Tesla Talk

For those knowledgeable about Tesla’s FSD model: As Tesla continues to train their vision model to manage new scenarios, is it likely that the model’s performance might degrade in other areas? I would assume that as FSD’s capabilities increase, so does the local compute resources required to run the more complex model in real-time. I’m drawing a parallel to large language models, where increasing the number of parameters to improve capabilities also leads to higher inference resource requirements. A model that is simulated and inferenced in the cloud will be far more capable because the resources available is many orders of magnitude greater than a than one ran locally on resource limited hardware.

That's only natural, isn't it? HW3 (like my 2017 car that came originally with HW2 but was upgraded for free to HW3) can do FSD quite well, but the compute power of the current HW4 is far stronger. This in time will be replaced by HW5, etc.

The difference we will likely be able to measure by stats like accidents or fatalities per 100 million miles. HW3 could end up 10 times safer than humans, HW4 100 times safer, HW5 1000 times safer, etc.

Tesla never needs to release another car again for the current share price to be justified. You make the mistake with many other people, that Tesla is just a car manufacturer like all other car manufacturers.

My point is just that the day to day volatility is a load of nonsense. Only if an event happened that would have significant impact on the net present value of all future profit, is a leap up or down justified. I.e. if Tesla announced that from tomorrow FSD is no longer beta, it is allowed to be level 5 autonomous (no need for a driver), people can get fully comp insurance with Tesla and if the car **** up, Tesla is responsible. Something like that has a huge impact on total future profit if it happened. Or another one: if Musk died tomorrow. Not if the company sold a few thousand cars more or less than expected in one quarter or if the gross margin was one percentage point lower or higher 🙄

FSD is in development with iterative incremental releases…it is naive to expect latest intervention data to be released between releases (perhaps when next version is released).

His job as CEO is not to overly focus on "wooing investors and get the share to surge" as you suggest. If anything he errs on the side of saying too much, in person and his X comments (tweets) and is criticised for it and lack of precision, given its a listed company and that rebounds (some examples below).

I can fully understand how there will be different views prevailing on new technologies but ..lets keep it real!

In my mind I do my best to seperate the "personal" traits from the business. I am not supporting his personal views (which may be more aligned with your views), but have made my points above and in earlier posts..so lets leave it there..

Tesla is a car manufacturer. What you're doing is telling people to invest in Spacex in the 1950s. Yeah it may pay off in the future but its likely decades away and very unpredictable. In the interim, it's just gambling because the price is speculative. The car figures do not justify the price. Adding in earnings from fsd, supercharging, and other nickel and dime stuff may help but it's nowhere near it.

So they either need to make the massive leap and be the only one to do so, in which case they will probably meet and exceed their current share price, or they won't make the leap and the volatility will continue until whenever mass consensus is reached that tesla won't reach the real fsd leap (or won't be leader/first to market) and the share price will settle and drop significantly.

This is why I say Tesla Motors and Tesla Innovation should be split. Keep the stable price on the car company and put the speculators like full "waymo style" FSD, robots, whatever other nonsense in the Innovation company. Otherwise you're literally staking the whole company and brand on the fate of some humanoid robots.

I know you put the "not just a car company" line in every post about the share price, and it's not fully incorrect, but the vast vast majority of their current, past and realizable future performance comes from cars.

Buying Tesla stock based on gain expectations based on the innovations stuff is like buying bitcoin before 2012. Yeah it turned out to be massively the correct thing to do , but anyone giving you a rational quant based answer for their investment decision is lying.

Until Tesla's regular adaptive cruise control can show the speed limit right, I'm not sitting in any of that self driving malarkey, let alone paying for it. Is there any link to the gps/map/database at all?

If it's smart enough to detect and respect the variable limits it sees on the overhead gantries on the M50, it shouldn't do that unless it doesn't jump up to 100 on seeing the roadside posts between them. And as for consistently detecting a change to 80 as you enter the Phoenix Park at Parkgate St, one of the most infamous 30 zones in Dublin!

Was that system actually made by Tesla or bought in? I believe first models used a system bought in an used by Tesla. Israeli I think.

until the owner gets arrested for espionage and engagement with foreign enemies. (see Wall Street Journal)

Why he dances in these circles is incredible.

Will we see the Space X taken over entirely by the US government for national security... who knows!

people can get fully comp insurance with Tesla and if the car **** up, Tesla is responsible.

Tesla Insurance product is certainly interesting and probably shows where telematics & insurance is heading.

Been watching reactions for a while in the US and I would never take up Tesla Insurance if it was available Irish market. So many complaints about it on the interwebs. As I do 20K+ km a year on motorways alone, I would get hammered too much for the actions of others.

…and I might not be able to enjoy driving a rear wheel drive 'robustly' on some of Ireland's great back roads!!

who was that with. I am aviva. Waiting on call back for quote.

I have decided on M3. Lr might be value if the tariffs already push the SR into the VRT ranges.

Just read that Tesla has now received 10 billion dollars in emissions credits from other manufacturers. Money for nothing 😁

They have been doing this for years. As other auto manufacturers exceed their relevant emission's targets they are forced to buy credits or face fines. Its a nice boost to Revenues and goes straight to bottom line…but I seem to recall that someone said Tesla are only a car manufacturer!

A bit tongue n cheek when I mentioned "just a car company". Tesla has a diversified business model that has many facets, some in development some revenue generating. Of course at this stage cars are a huge part but the carbon credits also rely on combined efforts across the board, EV manufacture, energy storage (powerwalls and megapacks). It shows Tesla's ability to monitise the opportunity, generating profits from its leadership position in drive for clean energy and other car manufacturers challenges inmeeting their emmissions targets.

These regulatory credits account for approx 3% of Total Revenues but because they go straight to bottom line they contribue a massive 43% of Tesla's Net Income for the first 3 quarters of 2024, which is perhaps not always widely known.