Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

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

This site may earn commission on affiliate links.
The consumer price index, flawed though it is, has increased by 6.34x since 1974, whereas 2400/35 = 69. Using the price of gold as a proxy for overall inflation is ludicrous. Off by an order of magnitude. Prices in general are obviously nowhere close to 69x higher than the were 50 years ago. If that were the case, gas would be $36/gallon, the median US household income would be $780k, and the median US home price $2.7M.

An ounce of gold is still an ounce of gold, but the demand for ounces of gold has changed drastically in the last 50 years, while supply from mines is only about double what it was in 1974. The majority of gold is used for jewelry and electronics. It looks pretty, doesn't oxidize, and has excellent thermal and electrical conductivity. As the population of Earth has grown and technology has advanced, demand for jewelry and electronics has grown. Even more of the demand growth came from central banks and investors buying much more gold than they used to.

And why is gold special? You could arbitrarily select any other commodity that still has the same physical properties as 50 years ago and get very different results. Even for other precious metals. Since 1974, silver is up 5x, platinum is up 6x, palladium up 20x.

Macroeconomic inflation in the result of a variety of factors, only some of which are within the control or influence of governments. Aggregate demand, aggregate supply, money supply, velocity of money, net investment, etc. THEY would love to have precise control over inflation, but THEY don't possess that power.
Clearly macroeconomic inflation is much more complicated than looking at the price of gold in dollars. That was not my intent. However, gold is important because it was used to peg the dollar (i.e. the Gold Standard) up until the 1970s, and the numbers are what they are.

It’s probably more accurate to say the change in the price of gold ($35 to $2400) is “dollar devaluation”, not inflation. Even that is not exact because of supply, demand, mining efficiencies, etc., but it is more accurate than simply calling it “inflation”.

And the USG largely controls the value of the dollar by how much it pumps into the economy. I say largely” because again there are many more factors at play when it comes to pricing goods and services.

Suffice it to say today’s monetary policies are complicated. To be clear, I’m super happy we moved away from the Gold Standard, but in a world of fiat currencies it’s important to recognize the goal posts for currency valuations move.
 
These are just wild guesses from Farzad. But the basic idea is likely to be correct. That is, there will eventually be tiered pricing based on capability.

Given that tiered pricing is likely to happen, Tesla needs to clarify what you are buying if you purchase FSD outright for $12K. As of now, those of us who purchase or purchased FSD outright are not even entitled to hands-free driving. We are not actually entitled to autonomous operation at all.

Tesla needs to assure customers that if you purchased FSD outright then all future autonomous improvements are included.
 
Last edited:
Quick thought, this will further erode the value prop of the Foundation Series CT, going to see some unhappy people there that already didnt like they were "paying" more than the locked in $7k FDS price for early res holders.

Cyberbeast Foundation option package is ~20k if you don't currently have Powerwalls, utilize the full $4k of installation credit, and value FSD at the $7k lock in price.

AWD Foundation, however, is even less reasonable than before, especially if one qualifies for the Clean Vehicle Credit and doesn't plan on using FSD a lot. Supercharging credits with an expiration date in lieu of Powershare equipment is also likely a net loss.

Right on red into the path of (possibly speeding) cross traffic isn't good though...

I have a really hard time believing:

1) That they don’t already have a ton of that feedback, and
2) That they are not already aware of 100 significant issues known to need work.

Take a ton of feedback, multiply by 1/10, then again, then again...
Take 100 issues, multiply by 10, then again, then again...

The march of 9 requires collection growth to maintain rate of improvement. It also requires as wide a collection set as possible to find the edge cases that were not yet handled.
 
Take some drives away from California or wherever you are. We are in no way on the march of 9s yet. Trust me I want to be. And we're a lot closer now than we were a few months ago. But we're not there yet.
I agree. I just had it do a couple dangerous things and had to turn it off. Scared my wife. maybe it’s a lot different in California but at least for me I don’t see how it would be worth $100 per month.
 
These are just wild guesses from Farzad. But the basic idea is likely to be correct. That is, there will eventually be tiered pricing based on capability.

Given that tiered pricing is likely to happen, Tesla needs to clarify what you are buying if you purchase FSD outright for $12K. As of now, those of us who purchase or purchased FSD outright are not even entitled to hands-free driving. We are not actually entitled to autonomous operation at all.

