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

[vikings123]

In the process of ordering a new LR+ S, I noticed that a month or two ago it was 4-6 weeks for delivery, last week it was 3-5 weeks, and now 2-4 weeks. Not sure if there is anything to read into that, or if typical EOQ push

I see at least 11 inventory model S cars in my local area, no additional incentives. I’m assuming this is to be expected but also want to say I have not seen this much inventory in a while, it’s usually 2-3 cars.

I realize I cannot make any inferences from this one data point but are there any indications that model S demand is lower?

 

[HG Wells]

Interesting report on factories restarting in China.

The Impact of Coronavirus on China’s SMEs: Findings from the Enterprise Survey for Innovation and Entrepreneurship in China
SEE
The Impact of Coronavirus on China’s SMEs: Findings from the Enterprise Survey for Innovation and Entrepreneurship in China

 

[Tslynk67]

You are presenting a false equivalence fallacy here. Pierre, has a university degree in computer science, but worked his life as an auto analyst... he has little to no exposure to cutting edge AI. Then there is Lex Fridman, who is an enthusiastic albeit weak AI researcher, with a bunch of biases towards his pet research projects like driver monitoring. Then there is Jim Keller, one of the best chip designers who has worked on the most cutting edge FSD systems and teams at Tesla. And then there is the Tesla team with Karpathy at the top of the FSD field - who informs Elon's estimates about the field. And we have Elon giving us information about FSD feature complete very soon.

Who are you going to believe? The best team and most sought after team in the world including Karpathy, Jim Keller, Elon and the huge amount of other talent they have hired, or Pierre with his idea that FSD won't even be able to do a 20 min drive? Don't you understand that there is a massive gulf of ability/talent/raw intellect that separates people like Pierre from people like Elon/Karpathy/Keller? (With people like Fridman somewhere in the middle)

Also I don't think his methodology was done well, simply because he thinks the probability of FSD is 0%. FSD is the single most important upside driver of Tesla, and to miss the main aspect of Tesla's future business is really poor imo. He is directly contradicting what Tesla expects to happen. Plus his idea of energy being 30billion vs the auto side being 5x that, also misses the point about the direction Elon intends to take the company. When Elon said energy will be the same as auto, this is because he intends to massively scale up battery/storage over the next 5 years. After battery day, and once Tesla starts producing their own batteries, I think analysts will have to start taking Tesla energy more seriously. For the most part, analysts are very over-conservative with their predictions, and essentially backwards looking. They are trailing the technology, not predicting it. Just because he is more bullish than almost all the analysts, doesn't mean he is a good analyst or correct.

On a related note, one of the aspects of chess I found most interesting was the ranking system. These ranking systems really put in place how vast the difference in skill and raw talent is between people. It is very difficult to beat someone with more than 100-200 points higher. These heavily right skewed distributions of ability will exist in every field. If you want to learn the most about a field, just work out who is the best in that field and listen to them. If you listen to people weaker than those at the top of their fields, you will be misinformed.

My point was that you stated you lost all respect for him simply on the fact that his opinion is that he doesn't believe in FSD, which I think is a very unfair stance - excluding their stance on FSD, the rest of their assumptions were very good and seemed quite comprehensive.

For what it's worth, I think he's wrong too, I do think that the vast majority of drives will be fine under Level 5, but Pierre was referring to edge-cases where human intervention - the "intuition" aspect or philosophical situations is needed to intervene to solve. In this regard I think he has a point and there will be scenarios where the AI will not know the correct choice. Now the escape-route in these situations isn't known right now, I could imagine a battery of remote drivers that could take over control to negotiate the deadlock for slow-evolving situations, could be a way out, after which presumable the fleet will have learned something to help avoid it the next time. Obviously that won't fly for real-time scenarios, like whether the car should swerve away from the old man crossing the road, but then run-down a child instead.

Note he also said that when he sees more evidence of FSD he will revise his model. I think that's fair, no?

High ELO rated chess players can still be beaten by lower ones, I know this because I've done it, so it's not a given.

Anyway, don't want to pick arguments.

 

[jerry33]

For what it's worth, I think he's wrong too, I do think that the vast majority of drives will be fine under Level 5, but Pierre was referring to edge-cases where human intervention - the "intuition" aspect or philosophical situations is needed to intervene to solve. In this regard I think he has a point and there will be scenarios where the AI will not know the correct choice.

