Tesla, TSLA & the Investment World: the Perpetual Investors' Roundtable

[cliff harris]

I must be on your ignore list and you may never see this, but have you even looked at Wayve? They just drive with humanlike AI and don't use millions of video. If they can do it, but in a different way without all the videos you claim, then no, Tesla does have competition. Add in Nvidia/MSFT/Softbank just invested a 1 bil+ with Wayve, and they already use OEM cars now, I see them as a viable hot/"startup" threat to Tesla who can actually just leapfrog pure vision self driving.

Sorta like if a smarter AI/human drives you around.

Watch some of their videos already in testing and ask yourself it they have 0 competition. There's also MobileEye driving around I see.

I am not ignoring you. Wayve is a TINY company (based in London) compared to Tesla. They are testing their systems on Ocado delivery trucks. Whoopdy-do. I'm sure there are millions of those (there are not).

If chatgpt has any lesson whatsoever, its the AI improves dramatically with scale of training data. There is a HUGE jump when you go from tons of data to ludicrous amounts. Nobody cared about openai until chatgpt 3.

Also, have we learned nothing from Tesla's early attempts at FSD? It is VERY easy to put together a decent tech demo, and impress people by showing an intervention-free drive. Tesla were doing that a long time ago. It turns out that going from 'we can get a decent video if we pick the day/weather and choose one out of 50 runs' is many orders of magnitude easier than 'drive anywhere on FSD. have fun'.

If FSD V12 was private and internal to Tesla, I would 100% agree with you, and would be very worried that cruise/waymo/wayve will be serious contenders. The fact is, there are thousands (tens maybe?) or people using FSD V12 all the time and sharing those clips on X and youtube. This is not cherry picked data, its real-world data.
I suspect wayvee are enjoying their moment in the sun, when they can raise ludicrous sums of funding from investors who have no idea how neural networks are trained or work.

 

[cliff harris]

Tesla wouldn’t have a monopoly on autonomy based upon data collected because the data collected from its fleet of 2million+ internet connected vehicles is applicable and usable only in vehicles with the same dimensions, camera geometry, etc as those 2million+ internet connected vehicles.

I don't believe this to be entirely true.
Its not like FSD is a single neural network, but many, in layers. Some of those layers are doing very generic stuff, such as stop-light and road-sign recognition, basically making sense of what all the objects are that the car can see.
The very 'end' network that makes decisions as to speed, direction, braking etc is going to be specific to each car, but the vast majority of it will be applicable everywhere.
We already have cars that have cameras with different specifications in the fleet, and indeed different hardware for inference.
Neural networks work exactly the same way our primitive brains work. The driving position for me in my wifes lexus and my model Y are definitely different, and the cars are different sizes and shapes, yet I can happily drive both .

 

[SOULPEDL]

I don't believe this to be entirely true.
Its not like FSD is a single neural network, but many, in layers. Some of those layers are doing very generic stuff, such as stop-light and road-sign recognition, basically making sense of what all the objects are that the car can see.
The very 'end' network that makes decisions as to speed, direction, braking etc is going to be specific to each car, but the vast majority of it will be applicable everywhere.
We already have cars that have cameras with different specifications in the fleet, and indeed different hardware for inference.
Neural networks work exactly the same way our primitive brains work. The driving position for me in my wifes lexus and my model Y are definitely different, and the cars are different sizes and shapes, yet I can happily drive both .

Plus Elon said the NN would transfer easily into a new envelope (paraphrasing) - was during the Semi reveal as I recall.

No need for proof, humans can also drive different types of cars. "This one seems a bit harder to park..." is typically all we here as the result of a shape change.

 

[elasalle]

Don’t care about the tour, do care about the CEO apparently blackmailing the shareholders by saying he’ll take his ball and go home if he doesn’t get 25% of the voting shares. We already have a part time leader at Tesla - I happen to think the board does not do it’s fiduciary duty by just rubber stamping everything Elon wants.

Even if BoD did not do its duty, the advantage was with shareholders at all times ….either Elon did not get paid a cent, or he grew the company to unheard of levels.
Either way shareholders won.

 

[jeewee3000]

Anecdotal evidence: Tesla Belgium seems to be hiring.

 

[Christine69420]

Don’t care about the tour, do care about the CEO apparently blackmailing the shareholders by saying he’ll take his ball and go home if he doesn’t get 25% of the voting shares. We already have a part time leader at Tesla - I happen to think the board does not do it’s fiduciary duty by just rubber stamping everything Elon wants.

Deciding to vote against the board because of something he CEO apparently did is a very flimsy way of taking care of your investment.

I would myself try to be more certain of the facts before making such an important decision.

Unless you only have a handful of shares. Then it does not matter much.

But if you are seriously invested then

 

[AndrewZ]

I don't believe this to be entirely true.
Its not like FSD is a single neural network, but many, in layers. Some of those layers are doing very generic stuff, such as stop-light and road-sign recognition, basically making sense of what all the objects are that the car can see.
The very 'end' network that makes decisions as to speed, direction, braking etc is going to be specific to each car, but the vast majority of it will be applicable everywhere.
We already have cars that have cameras with different specifications in the fleet, and indeed different hardware for inference.
Neural networks work exactly the same way our primitive brains work. The driving position for me in my wifes lexus and my model Y are definitely different, and the cars are different sizes and shapes, yet I can happily drive both .

