One way never to fail is no never do anything. One of the outstanding characteristics of America is that failure is not the end of the line, you get up and give it another try.
The Captain
4 Likes
AI being the ultimate software to do just that. It is an endless task set in motion.
At the point that insurance premiums can come down on your EV because safety is better and less costly than in an ICEâŚthat is when AI crosses the threshold into common usage as FSD âin economic termsâ.
The hilarious quote from the quarterly call, being swallowed and endlessly repeated, is that Tesla is âbetween two waves of growthâ. Well, itâs certainly possible, but a more accurate way to portray it is âwe had tremendous growth, and at the moment it isnât so tremendous. Of course we hope for another wave, and thatâs what weâre trying to accomplish.â
The things arguing against include 1) the proliferation of competition, 2) the troubles at the Mexico site (not yet begun), 3) the reset bubblicious predictions of âno more ICE after 2030â, 4) trouble on the assembly line (in Sweden and now possibly the US) in light of his demand for even more compensation, 5) the possible extraction of AI from future Tesla evaluation, and, of course, 6) the mercurial attention being given the business by Musk himself.
On the other side, 1) introduction of a new lower priced model two years hence, 2) walking, talking robots (?), 3) Full Self Driving is suddenly solved, 4) margin compression reverses as dramatically as it imploded. Which of these things seem likely?
Again, I believe Tesla to be a major, able competitor. But the conditions that led to its first great growth wave: unimpeded market, large subsidies, EV bubble mentality, low interest rates, technological advancements - seem unlikely to be repeated - or replaced by similar advantages.
Time will tell, but as for now, âWe had a great growth wave, and now weâre on the other side of thatâ seems more honest.
8 Likes
You left out a couple of important factors.
- EUâs ban on new internal combustion vehicle sales in 2035.
- Rapid decline of internal combustion car sales in China due to new emission standards. Gas Car Sales Collapsing in China Now and Will Collapse in Europe By 2025 | NextBigFuture.com
- California ban on new gas powered car sales by 2035.
India will also very likely limit internal combustion vehicles in its major cities for air pollution reasons.
The globeâs biggest auto markets will either have to compromise on their environmental goals or create an enormous demand for the next generation affordable Tesla, a demand that will increase every year as the deadlines approach.
As for Bots, Iâm not sure why the cynicism. BMW will be testing humanoid Bots from another manufacturer in its South Carolina plant. No reason I can see why these couldnât greatly increase manufacturing automation. Seems like a big deal. First Bipedal Robots With Hands Coming to BMW Factory
3 Likes
The reason for the cynicism is that thereâs no real reason to use a humanoid form for robots, particularly in auto manufacturing, since robot âbrainsâ are nowhere near the point where it would be useful for them to have a humanoid form factor. I suspect Figureâs move with BMW is far more likely an effort to raise their profile for their start-up capital raises and/or (eventual) IPO. Itâs an effort to sell equity, rather than robots, because thereâs probably nothing that these robots can do in a factory right now thatâs useful to the factory. No downside to BMW - having these things in your factory makes it look cool, and itâs a nice PR balance to what Tesla might end up doing.
As for the other stuff, I think that Goofy was suggesting that those bans on ICE sales in Europe and California are written in water, and will get delayed or watered down altogether as we get closer to them actually biting. Itâs one thing to pass legislation requiring something that folks are already going to do (the assumption that weâre on an S-curve), and quite another to mandate a switch to EVâs if theyâre too small a segment of the market for that to work seamlessly.
5 Likes
I imagine all salaried workers are cheesed off that they wonât receive their usual stock compensation bonuses. But Elon is pushing the board to DOUBLE his stock holdings! LOL
The hourly factory worker only received a bump because of the union gains at the big three. Musk has managed to keep the UAW out of his factories so far. If the union make another attempt it is likely they have tested the waters to gauge the amount of disgruntlement among those folks. And remember Musk claims those hourly employees will need to sleep on the factory floor to ensure the success of the Model 2. A call for sacrifice without compensation which apparently he is unwilling to do. Of course he has an âinâ with his friends on the board.
I suspect Musk along with other JCs are muttering that they can hardly wait until those âgreedyâ employees can be replaced with compliant robots & more automation & AI.
1 Like
As a car company yes.
As an AI company in timeâŚif FSD is less expensive in insurance costs and human and property costs in general, then there is another wave.
Granted just beating the insurance costs wonât matter that much.
Controlling AI is a valuable thing. I am still a doubter in its value.
