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

Off topic galore

This site may earn commission on affiliate links.
1657792687309.png
 
  • Like
Reactions: JimS
Responding to main
kaparthy is not needed.
Agree, in the same way foundation and framers are not needed once a building is dried in. Additionally, there are others out (and in) there, if needed.

Raw compute is not enough.
Raw compute can beat any algorithm.
Wait, which are you advocating? Local minima in approach cannot be overcome with more compute.

Most important are excellent annotated data.
Garbage In, Garbage Out, but without a correct processing solution in the middle, even infinite perfect data on infinte compute won't achieve the goal.
 
  • Like
Reactions: navguy12
Wait, which are you advocating? Local minima in approach cannot be overcome with more compute.
Both.
It depends.
I the early years at university computer vision was very much SIFT/SURF and other "clever" techniques. Then came CNN and raw compute beat the *sugar* out of those things.
On the other hand: If you have a process that is not doomed in a local minimum, clever algorithms are the main thing & focus to optimize.
There was a paper from Dijkstra(?) or someone similar where they made a case-study. All "optimization" tricks like "closer to the metal", "rewrite in C++" etc. lost massively against a simple "change tiny sub-algorithm X for Y" that does the same, but different.

But it is a big, big Pareto-Front .. with dozens of compromises/caveats/...

So a simple "Just build dojo and pump money into it and FSD will be solved" is as wrong as "don't develop the software-architecture further, just do bugfixes. Problem is solved"...

I mean .. fastest way to search using comparisons is O(n*log(n)). Actual fastest way to sort lists < 1000 is Insertion-Sort with O(n²).. Actual "fastest" way to sort is to cheat and transfer it to another domain (i.e. radix sort, which is O(n*m), with m << n)... :D
 
View attachment 829789

TL: DR

You led Mary...You led

View attachment 829790

This is so sad! My family has always been a Opel/GM family - only my weird uncle bought a Ford (Escort). So it's in my blood to cheer for GM. Growing up our dream was to one day buy an Opel Senator - the euro version of a Cadillac.

But now I want what Mary Barra is smoking. Because her management of GM is really far out there! She is running the company into the ground while she herself is in a total bliss. Just as well GM sold Opel to Stellantis because this is not going to end well!

GM might survive as a boutique car maker with low volume production of a few iconic models. That is their best hope. Unless the US taxpayers decide to throw a space budget their way. But why would they?
 
  • Like
Reactions: M|S|M MYP
🤨

So I have it upside down and the media are correct.

Tesla is working to make cells and batteries. Ford/ GM making batteries.

Thanks I'll try to be better.
FWIW, I'd greatly prefer them to say "assembling battery packs" than "manufacturing batteries". Heck, even "assembling batteries" is more precise.
And don't get me started on "recall".


Communicating
communicating_2x.png
 
  • Funny
Reactions: navguy12
@The Accountant, I'm kidding, you have nicer charts.
Question since I'm here, how do warranty claims impact service revenue vs out of warranty?

Bitcoin Simplified

It's easy to get tied up in knots trying to figure out where Tesla stands on bitcoin because of the silly impairment charges clouding up true cash impact.
Here is a simple way to look at it.

Tesla spent $1.5b buying Bitcoin (cash out) and sold twice for $272m and $936m (cash in) leaving them $292m in bitcoin from their original investment.
As of July 25 (today) their remaining Bitcoin is worth $265m so they are underwater by $27m.
Tesla reaches breakeven on their original $1.5b outlay if Bitcoin reaches about $26k per coin.

View attachment 832527

The full history of Bitcoin at Tesla
Q1 : purch $1,500M, sold $262, book asset: $1,331
Q2 : book $1,311
Q3-Q4 : book $1,260
Q1 : book $1,261 (Doge?)
Q2 : sold $936 book $218
Total sold: $1,198
Current book asset value 218/17,708*23,200 = $286
Total value: $1,483M
Currently down $17 million. Breakeven at $24,531
 
  • Funny
Reactions: The Accountant
@The Accountant, I'm kidding, you have nicer charts.
Question since I'm here, how do warranty claims impact service revenue vs out of warranty?





Warranty claims have no impact on service revenue. When Tesla sells a vehicles, it puts about 2% of the sales price into a warranty accrual (it is part of Auto Cost of Goods Sold). When the warranty claim is serviced, they charge this warranty accrual instead of expensing it on the P&L.
In the 2022 column below, the $319m is the warranty accrual (money set aside for the vehicles sales in Q2 2022). The -$187m are the warranty costs spent in Q2 for cars sold in the current and prior periods.

For Non-warranty (out of warranty): The invoice the customer pays goes to Service Revenue and the cost of that non-warranty service goes to Service Cost.


1658772476034.png
 
Warranty claims have no impact on service revenue. When Tesla sells a vehicles, it puts about 2% of the sales price into a warranty accrual (it is part of Auto Cost of Goods Sold). When the warranty claim is serviced, they charge this warranty accrual instead of expensing it on the P&L.
In the 2022 column below, the $319m is the warranty accrual (money set aside for the vehicles sales in Q2 2022). The -$187m are the warranty costs spent in Q2 for cars sold in the current and prior periods.

For Non-warranty (out of warranty): The invoice the customer pays goes to Service Revenue and the cost of that non-warranty service goes to Service Cost.


View attachment 832624
So on a warranty issue, the warranty tracker gets a debit with the labor and parts costs. I follow that. Where I'm fuzzy is how the labor and parts (inventory?) get adjusted. Parts, I assume, live under the automotive category, so that stays net neutral. Does service revenue go up based on the labor hours such that the serivce and other (S&O) category is compensated? Or does the warranty labor get removed from cost of S&O? Same overall effect by a different route.

Trying to understand warranty vs non-warranty work in the context of S&O Gross Margin/ profit.
Thanks!
 
So on a warranty issue, the warranty tracker gets a debit with the labor and parts costs. I follow that. Where I'm fuzzy is how the labor and parts (inventory?) get adjusted. Parts, I assume, live under the automotive category, so that stays net neutral. Does service revenue go up based on the labor hours such that the serivce and other (S&O) category is compensated? Or does the warranty labor get removed from cost of S&O? Same overall effect by a different route.

Trying to understand warranty vs non-warranty work in the context of S&O Gross Margin/ profit.
Thanks!
Yes - the warranty labor gets removed from the S&O costs. My guess is that all service labor at first goes to the Service Cost center for the month and then there is an entry based on warranty hours worked that takes to costs and moves it out from Service cost to the warranty accrual.
 
Yes - the warranty labor gets removed from the S&O costs. My guess is that all service labor at first goes to the Service Cost center for the month and then there is an entry based on warranty hours worked that takes to costs and moves it out from Service cost to the warranty accrual.
Ahhhhh!!
Charge codes/ cost centers, of course.

Thanks!
 
  • Like
Reactions: UltradoomY