Have you tried driving other cars with your eyeballs in different positions?

Certainly some stuff would be pretty easy to transfer over, but if all this was such an easy thing then we wouldn’t see these lags and differences in the other models within Tesla’s own lineup much less trying to work this out with other companies. Now you need to get cameras into all these other new brands/models in at least roughly the same positions, then you need to get enough of them on the road to begin collecting significant amounts of data to feed into training, then the training etc etc.

Bam you now have even more hardware iterations floating around out there — I’m old enough to remember the camera retrofit debacle. How all this would play out is so far from certain, if it even went in that direction in the first place, and competition would surface.

 

[DarkandStormy]

View attachment 1051349
Anecdotal evidence: Tesla Belgium seems to be hiring.

Why post this without bothering to look it up?

 

[unk45]

Well, as far as I am aware, the batteries themselves aren't responding in nanoseconds. The system may bid that fast, but the response times are in the 100s of milliseconds.

Quite correct. The Autobidder system response to changes is faster, but battery response follows that, even more slowly in cases of largely distributed VPP nodes. That means by the time actual responses are effective could be much slower, even more than a second.

The reference was to the Autobidder system, not the delivery of it's response.
Remember though, that pre-Autobidder operating with stationary storage response times were often measured in 15 minute increments when things worked optimally.

That is why Autobidder ended out revolutionizing the way utility fluctuations were managed, and spawned the continuing flood of peaker plant replacement with stationary storage.
It's quite hard to overstate the significance of the Hornsdale Power Reserve.

 

[GOVA]

Plenty of new pimples popping out lately over here. Bullish.
Same was observed in 2019.

 

[Leftlane22]

Deciding to vote against the board because of something he CEO apparently did is a very flimsy way of taking care of your investment.

I would myself try to be more certain of the facts before making such an important decision.

Unless you only have a handful of shares. Then it does not matter much.

But if you are seriously invested then

It’s not because of what the CEO did, it’s that the board is not independent by any stretch. I am a decent sized shareholder (for my net worth) I’m bullish on the stock, and I bought more this morning. Doesn’t mean I have to vote yes on everything the company wants, does it? Not sure how voting for massive dilution is taking care of your investment.

 

[sunwarriors]

I am not ignoring you. Wayve is a TINY company (based in London) compared to Tesla. They are testing their systems on Ocado delivery trucks. Whoopdy-do. I'm sure there are millions of those (there are not).

Just to clarify, it's not just delivery trucks. I see Mach-E, Jaguar, is that an Alfa Romeo? as well in videos so if anything, the "AI" deals with different vehicle types vs. Tesla having trouble launching Autopilot even on CT currently.

People talk that small/nimble startups are great when they talk about Tesla vs. incumbent, but when it's a competitor, that's a bad thing now? They simply don't need that much $$ now/yet.

 

[hobbes]

https://twitter.com/x/status/1795474823675572464

 

[hobbes]

Toyota doing Toyota things

https://twitter.com/x/status/1795452926426538118

 

[EVMeister]

It’s not because of what the CEO did, it’s that the board is not independent by any stretch. I am a decent sized shareholder (for my net worth) I’m bullish on the stock, and I bought more this morning. Doesn’t mean I have to vote yes on everything the company wants, does it? Not sure how voting for massive dilution is taking care of your investment.

You clearly have no idea what you're voting on here. Do us all a favour and abstain from voting instead of basing your decision on half-baked inaccurate FUD.

 

[Leftlane22]

You clearly have no idea what you're voting on here. Do us all a favour and abstain from voting instead of basing your decision on half-baked inaccurate FUD.

Hey - I found another one who thinks they are smarter than everyone else and only their opinion can be correct! No chance I read all the proposals and considered each one before voting - that would take the intelligence of a Lover of Tesla.

 

[Knightshade]

Plus Elon said the NN would transfer easily into a new envelope (paraphrasing) - was during the Semi reveal as I recall.

You mean the semi that, like the cybertruck, also lacks FSD today?

 

[elasalle]

Not sure how voting for massive dilution is taking care of your investment.

Short term high(no share dilution) != long term gain (+ Elon)

OK continue buying more stocks ...

 

[EVMeister]

Hey - I found another one who thinks they are smarter than everyone else and only their opinion can be correct! No chance I read all the proposals and considered each one before voting - that would take the intelligence of a Lover of Tesla.

You think voting "no" for Elon's compensation plan is a smart move to avoid share dilution? Imagine you work for a salary and despite being promised that salary it is not given to you. Does that motivate you to continue doing your job?

 

[M|S|M MYP]

You think voting "no" for Elon's compensation plan is a smart move to avoid share dilution? Imagine you work for a salary and despite being promised that salary it is not given to you. Does that motivate you to continue doing your job?

Reminded me of this: