8y ago

Been a software engineer majority of my life and if I worked on technology that could kill people and my company was not approaching how I thought made sense it would be difficult for me to stay. I do not know if that is what is happening here but might be. IMO, you can NOT be kind of driving. SDC technology should NOT go directly to consumers at this point. Should be done with a commercial offering as we work through testing. SDC - Self Driving Cars

ggg9990 8y ago

What they really need to do is have the decision on when to ship be put in the hands of someone other than the person who is managing the development of the technology.

Proven 8y ago

Rats jumping ship

loceng 8y ago

I would love to see the chain of command, along with specifically who, for decision making relating to the investigation of and up until the press releases relating to that last crash where someone died (where the roadway safety buffer structure wasn't replaced properly), where based on information and analysis in HN comments, it seemed like Tesla/someone at Tesla was disingenuous or had a thinking error in information they translated up the chain of command. The PR definitely felt like the marketing department - who we can assume doesn't have the same technical understanding or abilities - had sway in perhaps a situation they shouldn't have. The only way for Elon to survive the hole that's getting dug deeper is to own up to analyze their process, own up the mistake, and be very public/open about the process and improvements to their processes they will make - and highlighting if it was brought to light by information/analyze from the tech community; the liability also shouldn't be any higher than 50% for Tesla IMHO, because if that safety infrastructure was properly replaced then there's a very high probability the individual would have lived, and they would become a spokesperson crying foul.