Anyone feel this is everything Lucid but not Lucid !!!

[eelpout]

I'm a bit confused about the Nuro partnership to be honest, as it appears they are using their own test bed of sensor equipment (via the roof). Not sure how much of what they are doing, if anything, will end up native to DD2P. Though one can hope.

 

[The_Don]

I'm a bit confused about the Nuro partnership to be honest, as it appears they are using their own test bed of sensor equipment (via the roof). Not sure how much of what they are doing, if anything, will end up native to DD2P. Though one can hope.

The answer is none. Lucid's DDP is completely separate from Nuro. Lucid is simply providing Nuro with vehicles to fit it's tech onto. Once the vehicles are in Nuro's hands, Lucid is completely out of the picture.

Lucid's partnership with Nvidia, on the other hand, will have a direct impact on DDP.

 

[HC_79]

I dunno, I wouldn't write off Lucid here. Sure, this is a flashy presentation by Rivian. But autonomous driving is hard, and even a tech powerhouse like Alphabet has had to spend billions of dollars and years of time to get to limited autonomous driving. As research in AI progresses, the hurdle to clear for building an autonomous driving system is getting a bit lower, but only a bit. Lucid is partnering with Nuro for autonomy. Nuro hasn't solved the autonomy problem either, but they are 9 years into a pretty deep multi-billion dollar investment in their attempt to do so. The partnership with Nuro could easily deliver viable autonomy to Lucid customers years before Rivian's effort pays off. Or not. Time will tell.

Lucid's smartest decision was partnering with Nvidia. Marc said it on the last earnings call, they didn't want to invest billions into something when they've got other things to be focusing on. The R2 is Rivian's M3 moment and the M3 almost sent Tesla bankrupt. Rivian is going to need a tonne of cash to launch R2 and now building its own ADAS chips and developing its own autonomous driving in house is going to require significant cash also. Yes, they have VW money now but they could be stretching it very thing across 2 big projects.

Time will tell if it was a smart move or not. What it shows though, is on the software / autonomy side Lucid does need to get more aggressive or they’ll be left in the dust without being able to get a foothold in the market.

 

[eelpout]

Lucid's smartest decision was partnering with Nvidia.

heh, JLR has partnered with Nvidia since 2022 (local to me they've been loaning them i-Paces and other vehicles to test with). So far they have bupkis to show for it. But JLR has other issues.

 

[Steve181]

I absolutely LOVED my G37x.

I had a 2010 G37 and loved it. I also liked the European Salon style. That being said, I also love the styling and features of my AIR PURE. There is little I would change except to equip it with an old fashioned key fob/system.

 

[socktucker]

A friend of mine is a professor with a doctorate in automated computer systems as in he literally wrote the book on ADAS system programming. He said full self driving as currently described by Tesla is nothing more than a pipe dream, not necessarily hindered by tech capabilities but by other constraints.

The example he uses is this:

In our city there is a huge dairy distributor. Been there for decades. All the trucks pull in and fill at this same intersection every day and has for decades. All day long.

The problem? Every turn they make is illegal because it is technically within 100 feet of an intersection. On paper, illegal. In real life for decades: no problem. Would be impossible to move.

So then, how does he program the car to deal with that? If the car follows directions to the fill station, it breaks the law. He cannot program a car to break the law. As such, there are hundreds of thousands of these types of intersections and situations throughout the United States, it’s impossible. Something like FSD supervised is the best that will ever exist. Plus there’s the old trolly on the track problem. Dos the car decide who lives or dies if the death of one party is inevitable? How do you program that? What if it needs to kill the driver in order to save a school bus? Do you program it to so that? They can program it either way but who makes that decision? Does the driver have to opt in? Who regulates that decision? What are the criteria?

It was a very interesting breakfast lecture.

 

[TahoeMichael]

The answer is none. Lucid's DDP is completely separate from Nuro. Lucid is simply providing Nuro with vehicles to fit it's tech onto. Once the vehicles are in Nuro's hands, Lucid is completely out of the picture.

Lucid's partnership with Nvidia, on the other hand, will have a direct impact on DDP.

I think it's extremely premature to say that. Lucid has established partnerships with Nuro and Nvidia, and has room to double down on either one as technology development proceeds. And of course they could always establish a third partnership in the future if they choose. The current state of autonomous driving R&D prevents anyone from making a certain prediction of what might be necessary to offer autonomous driving for privately owned vehicles.

I personally think that Lucid is extremely smart to lean on partnerships for autonomous driving. The technology is brutally hard to develop, extremely expensive to develop, and there are clearly going to be multiple companies hoping to license autonomous driving technology to vehicle manufacturers that don't own their own tech, thus providing competition and keeping the price of licensing down.

 

[TahoeMichael]

A friend of mine is a professor with a doctorate in automated computer systems as in he literally wrote the book on ADAS system programming. He said full self driving as currently described by Tesla is nothing more than a pipe dream, not necessarily hindered by tech capabilities but by other constraints.

