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

[NicoV]

Note that the primary factor creating the fork is the two vastly different versions of the hardware: HW3 is 10x-20x faster than HW2 and can load much larger networks as well.

This is a physical fact that cannot be designed away, and splitting functionality out that can only run on the newer hardware is standard industry practice.

For shared functionality it's possible to keep a largely unified "codebase" by downscaling the HW3 networks to fit into HW2's computing constraints - which is a mostly automatic step.

The obvious explanation of why Tesla tried to max out HW2 for so long is to:

1. implement EAP and basic Autopilot functionality on a HW2 basis,
2. leverage the combined HW2+HW3 fleet to train HW3-destined networks as long as possible,
3. to wait for the HW3 fleet to become almost as large as the HW2 fleet,
4. to wait for a comparatively "quiet" January-March period when North American and European service centers have excess capacity to perform the mass-retrofits.

5. To delay the cost of the upgrade as long as possible (but not longer).

 

[Artful Dodger]

'Knock knock.'

Who’s there?

'Margin calling.'

Margin calling who?

'Margin calling Gollum.'

Oh, Sméagol!​

FTFY.

 

[MC3OZ]

Note that the primary factor creating the fork is the two vastly different versions of the hardware: HW3 is 10x-20x faster than HW2 and can load much larger networks as well.

This is a physical fact that cannot be designed away, and splitting functionality out that can only run on the newer hardware is standard industry practice.

For shared functionality it's possible to keep a largely unified "codebase" by downscaling the HW3 networks to fit into HW2's computing constraints - which is a mostly automatic step.

The obvious explanation of why Tesla tried to max out HW2 for so long is to:

1. implement EAP and basic Autopilot functionality on a HW2 basis,
2. leverage the combined HW2+HW3 fleet to train HW3-destined networks as long as possible,
3. to wait for the HW3 fleet to become almost as large as the HW2 fleet,
4. to wait for a comparatively "quiet" January-March period when North American and European service centers have excess capacity to perform the mass-retrofits.

With the shared code base, typically it is possible to do the following...

Have one version of the software which downloads to all cars, but detects which processor it is running on, and some functions are simply turned off on HW2...

My understanding is that the NN is encoded in the software, but the HW2 processor may not need to use all parts of the NN.

Again this is simply educated guess work...

 

[verygreen]

Much of EAP is completed already

check in the nearest Tesla car, it says "BETA". That means it's not complete. That in turn means they need to keep developing it until it's complete.

If you read through that list:

• match speed to traffic conditions: done
• keep within a lane: done
• automatically change lanes and transition from one freeway to another: done
• be summoned to and from your garage: done
• self-park when near a parking spot: "Smart Parking" feature promised but not yet done.

All beta. some more stable than the others none are perfect. FSD needs them to be perfect => shared development => why wasting time doing parallel development of the same stuff independently?

 

[wipster]

Yes, he has his NLP and computer vision bots Fact Checking and KarenRei processing this thread and immediately messaging him of any interesting posts.

Speaking of KarenRei, her absence is noticeable on 420 day... or did I miss her postings?

 

[keydiver]

This news of a 60+ car pileup was posted earlier here. My wife said “hey there is probably a Tesla in there somewhere”, and noticed that the vehicle to the far right being loaded on the tow truck sure looks like a Tesla Model 3. Anyone agree/disagree?

Police: 69-vehicle pileup in Virginia leaves dozens injured

View attachment 492126

I was going to suggest the same thing when I saw the picture earlier. Ouch!

 

[Chop110]

Just got back from an all day meeting to see my Tesla shares are at 419. I missed the 420 party!!! Looks like the next one will happen at the crack of dawn in the premarket. My wife is going to get up and ask me why I am making brownies for breakfast.

 

[verygreen]

With all due respect, it is asinine to asusme there isn't a biggert NNet out there for HW3.0 FSD simply because you haven't seen or heard of one.

We deal in observable evidence here. No super advanced NNs were observed. Not even at the April demo day where almost everything demonstrated except some visualizations was possible to replicate with hw2.5 car on customer firmware with some small limitations.

A batch of Tesla employees have been testing HW 3.0 FSD for a year! Do you honestly think that has been done with the same NNet that is on any consumer car right now, let alone ones with HW 2.5?

So.. show me the goods? Perhaps if Tesla is so amazingly forward they'd do a demo? You know they were not shy in Q3 2016, so why so shy now?