Iâm remembering âWeâre outlawing incandescent bulbs because the new twirly ones are so much more efficient.â Which was followed by âWeâre outlawing incandescent bulbs and twirly bulbs because LEDs are so much more efficient.â It took from 2007, when the rule was first passed, to 2023 for it to come true. And this for an item thatâs a dollar twenty-nothing and not opposed by some of the largest manufacturers in the country.
Almost none of the most meaningful environmental efforts have come to fruition: nuclear hasnât replaced coal, solar hasnât replaced nuclear, the plastic patch in the ocean continues to grow, natural gas production (yes, a fossil fuel) is at an all time high, oil production hasnât peaked, recycling (with a few exceptions such as aluminum and paper stock) is a joke, and so on. Not that we havenât made progress, we have. But weâre loathe to institute even the smallest changes - soda bottle/can deposits, for instance. Or a tax on wasteful packaging.
And people think theyâre really going to outlaw all internal combustion cars in a decade? Donât make me laugh. OK, make me laugh. Itâs the funniest thing Iâve heard this year, although I admit the year is young.
EVs top out in the US at maybe 50% within any reasonable lifetime. Eventually, someday, with enough ifs and butts and candies and nuts higher, but not that anyone here is going to see it. Heck, Norway, for all its storied publicity, sales figures, advantages and progressive legislation, still has 75% of vehicles on the road burning gasoline and diesel.
Reality bites, maybe, but there it is.
8 Likes
And to clarify, thereâs considerable doubt that theyâre really going to outlaw all sales of new ICE cars within that decade, also. That might have seen plausible when it looked like Europe as a whole might follow the âS-curveâ rate of adoption of Norway. But adoption slowed mightily in 2023, with BEV sales actually declining in Q4 2023 compared to the prior year. Weâll see more this year and next, but the Norway âS-curveâ looks to exist only in countries that have extraordinary subsidies - and only for as long as those extraordinary subsidies last. Now that those subsidies are being cut back, future adoption might be much, much slower.
1 Like
You shouldnât make assertions like this without having some appreciation of the power of machine learning. There are probably hundreds of papers on the topic by now, but here is one.
A fundamental mechanism for rapid learning in humans is imitation: children routinely learn new skills (e.g., opening a door or tying a shoe lace) by imitating their parents; adults continue to learn by imitating skilled instructors (e.g., in tennis). In this chapter, we propose a probabilistic framework for imitation learning in robots that is inspired by how humans learn from imitation and explorationâŚWe present results demonstrating that a 25-degree-of-freedom humanoid robot can learn dynamically stable, full-body imitative motions simply by observing a human demonstrator. Learning Actions through Imitation and Exploration: Towards Humanoid Robots That Learn from Humans | SpringerLink
The advantage of a humanoid form is that the robot can learn directly by observing a human perform the task. That is a big advantage over conventional automation that relies on traditional programming. For one, the humanoid bot can potentially perform a variety of tasks, becoming a general purpose machine.
Sure, if it is demonstrated to be economically impossible to achieve the objective. However, if Tesla starts mass producing its $25K model in 2025 and BYD shows it can make a profit on its lower priced BEVs then there really isnât a reason to compromise environmental goals that many (other than you) think are really important. 2035 is a decade from now. Who would have thought a decade ago that Tesla would today have the best selling model in the world or that 25% of new cars sold in China today would be EVs?
Back in 2014, Tesla was mostly pushing the Model S and there was a lot of skepticism in China about EVs.
Less than 7,000 electric and hybrid vehicles were sold in the first quarter of 2014. https://www.cnn.com/2014/04/24/world/asia/china-electric-vehicles/index.html
At least try to be accurate. The plan is to ban sales on new ICEs by 2035. Meanwhile CAFE standards keep rising.
4 Likes
Except for, well, political pushback, same as happened with lightbulbs. Bottle deposits. And hundreds of other environmental issues that have been derailed along the way, with far less personal effect than âbanning the car you want to driveâ, or other issues like âcharging at an apartment building or condoâ, âcharging in a rural environment, and so on.
Yes, I clearly meant ânew car salesâ. You admit as much by noting âCAFE standardsâ, which apply only to new car sales.
2 Likes
You are conflating.
The ability to âobserveâ is not dependent on the shape of the observer. In other words, a machine can observe and learn just as easily without having a foot or a shoulder.
The physical form of the machine that does the work is not necessarily bettered by having a foot. See the Amazon warehouse robots for an example.
1 Like
I donât want to be repetitive - weâve talked before about how watching a human only allows the robot to acquire information thatâs conveyed visually, so that watching a human pick up a box doesnât teach the robot how much force to use on the box. And again, humans in auto factories (and assembly lines more broadly) generally arenât used as âgeneral purposeâ workers - they usually have one job on the line. So thereâs very little benefit to having a robot that can switch jobs or move around.
