Indeed TSLA has made a lot of money with BTC. So much money, in fact, that BTC profits are bigger than earnings coming from selling cars/energy/solar panels/what-have-you during the whole cumulative history of the company. Well done!
Let’s push the thought experiment a bit further:
Management realizes that there is much more money to be made in the crypto space than in its current business.
TSLA liquidate all its non-crypto assets : factories, machines, inventories, etc, and manage to sell them at the value carried on the books.
Management restructure the company in a crypto holding, and reinvest the liquidation proceeds in BTC/ETH/etc.
The market capitalization converges towards the net asset value of the crypto holdings, which would be around $21 billion. The share price is now $22.
In a subsequent SEC 8K filing, the company states that effective immediately, the titles of TSLA investors have changed to Archdukes of BagHolding.
I think the driving scenario from the first video (many lanes, many parked cars, pedestrian crossings, traffic lights) is already a quite demanding one. Put a tram somewhere in there, wipe out some lane markings and it’s close to the toughest driving conditions you can imagine.
Indeed it’s not impressive. I still think the right approach to solve it is through machine learning. You cannot program every scenario, you have to let the car software do its magic also in unimplemented situations. For this you need millions of miles driven with human feedback correcting erroneous neural network.
I’m puzzled as to why some automakers still try to achieve FSD with LIDAR. A human can manage with two cameras, why can’t a car manage with 8? (I wonder though, if they will have to implement some camera wiping mechanism to remove dirt/water) LIDAR is expensive and bulky. Recently Elon said, they will even try not to rely on radar, I guess all in the effort to save costs. The same question applies to pre-mapping a street. A human does not need that, doing this to the car is just cheating, and does not scale well.
By the way, the non-fsd Model 3 that I’m driving displays traffic lights and even reacts to them. When you stand at a red light and it turns green, the car will make a “ding” sound to let you know it’s time to drive. But often times, when waiting I look at the screen, and there the light rapidly changes from red to green to turned off, plus everything it renders shakes like it’s some hallucination.
I’m puzzled by the reverse Why cripple yourself voluntarily? (as expected lidar cost went down by an order of magnitude or more, it’s like 500 bucks now)
Sounds a bit like luddite thinking, refusing a technology out of principle. Imagine doing that for e.g. GPS (expensive, a paper map is fine to navigate around, humans handle it easily).
Good point about falling costs, didn’t know that. I read that right now you can get a LiDAR system for $1000, on the way to have it for $500, or even $100 in the long run. If LiDAR proves superior, that is terrible news for Tesla, which has equipped cars with cameras, and probably it would not even be possible to put LiDAR anywhere.
But the question is if LiDAR is really the right tool for the job. Yes, it can make a 3D map of the surroundings, of the shapes, but it’s not be able to discern a flying plastic bag or a falling leaf from a real obstacle. It will not recognize street signs, traffic lights, pedestrian crossings. It only answers the question: is anything in my way? (and it can easily give a false positive in case of the mentioned plastic bag).
That’s why people use a bunch of inputs: usually camera + radar + lidar. They work under difference circumstances, you usually merge all the inputs into one scene.
but heavy rain disturbs Lidar, then you have to have camera based system, and if you have solved camera based system in heavy rain, then you don’t need Lidar when it’s not raining. Heavy rain hinders LiDAR sensors | The Engineer The Engineer.
having Lidar does not mean that you can avoid solving camera-based problems.
exactly. But to ely on camera you need to solve the vision and AI problem. And once you solve that, you can recreate the same data as LIDAR by using stereoscopic camera and understanding depth from pictures, complementing it with radar and ultrasounds.
Lidar + Ultrasound + Radar cannot work because of hardware limitations. You need camera. But then you need to solve the vision problem with AI
Camera + Radar + Ultrasound would work without LIDAR if you solve the software / AI problem.
Which is not a given, and is not certain that Tesla can effectively solve that with that long tail of corner cases
To be honest I don’t have personal experience in the field, it’s only what I’ve heard from the engineers working on autonomous driving (that voluntarily degrading the inputs makes things harder the more you have the more feasible the problem becomes).
As soon as they’ve enthroned their own twittering spiritual guru to rival the Technoking
TSLA is going to be so toast, you can smell it from Grünheide to Wolfsburg!
This is very interesting. The latest video by Tony Seba from RethinkX.
Apparently, when governments decide to build another electricity plant, they use LCOE (levelized cost of energy) to decide if the investment is economically viable.
So, according to them, the conventional energy market will get heavily disrupted by SWB (solar - wind - battery) solutions in the coming years. If anybody’s building a fossil fuel power plant now, it’s just money down the toilet.
Now think, how many energy companies there are today. Oil extraction, refinery, distribution. All this looks overvalued today. And if that’s the case, which company do you think is best suited to benefit of this disruption?
I might agree with the conclusion, but not with the argumentation.
They present it like if it’s a conspiracy…but the capacity factors of coal plants in the western world are trending down simply because they’re already phasing them out, the UK will shut them all down within a couple of years I believe. They should not extrapolate anything from those data.
The random trend line they draw out of the gas plants chart made me giggle…where is this coming from??
To be clear I’m far from thinking that building new coal plants is a good idea, but their argumentation is just really far fetched.
If their analysis does not take this into account, then it’s really misleading, indeed. But still, if it’s true that solar/wind/battery costs will continue to go down in the following decade, they will displace all other energy sources. We’re at a point where we still need old power plants to provide power during the evening peak, but if battery infrastructure keeps expanding, these peaker plants will become obsolete.
By reading and partipating to this forum, you confirm you have read and agree with the disclaimer presented on http://www.mustachianpost.com/
En lisant et participant à ce forum, tu confirmes avoir lu et être d'accord avec l'avis de dégagement de responsabilité présenté sur http://www.mustachianpost.com/fr/
Durch das Lesen und die Teilnahme an diesem Forum bestätigst du, dass du den auf http://www.mustachianpost.com/de/ dargestellten Haftungsausschluss gelesen hast und damit einverstanden bist.
Why cripple yourself voluntarily? (as expected lidar cost went down by an order of magnitude or more, it’s like 500 bucks now)