I'd say there are cases where the human driver doesn't perform the correct choice, reacts to late, drives too aggressively, or just plain misses noticing there's an issue. I'd further speculate that the odds are better for AI to get it right than for the human--Not in today's AP, and possibly not in the first public release of FSD, but very likely in the second release.

 

[anthonyj]

Anyone else see that consistent increase in volatility starting in end of 2016...

...when Trump was elected into office compared to Obama's run as POTUS?

That’s because a 15% drop at S&P $1800 doesn’t look as steep as a 15% drop at $3400. Market was always plenty volatile

 

[goinfraftw]

That’s because a 15% drop at S&P $1800 doesn’t look as steep as a 15% drop at $3400. Market was always plenty volatile

...wouldn't a higher market-capitalization for an instrument like the S&P 500 (which is an index of ...500 top-performing companies) suggest better stability in that asset (if we're talking about a stable world/US environment where it trades in)?

 

[KarenRei]

Interesting report on factories restarting in China.

The Impact of Coronavirus on China’s SMEs: Findings from the Enterprise Survey for Innovation and Entrepreneurship in China
SEE
The Impact of Coronavirus on China’s SMEs: Findings from the Enterprise Survey for Innovation and Entrepreneurship in China

I'm confused by that article. They talk about companies not having resumed work by 10 February, but that was during the Lunar New Year holiday (which was extended for a week). And 10 February is three weeks ago.

 

[Fact Checking]

I'm confused by that article. They talk about companies not having resumed work by 10 February, but that was during the Lunar New Year holiday. And is three weeks ago.

I found that odd too: February 10 was a Monday, the first day of work resumption in select industries.

 

[lafrisbee]

I watched the whole thing and found his viewpoints on AI and the solvability to be reasonable. Is he right or wrong, I don't know, I'm no expert either, hell according to Jim Keller, it's just "ballistics", which Lex Fridman doesn't agree with at all - Pierre is in this latter camp.

To "lose all respect" because you didn't agree on this point of view seems very harsh to me, especially when the methodology, research and reasoning they put into their Tesla price targets appears to be extremely well done.

See if this could be correct...
First it isn't about whether it can be done or not.
The only issue being discussed is, does the person uttering for or against our ability of develop FSD have the imagination to believe we can or we can't.
Those in the "no" camp can't give a valid reason why FSD can not be a reality. Their decision hinges on their imagination, as do those that think FSD can be accomplished. It is just where someone believes, not what they know. No concrete barrier or path has been found.
I'd have told you a good electric car was not something anyone could create to rival the ICE autos 20 years ago. I'd also have said I have no real reason... just an imagination.
I hope FSD occurs. I won't question it for very long if it does.

 

[anthonyj]

...wouldn't a higher market-capitalization for an instrument like the S&P 500 (which is an index of ...500 top-performing companies) suggest better stability in that asset (if we're talking about a stable world/US environment where it trades in)?

No. Markets have been irrational since the beginning of time. They are never efficient. The last two corrections were due to rapid rate hikes and Coronavirus which can happen under any administration. S&P Volatility actually reached its lowest level ever in late 2017

 

[goinfraftw]

No. Markets have been irrational since the beginning of time. They are never efficient. The last two corrections were due to rapid rate hikes and Coronavirus which can happen under any administration. S&P Volatility actually reached its lowest level ever in late 2017

"Do or do not, there is no try."

- Yoda

(FYI - its my effort in being hilarious)

 

[davecolene0606]

I found that odd too: February 10 was a Monday, the first day of work resumption in select industries.

Doesn’t seem to be prescient at all, especially being a pub date of 2/28. Seems as though the info is from an annual effort for purposes other than tracking a restart of business after an event such as is happening.

IMHO info gathered on 2/10 is of little use in present circumstances. If they would redo the survey today, now it would contain some useful data that one could base some forward thinking on.

Fire Away!
(It’s the batteries, Stupid!)

 

[NakedLong]

My point was that you stated you lost all respect for him simply on the fact that his opinion is that he doesn't believe in FSD, which I think is a very unfair stance - excluding their stance on FSD, the rest of their assumptions were very good and seemed quite comprehensive.