The example he uses is this:

In our city there is a huge dairy distributor. Been there for decades. All the trucks pull in and fill at this same intersection every day and has for decades. All day long.

The problem? Every turn they make is illegal because it is technically within 100 feet of an intersection. On paper, illegal. In real life for decades: no problem. Would be impossible to move.

So then, how does he program the car to deal with that? If the car follows directions to the fill station, it breaks the law. He cannot program a car to break the law. As such, there are hundreds of thousands of these types of intersections and situations throughout the United States, it’s impossible. Something like FSD supervised is the best that will ever exist. Plus there’s the old trolly on the track problem. Dos the car decide who lives or dies if the death of one party is inevitable? How do you program that? What if it needs to kill the driver in order to save a school bus? Do you program it to so that? They can program it either way but who makes that decision? Does the driver have to opt in? Who regulates that decision? What are the criteria?

It was a very interesting breakfast lecture.

I don't exactly disagree with your professor friend. But his example illustrates the difference between academics and working engineers. The trolley problem is one of those thought exercises that academics engage in which don't actually have much resemblance to the actual problems autonomous vehicle engineers work on. The question itself is posed in terms that assume a deterministic driving environment, while real autonomous systems are built around a constantly evolving probability model, and so the vehicle controller is never actually faced with the clear cut decision of who it should kill.

 

[HC_79]

A friend of mine is a professor with a doctorate in automated computer systems as in he literally wrote the book on ADAS system programming. He said full self driving as currently described by Tesla is nothing more than a pipe dream, not necessarily hindered by tech capabilities but by other constraints.

The example he uses is this:

In our city there is a huge dairy distributor. Been there for decades. All the trucks pull in and fill at this same intersection every day and has for decades. All day long.

The problem? Every turn they make is illegal because it is technically within 100 feet of an intersection. On paper, illegal. In real life for decades: no problem. Would be impossible to move.

So then, how does he program the car to deal with that? If the car follows directions to the fill station, it breaks the law. He cannot program a car to break the law. As such, there are hundreds of thousands of these types of intersections and situations throughout the United States, it’s impossible. Something like FSD supervised is the best that will ever exist. Plus there’s the old trolly on the track problem. Dos the car decide who lives or dies if the death of one party is inevitable? How do you program that? What if it needs to kill the driver in order to save a school bus? Do you program it to so that? They can program it either way but who makes that decision? Does the driver have to opt in? Who regulates that decision? What are the criteria?

It was a very interesting breakfast lecture.

After doing multiple trips in a Waymo I’ve been surprised at not only how natural the driving feels but also how aggressive it can be when it needs to be. For example, it got stuck behind a delivery truck in SF, had its indicator on to move over and of course no one would let it in. It had no issues of making a quick move to get in from a complete stop. When it’s facing a red at traffic lights and the opposite light turns yellow, it’s starts to edge forward knowing a green is about to come. Someone opened their car door to get out, it didn’t stop, it swerved around them. Even when being picked up once it stopped in front of a fire hydrant which was surprised at as that isn’t legal.

Only criticism I have with Waymo is it can’t master pulling over properly yet. Of all things . I wouldn’t assume autonomous cars don’t have something in their algorithms to bend the rules a bit so it’s more human like.

 

[hmp10]

Rivian also promised Level 3 in 2020 then didn't deliver on that too.

I don't believe that's accurate.

With the Air introduction, Lucid announced that its sensor suite would allow the Air to reach Level 3 autonomy when the software became available, but Rawlinson was very careful to say the software was nowhere near that. In fact, he consistently painted one of the bleakest pictures of ADAS' near-term future of any of his peers at the same. In several interviews he said truly autonomous driving was at least 10 years away and that he didn't see anyone putting the money into software investment that it would take to get there in that time period. He even went so far as to speculate that Level 5 would never arrive and that even Level 4 might be a pipe dream.

This was in contrast to the years and years Tesla, in the person of Elon Musk, has claimed true FSD was just months away. The stories of people who bought Teslas and paid for full FSD, sometimes at prices as high as $15,000, without ever seeing it during their years-long ownerships are legion. And not only that, Tesla refused to keep FSD active for the subsequent buyer if the car was sold, though the original owner had paid for it without ever receiving it. Nor did Tesla allow an owner who paid for but never got the promised FSD capability carry the FSD over to another Tesla if they traded their cars for a newer model.

I'm no fan of Lucid software just at the moment, but I have never felt Lucid misrepresented the status or capability of DreamDrive.

 

[Gravitas]

Lucid's smartest decision was partnering with Nvidia. Marc said it on the last earnings call, they didn't want to invest billions into something when they've got other things to be focusing on. The R2 is Rivian's M3 moment and the M3 almost sent Tesla bankrupt. Rivian is going to need a tonne of cash to launch R2 and now building its own ADAS chips and developing its own autonomous driving in house is going to require significant cash also. Yes, they have VW money now but they could be stretching it very thing across 2 big projects.

