In other words, features beyond advanced summon / NoA.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Tesla, TSLA & the Investment World: the Perpetual Investors' Roundtable
- Thread starter AudubonB
- Start date
In other words, features beyond advanced summon / NoA.
mblakele
FSD Beta (99)
In the Lex Fridman podcast released today, Musk mentioned that he's been using a build that stops and starts for stoplights.
You think that they would hold an event for Adv Summon/NOA when those features are already active on Tesla cars......out in the wild....as in any investor can test drive the features on their own right now. Right.......
I believe they already explaned the reason for leasing:
50% of the car sale price will be paid off by the leasee by the time it's returned. If you subtract the margin, they'll get the car with 30k miles on it for some 30% of the original price. I'm not proficient with leasing terms, but this seems as a better deal than building $35k cars specifically for the Tesla Network.
Btw, this interview that was just posted
Check this out @22:22 :
So, here we have FSD matching human capability at EOY 2020. If they have 200%+ safety, they apply for regulator's approval. He stated exponential rate of improvement. So, this explains why they think 3 years from now they can use the cars in the Tesla Network.
Paraphrasing Musk: "I don't want to sound overconfident, but we are beating everyone's arse to FSD".
sundaymorning
Active Member
If you were worried about demand you should be glad leasing is here. Dropping the price of a vehicle by $100-200 off monthly payments will attract a lot of buyers who don’t want to own a car past 50k miles due to maintenance. Lots of business owners would love to lease for tax write offs; lastly there will also be lots of buyers who don’t want to commit long term to a new technology they don’t yet understand, so leasing is the right move.
Keep in mind that the Model S went through multiple price drops in the past 6 years and yet we still got ourselves to 25% margins, why? Because as time passes, cost savings from efficiencies, supplier renegotiations as well as high volume reduces the costs of M3s and offsets the price reductions. Pushing back $35k standard also buys us some time until Shanghai gets setup (Munroe thinks margins in Shanghai could be as high as 40%). Remember when the headlines use to be “Tesla loses $80k for every model 3s produced”? Or Model s losses $200,000 per vehicle sold? So we’ve come full circle to Model 3 now, and guess what? It’ll be the same cycle with model y.... what Tesla lacks currently is variation amongst its cars, we’re still only selling 3 models but we do have 3 more in the pipeline slated to arrive sometimes next year (we’re buying ourselves time until more products arrive).
Currently the Model 3 is so good that’s it’s osbourning the S, what this means is that Tesla will need to refresh the S and make it that much better. Do you think they won’t be able to outdo themselves? Do you think that it’s not in their DNA to produce a even better product? Give it 3-4 quarters and we can have this conversation again. Like everyone else I think a refresh is coming, Elon isn’t going to go out there and drive demand down himself by openly declaring a refresh. But I do think Tesla misjudged how severe the Model 3 will osbourn the S. Another mistake is that Elon also misjudged how severely he has damaged morale by his incessant tweeting, how much he has twisted the knife on his own brand... I hope he learns from this and turns his image around and starts to rethink about how he approaches Tesla’s image. Instead of fighting the media he should be courting them. He doesn’t need to court every outlet, just a few to offset the negative Nancy’s and buy us some time before real mass adoption occurs. The Model 3/SX was suppose to finance all future projects, if it appears like it doesn’t, then raise capital. Don’t be stubborn and raise some funds to get us to Y and truck sooner.
Last edited:
No, if it’s Tesla HQ, you won’t see any other car brand, they either drive a M3 or take the shuttle.
Krugerrand
Meow
For some people work is too much work. You have to arrive at a certain time, stay a specified period of time, do the same work all the time, rinse and repeat week after week, sometimes do overtime. I’m exhausted just thinking about it. Way easier to get live on the freebies provided by others.
sundaymorning
Active Member
I think Tesla needs to turn on leasing so they can have enough data to tell investors emphatically during CC/Earnigns that demand is strong. Knowing that 80% of customers opted for standard plus was a confidence boaster. Eventually leasing was going to be unleashed, might as well do it now so the company can extrapolate into the future. Waiting too long might be a detriment. They really neeed to get in front of this and aggressively grab as many leases from BMW Mercedes, Lexus, Audi as possible. Why give up months of leasing to competitors when you can steal them now?
The demand question calls for it.
The 3-year timeframe on the lease is probably when the HW4 computer will come out...
Sparky
Member
Thanks for this.
OT: This was how an interview of Musk should be done. Good questions, focused on a critical topic of which too little is understood.
