Thursday, September 21st 2017
Tesla Motors Develops Semi-custom AI Chip with AMD
Tesla Motors, which arguably brought electric vehicles to the luxury-mainstream, is investing big in self-driving cars. Despite its leader Elon Musk's fears and reservations on just how much one must allow artificial intelligence (AI) to develop, the company realized that a true self-driving car cannot be made without giving the car a degree of machine learning and AI, so it can learn its surroundings in real-time, and maneuver itself with some agility. To that extent, Tesla is designing its own AI processor. This SoC (system on chip) will be a semi-custom development, in collaboration with the reigning king of semi-custom chips, AMD.
AMD has with it a clear GPGPU performance advantage over NVIDIA, despite the latter's heavy investments in deep-learning. AMD is probably also banking on good pricing, greater freedom over the IP thanks to open standards, and a vast semi-custom track-record, having developed semi-custom chips with technology giants such as Sony and Microsoft. Musk confirmed that the first car in which you can simply get in, fall asleep, and wake up at your destination, will roll out within two years, hinting at a 2019 rollout. This would mean a bulk of the chip's development is done.
Source:
CNBC
AMD has with it a clear GPGPU performance advantage over NVIDIA, despite the latter's heavy investments in deep-learning. AMD is probably also banking on good pricing, greater freedom over the IP thanks to open standards, and a vast semi-custom track-record, having developed semi-custom chips with technology giants such as Sony and Microsoft. Musk confirmed that the first car in which you can simply get in, fall asleep, and wake up at your destination, will roll out within two years, hinting at a 2019 rollout. This would mean a bulk of the chip's development is done.
33 Comments on Tesla Motors Develops Semi-custom AI Chip with AMD
Did anyone even ask for this? Serious question.
I especially don't see how it'd kick off in much of America. Maybe in a few places (NYC) where people identify as commuters.. and definitely in Europe.. but driving yourself is about as American as apple pie. And I, personally, barely trust anyone to drive. Let alone a machine.
I also imagine a rule where you are not allowed to sleep alone in an autonomous car.
That said, I wouldn't trust it either unless the entire road system ran on it.
To be fair, I kind of have trust issues in general :P
And this goes beyond just tight situations in traffic. I've peeled out in a parking lot, as a robber pulled a gun right at my window... turned 180 Starskey and Hutch style and made a break for it. I probably even impressed him, because he didn't even fire shots as I drove away. An AI couldn't do any of this.
Not yet anyway.
Back to traffic though.. I'd say it's even more important when riding bikes. There's a lot more tight situations that need human vigilance. So I hope there aren't self-driving bikes to boot.
1) Tesla at least partially dropped their partnership with NVidia, which is kind of weird, because Nvidia still has a proud Drive PX2 on Tesla demo on their website.
2) AMD grew a pair to get into embedded and automotive market, which they did not want to do last year. Good for them. Of course there is market for self-driving cars and people who want it.
It's not going to start with 100% self-driving cars. So far both Tesla and Mercedes have demonstrated what they did with older tech, like NV Tegra X1, and it looks impressive. Fully autonomous highway driving, parking, etc.
Almost half of the US has already passed laws and regulations allowing autonomous vehicles w/ driver present. So, even hard-headed politicians see an appeal in reducing human error on the road.
Other countries are going as far as already commercializing self-driving car services (taxi, autonomous truck deliveries etc.).
Google did a very lengthy trial on their system , and here's what they got: That's 14 minor incidents on 23 cars in almost 8 years. Only one of those was a software glitch, while the others were caused by a human.
dont you think someone like elon thinks about this pretty much every day? he is one to say AI can become a threat to the human species, so then why would he do the opposite & be reckless with car AI?
AI cant really replace good drivers or solve tricky situations, but there are so many bad drivers that cause problems or cant get a handle on a situation
plus it's their own chips (AMD only helps) so now, they will only be able to blame themselves if something goes wrong (what would go wrong? right? )
Do machines need the thrill of speeding? Not likely and that accounts for roughly 66% of all accidents. The other 33% is human error.
Personally I like to drive too but if I am being assisted by AI and it helps me from making a mistake while driving, then why not?
I expect the same amount of error here as I do with spellcheck/autofill. I'm not sure it's going to be much better.
Reigning is what you were looking for.
Even drunken russians, finns, the entire eastern and central Europe combined can't beat that.