Time will tell if it was a smart move or not. What it shows though, is on the software / autonomy side Lucid does need to get more aggressive or they’ll be left in the dust without being able to get a foothold in the market.

I am very suspicious of VW's partnership with Rivian.
I think VW is strategically positioning themselves for a hostile takeover of Rivian.
Even their Scout SUV and Truck are eerily similar to Rivian's.

 

[Leftlane]

The banging on Infiniti is crazy - we’ve had 5 of them, they’ve all been awesome to own, nothing ever breaks, and their tech has usually been way ahead of the competition (early on this was more accurate than now). My wife’s QX80 can put plenty of luxury SUVs to shame with its interior, and it drives fine for a big vehicle - of course it won’t feel like a sports car because of physics. Design is personal for everyone but from my first G35 to our latest QX I’ve liked them all.

 

[eelpout]

The banging on Infiniti is crazy -

my wife's G37 is a 2009 and it's been problem free.

I think some automatically bang on them because they are so tardy to EV's, which is a no-no in many circles

( welcome to the Infiniti thread )

 

[socktucker]

After doing multiple trips in a Waymo I’ve been surprised at not only how natural the driving feels but also how aggressive it can be when it needs to be. For example, it got stuck behind a delivery truck in SF, had its indicator on to move over and of course no one would let it in. It had no issues of making a quick move to get in from a complete stop. When it’s facing a red at traffic lights and the opposite light turns yellow, it’s starts to edge forward knowing a green is about to come. Someone opened their car door to get out, it didn’t stop, it swerved around them. Even when being picked up once it stopped in front of a fire hydrant which was surprised at as that isn’t legal.

Only criticism I have with Waymo is it can’t master pulling over properly yet. Of all things . I wouldn’t assume autonomous cars don’t have something in their algorithms to bend the rules a bit so it’s more human like.

I was impressed with Waymo too!

 

[Miami_Lucid_Man]

The banging on Infiniti is crazy - we’ve had 5 of them, they’ve all been awesome to own, nothing ever breaks, and their tech has usually been way ahead of the competition (early on this was more accurate than now). My wife’s QX80 can put plenty of luxury SUVs to shame with its interior, and it drives fine for a big vehicle - of course it won’t feel like a sports car because of physics. Design is personal for everyone but from my first G35 to our latest QX I’ve liked them all.

They were one of if not the first to roll out surround view cams on I believe the Q50. Then they trickled it down to Nissan and the rest is history. Their tech was way ahead but they stagnated for years and fell way behind.

 

[Ktw13]

When Lucid released good news people say they are gonna dilute, not true for this year.

Rivian needs money for R2, spend a few million to get the auto drive campaign going to pump up a few billion in stock is a good deal, why no fanboyz speculator mentioning this?

 

[HC_79]

Rivian needs money for R2, spend a few million to get the auto drive campaign going to pump up a few billion in stock is a good deal, why no fanboyz speculator mentioning this?

Rivian has about 2 years of cash before it runs out which is about the same as Lucid. They therefore have money to produce the R2 and guess they’re banking on it being successful to seek future investment which is exactly what Lucid will need to do.

 

[Daniel2022AT]

The Rivian product has a cool factor. But 2 mi/kwh and 400v architecture is unappealing to me. Does anyone here lurk on the Rivian owners forum like I do? The product problems seem more severe than Lucid’s, which are 95% software related. Rivian seems to have a lot of hardware problems too. Rivian owners discuss the importance of living near a Rivian service center. The grass is always greener on the neighbors side of the fence.

 

[Aka-none]

A friend of mine is a professor with a doctorate in automated computer systems as in he literally wrote the book on ADAS system programming. He said full self driving as currently described by Tesla is nothing more than a pipe dream, not necessarily hindered by tech capabilities but by other constraints.

The example he uses is this:

In our city there is a huge dairy distributor. Been there for decades. All the trucks pull in and fill at this same intersection every day and has for decades. All day long.

The problem? Every turn they make is illegal because it is technically within 100 feet of an intersection. On paper, illegal. In real life for decades: no problem. Would be impossible to move.

So then, how does he program the car to deal with that? If the car follows directions to the fill station, it breaks the law. He cannot program a car to break the law. As such, there are hundreds of thousands of these types of intersections and situations throughout the United States, it’s impossible. Something like FSD supervised is the best that will ever exist. Plus there’s the old trolly on the track problem. Dos the car decide who lives or dies if the death of one party is inevitable? How do you program that? What if it needs to kill the driver in order to save a school bus? Do you program it to so that? They can program it either way but who makes that decision? Does the driver have to opt in? Who regulates that decision? What are the criteria?

It was a very interesting breakfast lecture.

Not sure I agree that Tesla won’t program to break the law. Don’t they have a new mad max mode or is that a lie?

 

[gaskiller]

Not sure I agree that Tesla won’t program to break the law. Don’t they have a new mad max mode or is that a lie?

Saw a mad max mode video on Reddit where Tesla was overtaking at night at 78mph in a 65 mph zone…..that stuff should not be allowed