Separate names with a comma.
Discussion in 'Autonomous Vehicles' started by davidc18, Mar 13, 2017.
Intel buys self-driving tech firm Mobileye for $15.3 billion
More Intel lunacy. After the failure of "the network of things" they have been flailing and this is another "flail".
I don't know, I think a better example of "flailing" would be AP2. Talk about regression. I feel sorry for AP2 buyers who were sold one thing and delivered another.
I think Tesla's very public flailing with AP2 has put a bigger spotlight on the value of Mobileye and their proven tech which probably helped drive the price up to an astounding $15.3B.
Seems pretty smart to me. Intel clearly needs to respond to the threat from NVIDIA with respect to AI and self-driving tech. I must admit I'm slightly worried, Mobileye claims to be working with 27 car manufacturers, meanwhile Tesla is on its own and still hasn't caught up to where it was before with Mobileye tech on AP1. Could Mobileye become the standard and leave Tesla behind?
Intel to buy Mobileye for $15.3 billion
In business news this morning.
Well I guess they couldn't buy NVidia, so somebody else's cast off will have to do
The guy at Mobileye, anon shashua , mit phd, is second to none in that field.
Elon and his team are not in the same league.
I agree, never look a gift horse in the mouth and they did that with ME. I feel sorry for all the AP2 owners out there that dumped their 1.0 cars to get in before the free SpC'ing ended.
I just don't understand how Tesla can "cripple" all these new owners coming from 1.0 and not be expecting huge lawsuits. I think there's too much focus on the M3 and they dropped the ball big time on this.
Too bad the M3 is also going to use nVidia
Better make it 10 cameras now to equal ME's 1 upfront LOL
I am a hw2 owner and while I feel Tesla screwed up and misled buyers in Q4 2016, I feel that TeslaVision, at fw 188.8.131.52.3 is, other than speed, at or above AP1. Speed will come soon. Mobileye ain't *sugar* if Tesla can panic replace them with Nvidia completely new products in 8 months. They just should've told customers honest timely info in order to make informed decisions. I could've bought an inventory AP1 for a decent discount and had the safety features we are still waiting for from the get go. In six months I'm confident buying hw2 will be a no brainer. Even now hw2 AP can do some tasks better than hw1. When it starts stopping at stop signs and lights it won't be close between hw1 and hw2. Then we can see what mobileye is actually allowing competitor vehicles to do. I'm not seeing better level 2 systems in the market compared to hw1 and now hw2 is poised to tie and exceed it and somehow mobileye is coming out roses? They lost a major client. Intel is drunk.
With the backing of Intel, Mobileye are the hare in this race.
They better not lose to Tesla's tortoise.
I think it might be bad news for the diversity of the market. Other than Mobileye, Tesla had considered Nvidia and Intel (plus AMD to a lesser extent), but with Intel acquiring Mobileye, that leaves one less choice in the market. Shows the importance of Tesla developing software that is agnostic to the choice in processing chip.
And before people cry joy for Tesla's competitors, consider the fact that none of the other manufacturers using Mobileye has anything even approaching AP1. Mobileye has been very clear that they want their clients to go at the pace that they set and the other manufacturers have been obliging.
This is a very, very smart move for Intel...
Whilst we tend to think of Mobileye as "the guys that made AP1 possible"... the reality is that the Deep Learning breakthroughs that occurred around 2012 made the 10+ years of investment in their traditional, human-trained computer vision algorithms obsolete almost overnight. DL + GPU acceleration quickly exceeded the best human trained networks in a just a few months.
Of course, Mobileye know all of this stuff. They're incredibly smart people. Their CEO Amnon is super, super smart. They already have existing deals and relationships with almost every major car manufacturer in the world, plus trust with regulators. They have a unique knowledge of the space. Intel have lots and lots of money and resources, they're incredible chip guys, plus their push into 5G is significant for connected cars. It's a formidable pairing.
Meanwhile, whilst Tesla was busy injecting steroids into the 7+ year old technology that forms the base of AP1... Mobileye have been busy uilding out their own next generation stuff: deep sensing, crowd sourced mapping, road management, driving policy and custom silicon. I'm pretty sure that's why Musk said that, at some point, it would became 'inevitable' that they would part ways. Their roadmaps just... diverged
Nvidia, however, have definitely gone all-in on AI... and their progress is rapid - both in software and hardware. Hardware wise, the DrivePX2 supercomputer found in our HW2 Teslas is already out-of-date, with Xavier coming later this year - a smaller, faster, lighter and more power efficient package. Their DriveNet reference network is getting better every day.
Of course, Mobileye specialise in driving systems exclusively... have their own human supervised network trained specifically for recognising driving related features; they can augment with an additional layer of training via DNNs for even better accuracy. It's all they think about, all day every day. They also have a fundamentally different core business belief to Nvidia: Nvidia want to build the platform and the reference standard, but believe it's up to the car companies to implement their own individual networks on their platform. They want that competition to drive (heh) innovation in the space. Mobileye want to build the hardware and the software, in one time-tested package, then sell that into the cars.
I suppose it's Apple vs Microsoft all over again... and I suppose Nvidia is Microsoft in this case.
I hope it crashes less.
If Intel is quick, like real quick, we could see a BMW out-driving a Tesla before AP2 is ripe. ...in a head-to-head test. Doing all the safety things of AP1 that is lacking AP2.. out-driving AP2 on its current software.
That would take some fluff out of Elon's wings!
Windows Server ported to Qualcomm's ARM server chip. Repeat, Windows Server ported to ARM server chip
Intel might need automotive.
Microsoft demoed Windows Server for Qualcomm ARM. Internal right now, but if they're serious Intel faces a real prospect of losing a big share of data center markets. They've already lost alot of personal computing.
So it is your contention that because Tesla wasn't able to reach parity and or exceed Mobileye's decade+ of R&D in a matter of months, that this somehow made mobileye worth 15B? If anything Tesla coming close at all to parity with mobileye's system in the span of a few months probably freaked mobileye out and made them go looking for a deep pocketed bag holder to buy them out before Tesla system becomes clearly better than mobileye with less than a year of development vs mobileye 15+ years to reach mediocre TACC in production vehicles.
Maybe TM can start buying ME tech again.
Underscores the.error made with the divorce. Why do in house what Mobileye has spent years perfecting and will perfect further?
Because you own and control everything.
No need to beg for scraps.
Out of curiosity, does anyone know how far (if at all) the ME tech has progressed since the divorce?
If Tesla Vision produces a good EAP system on AP2 Tesla's tech will be worth 15 billion. It would be a brilliant financial move to divorce with Mobileye. Not that they would but they could license their tech to competing companies in the auto industry. If they can produce FSD and convince regulators to approve even level 4 FSD it will be worth tens of billions more.