The others are much further along than Tesla because they aren't making hardware decisions to sell profitable cars. Waymo and Cruise both use radars, lidars, and more cameras than Tesla.
I think about this regularly just don't have the time to pursue it:
Couldn't you build your arm in Nvidia Omniverse by also adding feedback like a cheap hig resolution distance or angle detector and train an ml model to compensate it?
Making and animating a 3D graphics robot arm is trivial compared to building it in real life. So not so much Omniverse, you would want to use a proper simulator like gazebo.
But beyond that, the kinematics as well as the force dynamics for controlling a serial manipulator are very well understood. So there aren't too many gains to be made by AI. It is difficult to implement in software due to some tricky situations about the nature of motion planning. Discontinuities around orientation approaches in 6-DOF systems for instance. But widespread use of serial manipulators is proof that, although challenging, they are relatively solved. It is always interesting to watch an AI model or genetic algorithm do some path planning, but this is a pretty well trod area of research at this point.
Now, when you want a robot to walk and pick things up at the same time... that is when AI becomes something to consider in order to figure out how the dynamics should work.
I'm not super familiar, but they say in the webpage that it is specifically for Universal Scene Description, which is formally for graphics. Although, after a quick google, it looks like they do have a simulation package which then runs on top of Omniverse (Isaac Sim?), so I guess that is Nvidia's robotics offering.
My general experience with other commercial offerings for simulation... is not great. In my experience, people usually end up migrating to gazebo, but I have been away from the field for a while now so it could be different. It is probably a situation where Nvidia will have a few coporate clients that they prioritize, and you are on your own to get it set up if you aren't on that lists. Pretty normal.
But motor and motion control isn't exactly so mysterious that we need AI for it. Inductance in electric motors can have some odd effects in the acceleration domain, but it generally boils down to a second to third order differential formula. Even when linking multiple together in a serial manipulator, the math is really well understood for modeling the motion output. Maybe there are some outputs to be gained implementing different drive trains in arbitrary circumstances, and monitoring how they fail and stuff like that. At that point you are really getting into the weeds of operations and maintenance more than actual motion control.
The situation that arises into a very complex n-dimensional problem that you would want AI to search through is the coordinated motion of multiple actuators to achieve a very complex output. Like, picking something up of unknown weight, running while carrying it up a steep hill, waving it around while doing all this. We take it for granted as humans with brains that can perform all this stuff trivially, but it is extremely complex motion.
Well, that doesn't really work. There is only so much electrical efficiency you can crank out of these motor. Essentially, there is a relationship between the current you pump through them, which is limited by their thermal characteristics, and the inductance of them. So you are trading off building more inductive motors that are more powerful but less reactive, and current draw which you can increase by putting more material in the motor and making it dispose heat much better but also bulkier. There are diminishing returns in many places in this process, and at a certain point you have to consider switching to hydraulics if you more force at a high reactivity, under essentially much less energy efficient conditions.
Maybe you could make a model that sizes motors correctly per application? But you are still much better hiring an engineer that knows what they are doing that can explain what is going on and trouble shoot things when they go wrong. At a certain point you are trying to figure out how to completely replace an engineer with a machine learning model, which I would like to think is a bad idea.
Can someone explain why people try to make new cars from the ground up, rather than "fork" a base platform from GM/Toyota/KIA/whoever with a different design and little features?
This happens in literally every other market and industry. hell, just look at PCs and laptops.
Like why can't a Canoo or Rivian license a base EV platform and come up with their own chassis design, infotainment system, and other little things to differentiate themselves?
Even doing that is going to run you at least $100 million and you still have many of the risks associated with developing your own platform without anywhere near the number of escape hatches due to significant lock in.
This has recently happened to a well known EV company relying on just the electrical platform from a well known and highly respected auto maker that they have a good relationship with. It means they’ve had to go looking at other platforms (again, just the electrical platform) and has put the whole company at risk.
Some companies will use a contract manufacturer to produce their cars for them. For example Fisker (about bankrupt) used Magna Stehr in austria to produce their EVs.
When you talk about a base platform I assume you are meaning an entire drive train? Most pieces of a car are going to be produced elsewhere and then assembled. Canoo isn’t making their own batteries or motors. Those come from battery and motor manufacturers. Canoo or a similar company might even buy structural battery pack built to spec by CATL.
Tesla's original was a fork of a Lotus. They got the whole chassis/etc from Lotus, dropped in their own drive train and other stuff. Other reply mentioned paccar.