For what it's worth, I think he's wrong too, I do think that the vast majority of drives will be fine under Level 5, but Pierre was referring to edge-cases where human intervention - the "intuition" aspect or philosophical situations is needed to intervene to solve. In this regard I think he has a point and there will be scenarios where the AI will not know the correct choice. Now the escape-route in these situations isn't known right now, I could imagine a battery of remote drivers that could take over control to negotiate the deadlock for slow-evolving situations, could be a way out, after which presumable the fleet will have learned something to help avoid it the next time. Obviously that won't fly for real-time scenarios, like whether the car should swerve away from the old man crossing the road, but then run-down a child instead.

Note he also said that when he sees more evidence of FSD he will revise his model. I think that's fair, no?

High ELO rated chess players can still be beaten by lower ones, I know this because I've done it, so it's not a given.

Anyway, don't want to pick arguments.

Don't see it as an argument so much as a constructive debate - I have seen your other posts and learnt from them...

I think by watching AlphaZero you can tell that weak AI can have massively better intuition than humans in the realm of chess... you can extrapolate this to self-driving. Tesla's FSD will be exactly what AlphaZero is to chess - in 100 games vs the human world champion, the human will not win a single game. Magnus Carlen quipped - 'I kept asking myself, what would AlphaZero do, then I realised I am not AlphaZero'. As Elon has said - FSD will appear to improve very rapidly, and then it will have superhuman driving ability. The same as what has happened in chess/go/many other areas.

Regarding this idea of remote human drivers, or humans having to resolve edge cases - I think this is not relevant. By the time level 5 FSD is good enough for wide spread adoption at 99.99999% reliability, how they handle failures in irrelevant, because they happen so rarely.

Regarding this artificially contrived example of an AI having to chose between an old man and a baby, implying a FSD computer will never be able to drive well - again this is an irrelevant point. How often does this situation come up in driving? Even if it came up a significant amount, would the AI choosing the wrong answer negate the accidents saved from the system? And even if this was an actual valid use case, do you really think an AI would not be able to be trained to kill the old man instead of the baby?

Regarding being able to beat higher rated players - of course you can beat them - that is how you climb the rankings. It's a question of the probability of beating them. If you take Magnus Carlsen's ELO of 2800ish vs AlphaZero's 3400ish rating - this 600+ ELO gap means he has <1% chance of winning. There are lots of tables and calculators for converting ELO to probability of winning out there.

Regarding whether Pierre's other info was correct besides his FSD analysis - yes except for energy. But like most bulls on TMC, he didn't really tell us anything we don't already know. I think it is telling that he came into analysing Tesla with a heavy bias against the company. It shows how out of touch the mainstream analysts are with new tech, and it shows he is susceptible to herd mentality, but not so much that he was unable to change his mind after further research.

 

[Sherlo]

I know this isn't the correct forum for this, but I feel the most knowledgeable posters on this subject are here. But, mods I understand if you need to move it. Essentially, my neighbor is telling me that scientists have proven that renewable energy is far worse for the environment than simply using oil. He provided this article as proof. I am very much a layman on the facts in this area and don't want to present a weak reply. Can you guys help me by pointing out why the ideas presented in this article are not correct? I would like to compile the rebuttals and present it - in one big slam dunk.

If You Want ‘Renewable Energy,’ Get Ready to Dig | Manhattan Institute

Good question, suggest starting here:

Solar, wind and nuclear have ‘amazingly low’ carbon footprints, study finds

Edit: source study cited in the above article can be accessed: Understanding future emissions from low-carbon power systems by integration of life-cycle assessment and integrated energy modelling - Nature : NuclearPower

Mod: Good reply to a question that should never have been asked here. No more replies. --ggr
Entire other sub-forum full of good rebuttal information: Energy, Environment, and Policy

 

[JBRR]

If you are itching to see what the futures market will open up at - there's a betting market that's open in the weekend.
Trade Weekend Wall Street | Live Weekend Wall Street Stock Index Price | Weekend Wall Street | IG UK

Should be reasonable indicator.

 

[gerebgraus]

I watched the whole thing and found his viewpoints on AI and the solvability to be reasonable. Is he right or wrong, I don't know, I'm no expert either, hell according to Jim Keller, it's just "ballistics", which Lex Fridman doesn't agree with at all - Pierre is in this latter camp.

