But the software solution is inextricably tied to the physical hardware. If BEV batteries become commodity products, then the software that optimizes those BEV batteries will become commodity software packaged to match each BEV battery. If I’m GM buying CATL batteries, I’ll be working with the software that CATL has designed to make their batteries work. Like air bags - the air bags in your car have electronics that are run by software, and the software is necessary, but not really the main driver of value for the airbag. My ICE engine has software - it’s got an ECU that runs all the ignition timing and whatnot - but all that software has to do is be properly matched to that engine. If another car company bought that engine “off the rack” and put it in their car, they’re not going to be re-writing the software package for that engine to do anything any differently than what it’s currently written to do.
I have bought a washing machine recently - and no, they don’t. Even though it’s “just a washing machine,” there’s a wide variety of different hardware choices. Top loading vs. front loading, a wide variety of different sizes and capacities, different impeller options, different power motors, different materials for tubs (steel vs. plastic, load-balancing vs. not), single-bin vs. dual-bin options, different physical containers for detergent loading…and that’s before you get into all the different design and finish and cabinet options, or combo washer/dryers that do both functions in the same unit.
Sure, there’s software involved - if you install a steamer in the washing machine, you need to program the display to have a “steamer” button and have software that tells the steamer what to do. But no one’s picking between an LG or GE washing machine based on how good the steamer software is, because all they need is for the software to work with that steamer hardware.
Something about this argument has long bugged me but it finally clicked today. What is it about software that leads you to think it cannot be replicated (or duplicated) by a competitor? In other words, how is code, especially in an age where the cost of doing business in China is information sharing, such a proprietary advantage as it pertains to how it controls hardware?
I am reminded about a recent announcement that Benz will be bringing the first lvl 3 autonomous vehicle to market this year - leap-frogging Tesla in the process. Honda already has lvl 3 but not in the US. Tesla of course is still lvl 2.
I would think getting from 40mph to 60mph would be an easier lift than from “can’t recognize a fire truck” to “doesn’t drive into a fire truck” would be.
Depends how smart the elevator mfr is. We had all six elevators in three high rises replaced in Summer 2019. They removed one elevator from each building (leaving one per building still operational) and replaced the first ones three months later. They then removed the second elevator from each building and replaced with new ones a month or two before 2020 began. Problems were significant for a while during the first months of 2020. Then the REAL fun began.
In one building, a water leak on a middle floor flooded the elevator shafts with water–and, OF COURSE, all the electronics on both elevators got zapped. Cannibalized one elevator from another building to get one elevator working. Then Covid, so took six months to get parts (I think electronics) for the second elevator. Both elevators working for maybe two weeks–and ANOTHER broken pipe two stories down from the first break (different apt)–and repeat the above scenario.
Why the electronics are NOT in a watertight compartment is unknown. Now, the elevators go to the TOP floor and wait to be called for sevice (not waiting on the first floor, which is most frequently used call location). Nobody has a clue–but me, it seems–as once there is another leak that flows into the elevator(s), the elevators are STUCK because they can NOT go down to–or below–the floor where the leak is. Just had ANOTHER leak on my floor this past summer–building main water supply broke at this level. Luckily, no problem with the water and the elevators as the water flowed away from the elevators (and into the units on my side of the building). Had two inches of standing water in my apt. As most of my stuff is on wheels or casters (3"-4" off the floor), not a lot of damage. Replaced one carpet (it needed to be replaced anyway). The other one dried completely and no mold/mildew.
still better than every single other car on the road.
Clearly, the leap from L2 to L3 is the most significant leap on the spectrum. Whereas L2 is considered as advanced driver support features, L3 marks the beginning of conditional autonomous driving, where drivers can legally take their eyes off the road when conditions are met. Strictly speaking, only vehicles classified as L3 and above are truly autonomous vehicles.
The only car where you can legally take your eyes off the road.