Idk why it doesn't happen more often. I suspect most manufacturers aren't willing to license like that.
That wasn't really a fork, they didn't copy the design, they actually bought the chassis from lotus. It's not really that different from a manufacturer buying brakes or wheels from another supplier.
I suspect, starting from the ground up is part of why people are so interested in these companies.
The ICE car market is a massive market, which taking a part of required massive long term investment, and would result in you being relegated to just some minor side note.
Even Toyota one of the longest biggest players is still seen as kinda "new" compared to BMW and VW, let alone hyundai which is basically a massive state backed effort.
Now consider, all of that 100+ year history becoming completely irrelevant. An opportunity to time travel back 100 years and try to become the BMW of the future. I think this is what is exciting people.
when is the last time an entrenched trillion dollar industry with hardly any shakeups in decades suddenly had the ability for anyone to enter and say their biggest advantage is that they are new?
I am not saying I agree with it, I think knowing how to make a luxury car is a skill... But I don't think the goal of these CEOs and investors is to make a good car. Their goal is to make a 100+ year global legacy like BMW. If they could do that without making the car I am sure they would.
Why would anyone think that the current ICE manufacturers will just sit idly by and not get out EV designs at some point? Either on their own, or in collaboration with other legacy manufacturers? They have existing brand loyalty, existing relationships, supply chains, dealership networks etc. etc. that they can make use of even if their designs aren't as good as brand new entrants (and that's assuming brand new entrants can actually scale up quality manufacturing, which so far they haven't been able to).
Because for a long time the ICE manufacturers were making half-hearted play attempts at EVs, and Tesla came in and made real EVs and had a huge stock play related to that.
So now the ICEs are scrambling to catch up, and are doing so, but lots of money is flowing around hoping to be the next Tesla.
ICE is scrambling because ICE has 10 years left of new car sales in California (so the US basically) and Canada before laws prevent them from selling ICE. They aren't chasing Tesla, they're trying to prepare for the outlawing of their 100 year old products.
> Because for a long time the ICE manufacturers were making half-hearted play attempts at EVs, and Tesla came in and made real EVs and had a huge stock play related to that.
And their sales were still dwarfed by legacy manufacturers, who then started paying attention and practically all of them have at least one EV model out there, often with new platforms/models being in the works.
I suspect most of these EV startups are trying for that TSLA bump, and don't really care much that TSLA is almost inevitably doomed to stagnant stock returns for decades now.
Meanwhile the legacy companies will keep moving slowly along and produce vehicles. Some will die, some will merge, some will have wild successes.
But it's much different than Google et al because cars actually cost about as much to make and sell as they do to buy.
I'm not even sure an EV car is the answer for "Green". I remember a Joe Rogan ep where he and his guest talk about the giant holes and bad mining condition in Africa because of the current EV rash to get materials for car battery.
EV cars seem to be the go-between for ICE and the real answer for Green.
You need to separate the concept of EVs from the current state of battery technology. EVs are almost certainly the future, and as our battery and charging technology keeps improving the whole package will become greener and greener.
I'm not aware of any pure platform available off the shelf, though I imagine a big enough checkbook and a conversation with Magna might get you pretty far, but you can buy from Ford the Mustang Mach-E motor https://performanceparts.ford.com/part/M-9000-MACHE
There are lots, but most are not road legal. All privateer racers in many race series use cars that are manufactured by established companies, then heavily modify them.
Alpine makes forks of BMWs.
Braubas makes forks of Mercedes.
RUF makes forks of Porsche.
The other example is when 2+ large manufacturers work together to cost share a platform.
Examples:
- BMW Z4 / Toyota Supra
- Subaru BRZ / Toyota GT86
- Mitsubishi 3000GT / Dodge Stealth
Several Aston Martin coupes were built on Jaguar platforms.
The NUMMI car factory in Fremont, CA was jointly owned by GM and Toyota before Tesla bought it. For a while, it made the Toyota Voltz / Pontiac Vibe.
I think people get the impression from the wording that the “platform available off the shelf” means the consuming manufacturer can just grab some units and modify them, but in practice, there is a massive contractual negotiation. And since car manufacturers are heavily regulated, it makes sense to cost share both the manufacturing and the standards / crash testing.