Plus, I loved that Lex seems to be auditioning for the next Rod Serling / Twilight Zone remake. He was fantastic.
bdy0627
Active Member
Did you check out the podcast interview of Elon by the MIT researcher? If not, you gotta check it out. It's a great interview and made me feel rather bullish again on the self driving tech and therefore Tesla. If that hits over the next 6-9 months, which I now think is quite possible, the market is going to be pretty shocked. Plus, Elon looks and sounds great. Wayyy better than last year when he seemed on the verge of losing it. He's relaxed and confident, and that's great to see. I really do think it will help you feel more confident in your TSLA investment. The market believes the FUD that Tesla's autonomous driving tech is vaporware. Meanwhile, that Tesla neural net is exponentially learning how to drive, and it seems to have crossed the tipping point. These issues Tesla is having with production ramping, delivery logistics, and figuring out the right trim mix and prices are really frustrating and disheartening. I am confident they will figure it out eventually. I actually like the current trim mix and prices. I don't know a lot about leasing but it sounds like Tesla has set it up so that it is very favorable to them. That will probably limit the popularity of leasing for now, pushing people to buy instead. If they were absolutely desperate to increase demand, I think they would be doing leasing on favorable terms to customers.
I’m a big fan of Lex but disappointed in his questions.
Apparently he had a bunch of them tee’d up about the hand-off between FSD and the human, but I would have thought he could have pivoted better when Elon basically told him that’s all moot, because FSD will surpass humans in 6-18 months.
I’m no AI expert, but it seems like one central issue is that each 9 in the “March of 9’s” becomes more and more difficult because of edge case explosion. I wanted more detail on how Tesla is handling this edge case explosion without it involving increasingly intensive amounts of human labor in labeling...
To take one example of road hazard detection:
It would seem that the car has to classify every object in the road along a spectrum of how safe it is to hit, vs. how imperative it is to avoid.
To get to 99%, it may be enough to identify pedestrians, wheels, plastic bags, roadkill and potholes.
To get to 99.99%, there may be hundreds of different road hazards, going up to 100’s of thousands as you march to 99.999%.
Am I missing something? Does someone know how Tesla can handle edge case explosion?
Last edited:
I'd like to highlight 1 point Elon made around 8:30 in the video (which is not new information per say, but the way he puts it is eye-opening): Tesla having close to 500K cars with sensor suit collecting data on the road vs all autonomus driving competitors all-together having in the order of 5K cars or less means, that Tesla collects about 99% of the data in the field. That is a huge advantage! Some would even say YUUGE
Krugerrand
Meow
So that’s a no? In which case Tesla needs an alternative location. If a European country won’t move fast enough for Tesla, Elon will do something different. What are the most logical possibilities?
While the interview mostly lacked detail (but high level of confidence, must have pissed Blader off! lol), there was again validation that Tesla uses data from disengagements of Autopilot to send back to improve edge case learning. It sounded like only in specific instances though.
Made me think, I think there is going to be rapid iteration on what events are triggered to send back based on the current state of the FSD algorithm. For instance they only probably started collected some triggered disengagements of lane changes when autopilot suggested it, and the user declined (probably not accurate enough) or when the user accepted, but then disengaged (realizing some object /problem).
Meanwhile, stuff like traffic light detection has been developed with in-house data, but once it's deployed to a certain level of accuracy, only then will certain events that the user disagrees with the algorithm will a trigger save the data and send the edge case back to Tesla.
This means that Tesla's incremental development / deployment cycle is not just about making money / providing a nice L2/L3 product, but it also directly is beneficial to iteratively collecting edge cases over time as the algorithm is in good enough shape to trust the algorithm estimate / user input divergence.
Made me think, I think there is going to be rapid iteration on what events are triggered to send back based on the current state of the FSD algorithm. For instance they only probably started collected some triggered disengagements of lane changes when autopilot suggested it, and the user declined (probably not accurate enough) or when the user accepted, but then disengaged (realizing some object /problem).
Meanwhile, stuff like traffic light detection has been developed with in-house data, but once it's deployed to a certain level of accuracy, only then will certain events that the user disagrees with the algorithm will a trigger save the data and send the edge case back to Tesla.
This means that Tesla's incremental development / deployment cycle is not just about making money / providing a nice L2/L3 product, but it also directly is beneficial to iteratively collecting edge cases over time as the algorithm is in good enough shape to trust the algorithm estimate / user input divergence.
It also reminded me how brilliant Elon is.I'd like to highlight 1 point Elon made around 8:30 in the video (which is not new information per say, but the way he puts it is eye-opening): Tesla having close to 500K cars with sensor suit collecting data on the road vs all autonomus driving competitors all-together having in the order of 5K cars or less means, that Tesla collects about 99% of the data in the field. That is a huge advantage! Some would even say YUUGE
Plus their incremental approach is front loading the road with hardware platforms. Let's just say that Waymo get's full FSD figured out in 2021 and Tesla is 1 year behind. At that point Tesla has 1 million FSD capable cars on the road. In 2022 Waymo has a fleet of 100k cars running FSD, Tesla has a fleet of 1.5 million running FSD. Who owns the market?
They don't need HW4. HW3 with double redundancy is only using 10% of compute capacity at full resolution, full frame rate.
They can always throw them in containers stacked up like the Ukranians do with the LEAFs or ship them take free overseas as used cars. Should be plenty of demand
You seem to be assuming the car would need to classify exactly what each object is. But that's not really necessary. It just needs to classify how an object will likely behave. The latter tends to be an easier problem and less prone to edge cases than the former.
Similar threads
- Locked
- Locked
- Locked