More:
Therefore, if an OEM wants to officially introduce an L3 vehicle, it must be liable for all potential accidents that occur while the vehicle’s L3 systems are switched on. That is, no matter how advanced and sophisticated the technology inside a vehicle might be, if the OEM is not ready to claim responsibility for accidents caused by its systems, the vehicle can only be classified as high as L2.
Didi began to develop and test autonomous driving vehicles (AV) in 2016 and its AV unit has raised hundreds of millions of dollars in investment from firms such as IDG Capital and Guotai Junan.
Didi allows users in some parts of Shanghai and the southern city of Guangzhou to hail self-driving cars through its main app. Swedish carmaker Volvo, owned by Geely (GEELY.UL), supplies Didi’s self-driving fleet.
You do understand that the car is not limited, right? The only limitation is that the lvl 3 autopilot will turn off once you go over 40. You are welcome to continue lvl 2 autopilot at higher speeds.
Tesla doesn’t have lvl 3 autopilot that works at any speed.
I don’t think that the 40 mph is a limit on the car. It’s a limit on when the Level 3 autonomy can engage. When the car’s going slower than 40 mph, you get to hand over the driving to the car - you don’t have to have your hands on the steering wheel or be monitoring the road.
Of course, I wasn’t suggesting that the car be limited to 40 mph, but that the L3 features were limited to 40 mph … among other limitations which don’t come to mind at the moment. I was merely suggesting that if one imposed the same limitations on a Tesla, it would probably do pretty good too … as well as doing pretty good at highway driving as well. At this point we only anecdotes and broad summaries to use in judging how well Tesla is doing and we have even less data for MB, except that they managed to convince someone somewhere of something.
I don’t think there’s any real factual basis for suggesting that, though. Tesla’s been pretty adamant that its system is Level 2 only. Mercedes’ system clearly has different capabilities than Tesla’s if it is capable of operating completely autonomously in certain defined situation, which is what a Level 3 system can do.
That’s not a knock on Tesla - they’ve been pretty clear that they’re not trying to use LiDAR and other sensors to get to “lower” levels of autonomy, and are instead going straight to "full self-driving. But I don’t think there’s any basis for suggesting that Mercedes’ systems don’t have capabilities that FSD does not yet have.
Next gen (geiger) lidar improves the ranging from 3,000 to 15,000 feet. It exists today but is not common (yet). This is a manufacturing, miniaturization (cost) issue but as we see with most technology the price comes down over time, sometimes quickly.
Likewise with processing, which is still abiding Moore’s law, plus or minus. So going from 40-to-60 would seem to be within reach within a year or two. Just a wild guess, obviously.
Lvl 3 is largely a liability hurdle, not a technology one. So, MB may already be able to handle lvl 3 at higher speeds but they don’t wish to take on the liability of such - and it very well be a very long time before a company has full lvl 3 (at all legal speeds) because of that liability hurdle.
To copy a quote from above:
Therefore, if an OEM wants to officially introduce an L3 vehicle, it must be liable for all potential accidents that occur while the vehicle’s L3 systems are switched on. That is, no matter how advanced and sophisticated the technology inside a vehicle might be, if the OEM is not ready to claim responsibility for accidents caused by its systems, the vehicle can only be classified as high as L2.
My understanding is that the Mercedes Level 3 is limited to daytime driving on freeways where surrounding vehicles are moving less than 40 mph and there is a vehicle in front within the range of the sensors. No lane changing is allowed.
Essentially it is a system for rush hour traffic jams on freeways where there are no traffic signals, pedestrians, left turns, etc. All the car needs to do is keep a proper distance away from the car in front and stay in its lane
No one is making a fuss about it because the technology is at least a year old. Mercedes has been allowed to use Level 3 on certain highways in Germany for about a year now.
My guess is that there are probably several programs that can work under these restricted conditions. The usefulness is pretty limited though so I can see why there is no rush to get approval by other companies. I doubt it moves the sales needle much and as we are seeing, the PR value is modest. https://www.caranddriver.com/news/a38475565/mercedes-drive-pilot-autonomous-germany/
still better than every single other car on the road.