Oh, I think those environmental goals are very important. I just donât think that politically, either Europe or California are going to tick off their voters by telling them they canât buy their ICEâs any more. If consumers actually were already buying 85% EVâs on their own, that would be one thing - but I donât think that level of background adoption is at all feasible in a world where subsidies are being cut back and interest rates are back to historical norms. Sure, Tesla might come out with a $25K subcompact or city car (though color me skeptical on Tesla hitting their price target), but that model segment is only a piece of the overall market (even in Europe). And I think BYDâs going to get quite a bit of trade barrier resistance - the EU is very likely to regard protecting their domestic auto industry as a pretty good reason to compromise environmental goals, much as the U.S. already has.
2 Likes
Life is tough! Life is unfair! Life is a drag! 
The Captain
1 Like
Also not wanting to be repetitive but on this topic I believe you know far less than you think you do. Robots learn about force by seeing whether the box is damaged or not. This is a general public article but does have some links to papers. It describes how visual learning translates to nuanced object manipulation and heightened dexterity.
Basically, just as people learn by watching YouTube tutorials, these researchers are using videos to teach their robots how to do some things, and their approach has boosted success rates in new test settings by an impressive 58% compared to traditional robot data training. These groundbreaking studies collectively pave the way for robots capable of nuanced object manipulation akin to human abilities. Such advancements could redefine industries, from manufacturing lines to operating rooms. Imagine a surgical procedure where a robot, powered by AI, assists a surgeon, enhancing precision and outcomes. AI Researchers Are Teaching Robots to Mimic Human Dexterity - Decrypt
You are also thinking of the old model of linear assembly lines. Tesla is changing that to a more modular approach that has the capability of modifying production methods in real time. Itâs a long story but this somewhat out of date article gives some examples:
The bigger challenge in automating the Model 3 is that the Model 3 changes frequently, and Elon Musk wants the line to keep up with product iterations, which means the line has to be flexible enough to quickly adapt to new parts and changing processes. The traditional car companyâs manufacturing method cannot meet the requirements. In the original assembly line, each process repeated only one fixed process, and the goal of automated equipment such as robotic arms was to complete a single step quickly and precisely, not to adapt to change. https://www.topcarstesla.com/blogs/news/tesla-super-assembly-line-evolution-history
If the future of efficient manufacturing is constantly evolving processes that can change in real time, then general purpose machines/robots are a necessity.
It certainly helps. Much machine learning is based on how children learn, with mimicking being an important part of the process. Tough to mimic if you donât look like what your are trying to copy.
If the goal is to perform a variety of tasks similar to what a human can do then maybe having a foot helps.
The humanoid robot market is estimated to be a $13B industry in six years with CAGR of about 50%. Apparently lots of people think it is more than just a gimmick. Humanoid Robot Market Size, Share, Opportunities & Forecast
3 Likes
Well, this is what you said ââŚAnd people think theyâre really going to outlaw all internal combustion cars in a decade?..Heck, Norway, for all its storied publicity, sales figures, advantages and progressive legislation, still has 75% of vehicles on the road burning gasoline and diesel.â
Doesnât sound like you are just referring to new car sales.
2 Likes
He is on the autism spectrum, not ADHD.
3 Likes
The robots in the videos were not humanoid.
1 Like
A big part of the problem is activist (including people in power in various governments) overreach. They tend to want to panic people and then to make grand pronouncements like this (âeliminate ALL new ICE vehicles by 2035â, âsea levels will overrun all of the Florida peninsula in 15 yearsâ, etc). Meanwhile, I just read that about 40% of ALL carbon dioxide released into the atmosphere comes from coal burning, mostly for electricity. So I would think that there would be a much better place to start for the biggest bang for the buck. Instead of all the panicky pronouncements, how about instead trying to eliminate as much coal usage as possible. But no, instead they try to grab onto political things that can increase their power, like âbanning the sale of new ICE vehiclesâ or the various scare tactics.
Because the sources of CO2 globally are different from the sources of CO2 in the countries doing the regulating. Here in the U.S., only about 19% of our energy related emissions come from coal - compared to 46% from petroleum. So if you want to reduce global emissions, your low-hanging fruit is coal consumption for electricity generation in China and India; if you want to reduce emissions in your country, it might be something else. And because regulators have virtually no power to effect changes in other countries, that makes a big tilt towards regulating what you can in your own backyard.
1 Like