Tesla needs to assure customers that if you purchased FSD outright then all future autonomous improvements are included.
BTW, the relevant text on Tesla's FSD page looks like this:
The currently enabled Autopilot, Enhanced Autopilot and Full Self-Driving features require active driver supervision and do not make the vehicle autonomous. Full autonomy will be dependent on achieving reliability far in excess of human drivers as demonstrated by billions of miles of experience, as well as regulatory approval, which may take longer in some jurisdictions. As Tesla’s Autopilot, Enhanced Autopilot and Full Self-Driving capabilities evolve, your vehicle will be continuously upgraded through over-the-air software updates.
It sort of suggests that those who purchased FSD outright might get full autonomy eventually. But I can't find anywhere it says that when full autonomy is achieved, then full autonomy will actually be included with your purchase.

I just hope Tesla will clarify this.
 
Who said zero?

Who said 10x better than human?

Right now, it's about 100x worse than human, and I'm inferring based on some data on how these models scale that getting to human equivalent may take more than 100x in data and model size. Compute too, but part of that can be solved by training for a longer period of time.
Now it’s clear you have an agenda.
 
Quick thought, this will further erode the value prop of the Foundation Series CT, going to see some unhappy people there that already didnt like they were "paying" more than the locked in $7k FDS price for early res holders.
you are paying $20K to get it this year and to be one of the first CT owners in NJ ;) and we still don't know the long term value of FSD
 
Now it’s clear you have an agenda.
has been clear all along
1713013936871.png
 
I think it's much more likely FSD gets divided into two categories in the near future:

1) Consumer FSD
2) Commercial FSD

#1 would be for a Tesla owner who just uses it for their own car, no RT fleet use, and #2 would be for the RT's, people using FSD to generate income from ride hailing.
Bingo. That’s the only way it makes sense in my mind. Only a small portion of the population can afford paying several times a cable/internet bill for something they use a few hours a day at most.
 
And there is that TSLA Bull Greed again. 😥

I don't see this happening at all, nor do I think its feasible. At those prices most Tesla owners would never pay for FSD L3, L4 or L5 as its just too pricey. Plus it would make the entire FSD situation overly complicated, and Tesla as a company loves to simplify and avoid complications.

I think it's much more likely FSD gets divided into two categories in the near future:

1) Consumer FSD
2) Commercial FSD

#1 would be for a Tesla owner who just uses it for their own car, no RT fleet use, and #2 would be for the RT's, people using FSD to generate income from ride hailing.
And there is that typical human negative survival instinct brain activity unchecked and unawares.

To be clear, I don’t know if either of you are right or wrong. What I do know is that not having an opinion one has massaged and manipulated into a particular verdict allows one to walk through life with openness for any possibility, to ‘see’, and then to simply accept all for what it is without emotional drama or trauma. One also becomes a whole lot less wrong.

Yes, of course I understand the purpose of this thread. Do you? Speculation has gotten many people into financial discomfort and outright trouble. It’s also left them on the shore with the boat far out to sea.

I picked your post by random luck of the draw. You certainly aren’t alone.
 
Take some drives away from California or wherever you are. We are in no way on the march of 9s yet. Trust me I want to be. And we're a lot closer now than we were a few months ago. But we're not there yet.
I agree I had my 2017 MS with 12.3.4 drive me to my colonoscopy ( :oops: ) and it did intervention free and smooth as glass.....then got lost in the parking lot one business over from Dr. office.
 
OT OT OT
@DarkandStormy

What if my life was a series of imagined crises that never happen, but paralyzed you into inaction?

What if it isn't, and you keep moving forward, with ideas actualized that change the world for the better?

10.5 years example below, same driving force, expanding
(Gwen Shotwell is a true GODDESS)

may I remind you of
Monday, October 13th, 2013, 1st flight, and landing, of the Grasshopper
"Grasshopper is a 10-story Vertical Takeoff Vertical Landing (VTVL) ....,Grasshopper consists of a Falcon 9 rocket first stage tank, Merlin 1D engine, four steel and aluminum landing legs with hydraulic dampers, and a steel support structure." went up 744 meters
(read the comments over the years for illustrative insights over the years from 10 years ago and sooner)


Then a few short weeks ago
(BTW, the Washington monument is 169 meters tall, starship is 150 meters)
(in 10 years)
Thursday, March 24th, 2024, 3rd flight of the BFR/Starship/

"But that's not _Tesla_!!" <== true that, what's the driving force?,
a bit flawed perhaps, definitely not 100% correct, but successive approximations close enough

So, as for Tesla, Tony Seba posts 2 photos of New York City Easter parade, ~13 years apart of the switchover from horses to cars and the speed of the transition,
I would suggest the student go and review his work and perhaps revise your thought processes.

for wont of a better thought, perhaps we are in the midst of "punctuated evolution/distributed mini-singularities," others can be more precise

Those who can, do, those who can't, or won't, kvetch and kibbitz
Thanks for the great reminder even though you’re trying to talk sense and logic to someone who’s here for purely dishonest reasons. Still, someone else may learn from your post making it not a wasted effort.
 