There is no benefit to putting it on a consumer HW 3.0 car to just run effectively 2.0 EAP. I mean, Tesla is getting crap for not upgrading cars to HW 3.0 and they haven't even deployed the larger NN and its updated features yet. Imagine if HW 3.0 cars already had the "better' NNet? People would be pissed!

Based on my direct observation currently up to 19.40.2.1 firmware hw3 cars behave identical to hw2.5. only exception is cone display

 

[avoigt]

About a year ago I posted a longer analysis concluding that in 2019 given all the positives in the pipeline a new ATH will be seen after years of sideway moves.

Within 2019 I lost faith in my own prediction not because all the positive news did not happen, actually they did happen but we all remember well the Stock fell well blow $200 to a level where I was forced to buy again because it was simply said a bargain price given what Tesla had achieved. It fell despite of all good news with was nuts. Shorts and Bears have been clearly in control.

Now the SP did move in the direction of the truly value of Tesla while many more news are still pending as well as the market still to realize this is technology stock and not just automotive.

The daily increases in the morning hours and the almost flat behavior after, points for me to a squeeze in daily doses but still a squeeze knowing that if we trust Ighors analysis only a smaller portion of shorts did really cover.

The market finally realized that Tesla has been wrongly valued in 2019 and tries to anticipate a better valuation for 2020 but I don't hear people really understanding what they have in front of them.

Having said that, this stock is far away from any sort of over evaluation and although I never will exclude that we dive again one day maybe to test $400 or $380 again all signs point to a better sentiment in 2020 and more higher highs.

Right now there is an upwards dynamic that will find a stop somewhere but we are in not known territory therefore no one knows and frankly said I don't care anyhow simply because I don't intend to sell a single share for the longer future so why should I bother.

I've seen people selling today because of the number 420 which makes me speechless. An old rule is if as stock has a run get a cool beer and enjoy. Greed is in the way of many people you don't have patience.

We all should keep a cool head and don't overestimate the situation. There is still a lot of in front of Tesla to work through and nothing will happen without hard work.

Right now the market is ignoring negative news like people are leaving or the china credit which short time before has been weighting in but today no one cares. Obviously I like that but be prepared that can change again and will.

Anyway enjoy the run and X-mas.

 

[verygreen]

For shared functionality it's possible to keep a largely unified "codebase" by downscaling the HW3 networks to fit into HW2's computing constraints - which is a mostly automatic step.

if they did that that would still be the same codebase (I am not sure you can actually scale down NN like that)

there's another sensible thing to do - you may chose to not run some of the NNs on some platforms and run all the same NNs when the functionality is shared and that would still be a plus wrt develipment time and avoiding duplication of effort and testing. And they might do it eventually but so far it was hardly observed

 

[uselesslogin]

This news of a 60+ car pileup was posted earlier here. My wife said “hey there is probably a Tesla in there somewhere”, and noticed that the vehicle to the far right being loaded on the tow truck sure looks like a Tesla Model 3. Anyone agree/disagree?

Police: 69-vehicle pileup in Virginia leaves dozens injured

View attachment 492126

TSLAQ was actually talking about this and they concluded it is a Camry.

 

[ggr]

This news of a 60+ car pileup was posted earlier here. My wife said “hey there is probably a Tesla in there somewhere”, and noticed that the vehicle to the far right being loaded on the tow truck sure looks like a Tesla Model 3. Anyone agree/disagree?

Police: 69-vehicle pileup in Virginia leaves dozens injured

View attachment 492126

It's also the only one pointed the wrong way.

Old joke:
Wife calls husband who's driving home from work. "Honey, watch out, there's someone driving the wrong way on the 405!"
Husband: "Not just one, there are lots of them!"

 

[Leo9]

Just an FYI on this guy's posting history over the last nine months.

Not sure what's your point. I'm a very short term option trader and I was basically right in every istance you quoted but one. Did you check the chart?
That doesn't mean I'm not bullish on TSLA as a long term investment. I'm just commenting on the very short term price action.
If you find that funny, that's fine.

 

[pz1975]

TSLAQ was actually talking about this and they concluded it is a Camry.

Too bad for them. If it had been a Tesla, then clearly fraud...bankwuptcy...Teslemmings...TSLAQ... and the stock would be under $100 tomorrow.

 

[jhm]

Where is the price going next? I don't know, but here's a baseless conjecture.