How enthusiastic do you think those companies would be if someone came to them wanting to license their design perpetually and use it to compete with them? And even if that is something they're willing to do (which they might be), the terms would then make it difficult to ever do something that isn't based on that platform, since the manufacturer whose design is licensed would likely object that the new design was still based in some way on the old one.
The analogy in the PC world would not be "license a design from an ODM like Compal or Quanta", it'd be "license a design from Dell or Lenovo".
That in itself suggests a potential market: perhaps there's room for an ODM in the automotive space?
there isn't a single electric vehicle on the market that looks anything like the international scout, FJ40, or any classic rugged vehicle. Look at how much icon4x4 restorations go for. There are wealthy people out there with great taste, but no big auto manufacturer seems to want to produce a chassis with this aesthetic.
So why wouldn't Ford or GM be cool with letting a smaller label do this and testing the market?
It's a niche market. I love Icon's stuff but they sell very low numbers of vehicles. I would love to have an ev FJ40 or Scout, but I'm skeptical the general populace wants these as much as enthusiasts think they do.
I find this take to be so outrageously wrong. Do you think the public "wants" the new hummer? Do you think they want any of the garrish looking cars that are literally all over the roads? People will literally buy anything and everything that's put out there as the numbers have proven time and time again. This is what brands and marketing does. It shapes public opinion and public perception. There are countless generic, soulless, lifeless car bodies being sold every day. and icon can't put out big numbers because there's numericallyonly so many of those vintage cars left to restore, and they are a small operation.
Using an ICE base platform for EVs doesn't make sense, and there is no EV base platform for sale/license. Companies building the EVs are holding their cards tightly to their chests for competitive advantage.
> Using an ICE base platform for EVs doesn't make sense,
Kind of. Renault's small sized EVs share parts of the platform as some ICE models (e.g. Clio/Renault 5 E-Tech) which allows them to reduce costs of the EV model. Considering their target audience and market segment (small city-focused EV with city-scale range) the tradeoff is perfectly acceptable.
> and there is no EV base platform for sale/license
How do you know this? Renault for instance are in a pretty specific spot, market wise, and they were trying to merge with Fiat-Chrysler before the latter ended up with PSA (Peugeot, Citroen, Opel) in Stellantis. They have some EV tech, especially on the cheap city-focused side (Dacia Spring, Renault Zoe, Renault 5 E-tech), and would probably be open to collaborating / sharing risks. This is purely conjecture of course.
There are EV base platforms for sale and license. For example Nio licensees it's platform. 60% of EVs are made/sold in China so that's where you should look for a platform.
It’s a classic make vs buy decision. What base platform is offered by GM/Toyota/Kia/etc?
There are some examples where car makers (both EV and ICE) will sell components to competitors but in general this is pretty rare. Most companies want to control key aspects of their supply chain.
Rivian did use motors from Bosch but is now in the process of using their own internally designed motors. This allows them to build exactly what they need and avoid relying on Bosch for a key component of their vehicles.
It does happen, for example Paccar is a truck designer and sells licenses (apparently they also manufacture trucks but I think they used to only design them, although I'm not sure). I think some european makers do that as well and I would be surprised if Ford or GM doesn't. At least GM planned to buy (not sell) designs to Nikola if they ever managed to finish one (which I think they didn't in time for the deal to go through).
Paccar builds and sells trucks, not a license to build them. Other companies buy those add their own equipment (cranes, concrete mixers, etc) and resell them.
Light duty truck and van builders e.g. Ford and GM do the same thing but mostly for companies building RVs rather than commercial equipment.
> Like why can't a Canoo or Rivian license a base EV platform and come up with their own chassis design, infotainment system, and other little things to differentiate themselves?
I imagine anyone willing to license one at the moment only has a crappy one to sell you or it is some Chinese EV, which creates political problems.
>
Can someone explain why people try to make new cars from the ground up, rather than "fork" a base platform from GM/Toyota/KIA/whoever with a different design and little features?
That's sort of what Tesla did with the Roadster. It was a Lotus that they put an electric motor into.
Do any of the car manufactorers want to sell you their base model? Even if they did, how would Ford dealers feel about Ford selling rivian the rights to make better Ford cars?
Big companies don't use cloud just for fun.
And yes it's expensive but a ton of people working on that are good but not security experts or ops experts.
Managed shit takes the complexity out of this.
For everything non corporate yeah go with SQL lite if you like but you should have enough money anyway that those 'optimizations' don't matter.
One expert is not cheap and just coming up with that basic blog post requires a little tof expertise too.