Sooooo, I can't break 12.3.4. I had several things that 11.x either could barely do or not all. 12.3.4 has no issue that I can detect after nearly 200+ miles and I'll drive more, but this is sublime.

And I've tried, really really tried to break it. Throwing all kinds of traffic and complexities. It's been nearly a decade, to the month, when I started the team down this autonomous path and now, for the very first time, I'm left without words as my drives are without any reason to have a human behind the seat. Yes, there are corner cases and extraordinary circumstances which will be prioritized, but this is trivial by comparison.

Folks, ladies and gentlemen, members of the jury, we are on the march of 9s, at long last.

Expect 100s, 1000s, weeks and months, of 'human out of loop' driving in your vehicles.

I'm so proud of my former team.

So when we get earnings and they suck and the SP drops, I'll be buying calls/leaps as this is the beginning of the end. NFA😁
What year and model is your vehicle, and what hardware are you running?

I got 12.3.4 yesterday on 2020 LR Y with hw 3.0; did a drive yesterday and this morning. I do the same "challenging" routes to compare/contrast software versions. Can't say I notice any improvements from 12.3.3.

The car somehow got a school zone (15 mph) incorrect and now shows 40 mph speed limit, although at least it still drives at a speed that won't garner a traffic ticket.

Automatic speed offset is still too fast, will go 38 mph on a 25 mph road, for example. I think auto speed offset is a good idea in principle, but for now I feel that whenever the vehicle is in FSD it shouldn't go 5 mph over the posted speed limit. When it drives fast on a slow road it tends to take corners too fast and brake suddenly, causing excessive wear and tear on tires, brakes, suspension, etc. I realize this is a very subjective statement, but I want to maximize vehicle/tire longevity, not waste energy, and not waste money. Also, one would never drive this way with a stranger in their car, or their boss, or their elderly mother, etc.

It still messes up one intersection quite badly--I was really hoping it would be fixed because I give feedback every single time I drive this section, for years now--a right turn onto an in-town 40mph road (from a 25 mph road with a school), but the car still recognizes it as a 55mph highway after all these years. So the car just rips out of the corner, close to breaking traction, then will speed up to 55mph then suddenly back down to 40 when it comes up to the 40 mph traffic sign. Just a dangerous, illegal, hard-on-the-car, waste-of-money maneuver. Really wish Tesla could get the speed limit fixed here. I know they've got the ability to fix this.
 
I just downloaded FSD V12.4 and did my test loop a couple of times. My overall verdict is that V12.4 is definitely improved over V12.3. This was about one hour of driving with no disengagements and just one accelerator push to keep from annoying the person behind me.

There is a particular left turn lane that V12 refused to enter until now. It was really strange behavior. It was almost as if FSD thought that lane was not driveable space. With V12.4, it's a little hesitant, but it does get into the lane and complete the left turn. This was the one bad failure on earlier versions and now it's at least acceptable behavior with V12.4.

I'm quite happy with this result. It shows me that improvements are coming fast. It's getting to the point where I hate driving myself. I'd rather let FSD do it for me.
 
As a side note, my drive produced a little bit of evidence that the highway stack is still on V11. About 10% of my test loop is on interstate highway. As soon as it got on the highway, the wipers came on even though it's a sunny day. I turned auto-wipers off, but then turned it back on as soon as I got back onto the surface streets. The auto-wipers never came on while on the surface streets.
 
Now it’s clear you have an agenda.

My "agenda" is to assess when Tesla's robotaxi approach might actually become viable, and make investments based on that assessment.

What is wrong with those statements? Do you not believe it is currently 100x worse than a human? Does that statement feel aggressive to you?

The current software has a critical disengagement once every 300 miles. If we assume only 1/3 of critical disengagements would cause an accident, you are still at 1000 miles per accident, which is still more than 100x worse than human drivers.

And as for how much data, compute, and model size increases might be needed to overcome that 100x, I think I provided some research supporting it.

So how about attacking the material, not me?
 
Will geofencing reduce the size escalation rate? With everything I ever did, too long ago to be relevant, reducing the scope of any model, by defining and much as can be defined, reduces the model complexity. Although in AI when I was doing it, that was very primitive as was LISP, our only usable language at the time.

So, is that still true. Could Tesla simply use pre defined elements, and by geofencing reduce the unknowns in many cases. Of course those compromises might not matter were Robotaxi operate, as logic suggest it will, in specified geographical limits for each trip. That would make distinct version substitutions in every major market. They already do that with maps, although now in quite huge continental-sized chunks.

I believe geofencing can definitely reduce the error rate and thus reduce the need for how big the models would have to get.
 
  • Like
  • Informative
Reactions: primedive and Usain