Tesla tends to trade in a wide interval for several years before breaking into the next interval.

First, $20 to $36.
Second, $190 to $270
Third, $270 to $370
Fourth, $370 to ???

There was a transition from first interval to second. Likewise from the third to fourth, we've had a dal segno to $180 and back. Now we appear to be headed into the fourth interval.

So let's suppose that a certain interval will prevail over the next several years. What would be that range. Here's my totally baseless guess.

Fourth, $370 to $520.

I think there is a tremendous amount of attention on the $420, but that is too close to $370 to hold as an upper limit. We may get hung up at $420 for a while, but the real hype train needs much more range to roam. $470 could be a upper limit, but on a relative scale the range from $370 to $470 is really not that compelling. Musk's award is linked to a market cap of $100B or about $554/sh. So that will be a major battle with shorts. So I think $570 is too high for the fourth interval. So spitting the difference $520 seems about right and has some symbolic resonance with $420.

So there you go. Look for Tesla to be range bound between $370 and $520 for several years. The exciting part is that we could see the high $520 marked very quickly. Perhaps within 6 months.

Enjoy the ride!

Ok, I appreciate everyone's feed back on this. There were two sorts of criticism:
Some think $520 cap is too low for a couple of years growth.
Others think that Tesla will start to grow exponentially rather than in intervals that are fixed for a few years.

I certainly believe that the value of Tesla has been growing exponentially for quite some time and will continue. But how the market trades rarely intersects with value growth. I'd love to see it trade exponential in short horizons, but I still think there is potential for fixed interval trading. Both are exponential in the long run, but the basic difference is this. As Tesla transitions out of an interval, that can attract momentum traders. Moreover, longs start to buying into the momentum play as well. Before you know it the price can surge really high a few years before that price can become a firm floor. Then you go through a couple of cycles where shorts talk the price down followed by recovery. But the recoveries tend not to strike a new ATH well above the last ATH. So this is what keeps it range bound, which in turn causes sentiment to sour as it looks like that stock just can't get anywhere. So while the whole market is heating on Tesla, the underlying value continues to grow exponentially. Eventually, the stock is severely undervalued. The market realizes this, and we surge on up to the next interval.

So I still think that the stock is vulnerable to this sort of trading, but I do take the point that $520 may be a little too low to cap Tesla for a few years. We still have a sticky point near $555, so maybe $570 is the point where the thinks it may be getting ahead of itself. Imagine if we hit $570 in April 2020. Would we really be thinking that it can keep sprinting from there? Or will start to wonder if there is no more upside left for the next 6 months? I'm inclined to think in this situation we would tend to feel that in the short run there is more downside than upside. So that sets the ATH that will be hard to beat for the rest if the year.

So my revised interval is $370 to $570 through 2021.

(Of course, I do hope it is better than that in 2021.)

Merry Cybermas!

 

[ggr]

Just got back from an all day meeting to see my Tesla shares are at 419. I missed the 420 party!!! Looks like the next one will happen at the crack of dawn in the premarket. My wife is going to get up and ask me why I am making brownies for breakfast.

It's possible that with a gap-up opening, there will never be another chance. Good for me that I don't care.

 

[HG Wells]

check in the nearest Tesla car, it says "BETA". That means it's not complete. That in turn means they need to keep developing it until it's complete.



All beta. some more stable than the others none are perfect. FSD needs them to be perfect => shared development => why wasting time doing parallel development of the same stuff independently?

Gee, we always had two or more teams developing software in competition with each other. Sharing code was not allowed.
It often started as 3-4 teams and would shrink to two after six months. Early 1980's, everything written in assembler. Eight inch floppies (180K).
Good times.

 

[M3Rider]

Btw, have we heard what $1.5B Chinese loan was intended for? That's a lot of money, isn't it? They could probably build GF5 with it?

 

[Right_Said_Fred]

This news of a 60+ car pileup was posted earlier here. My wife said “hey there is probably a Tesla in there somewhere”, and noticed that the vehicle to the far right being loaded on the tow truck sure looks like a Tesla Model 3. Anyone agree/disagree?

Police: 69-vehicle pileup in Virginia leaves dozens

It’s not a Tesla

 

[winfield100]

MOD:

Now, where is that rusty razor I keep for Emergency Purposes?

@AudubonB
You keep it with the drunken sailor, in the lifeboat