To "lose all respect" because you didn't agree on this point of view seems very harsh to me, especially when the methodology, research and reasoning they put into their Tesla price targets appears to be extremely well done.

I am sorry to disagree but he was not in the Lex Fridman (Who is an expert, and pessimistic) camp. Ferragu had no clue what he was talking about. He was also worried at some point that Model Y would cannibalize Model 3 sales, and that is pretty silly. At this point, Model 3 +Y sales are limited by the battery supply, so if Model 3 is cannibalized by Model Y then Tesla will be forced to sell more of the more profitable Model Y. Too good for Tesla.

 

[JustMe]

Breaking news: corona sales plummeted in us. People said they think it might have something to do with the virus, but were not certain.

This was on a local newspaper here..

Good for us that it wasn’t called the “Tesla virus”

 

[Fact Checking]

Doesn’t seem to be prescient at all, especially being a pub date of 2/28. Seems as though the info is from an annual effort for purposes other than tracking a restart of business after an event such as is happening.

IMHO info gathered on 2/10 is of little use in present circumstances. If they would redo the survey today, now it would contain some useful data that one could base some forward thinking on.

Fire Away!
(It’s the batteries, Stupid!)

We have such a recent survey from China - got buried in the PMI news:

China's manufacturing PMI drops in February amid epidemic - Xinhua | English.news.cn

"As of Feb. 25, the work resumption rate in NBS surveyed large and mid-sized enterprises nationwide has reached 78.9 percent, and it is expected to climb to 90.8 percent by the end of March, according to the NBS."​

 

[generalenthu]

First let me say that MMs can and do abuse the rules and they deserve much greater scrutiny than the SEC has ever given. However, Options Market Makers do need to keep themselves near delta neutral, and buying and selling stock to achieve that is a bona fide market making activity.

In the case of selling puts, as the stock price drops the MM needs to become more delta negative because puts that they previously sold need to be kept delta neutral. In a balanced market with nearly equal puts and calls written by the MM it might be a wash where the delta shed from calls covers for the gain in delta from the puts, at least for small moves. TSLA has generally had a high put/call ratio such that option MM mostly needed to be net short the stock to remain delta neutral. The recent run-up in TSLA likely caused some people to insure their profits by buying puts (which you yourself suggested) and of course TSLAQ is always buying puts. So a MM may well have sold puts, in good faith and at reasonable prices weeks or months ago. Now the price is dropping fast and unless they want to go bankrupt they MUST short the stock (or sell stock if they have it) to remain delta neutral. They generally can't buy back the puts they sold - they are the ones making the market - so they must hedge instead.

I think it was @Papafox (or maybe it was @ReflexFunds ) who published tables of the "amplification effect" of option open interest on price moves - this could mostly be an example of that. The overall market melt-down did cause big moves which the MMs may have had no choice but to amplify in order to keep their positions neutral. That they are exempt from the uptick rule is legitimate in certain circumstances which may well have happened Thursday and Friday.

It's still possible that @Artful Dodger is right - for certain the MMs did do naked short selling, but depending on option open interest it might not have actually been illegal and improper. Digging further into the option open interest at the time to estimate what amount of short selling could have been necessary for the MMs to remain delta neutral is needed. Adding that analysis could prove or disprove if this was abusive or not. These days I'm so far behind on just reading the thread there's no way I can find the time to do the analysis myself, so all I can do is recommend that someone who can please give it a shot. I don't know how you and several others here are able to provide so much valuable work to this forum.

Here's the latest with a few days of historical info. Given the reinforcing effect of the options market, I was worried about larger drawdown, but we did pretty well during this sell off.

We're just back to prices as of a month ago, where as setback for the broader market seems to be 6-9 months.

There seems to be a combination of 2 things. Firstly there's good demand for stock at these prices. Secondly, which is probably more impactful in the short term is, put holders and call sellers are quick to take profits. Their theses have been utterly decimated. So they're supporting this market a bit better. This Corona thing ended up for them like a dog catching a car.

Longs on the other hand were hodling their calls earlier this year, leading to nonstop push higher. Will be interesting to see if this will continue as we recover. With so many catalysts, I believe so.

 

[hacer]

Look what we've found by comparing today's FINRA Short Selling Report with NASDAQ's 'per-minute' Trading log: the volume of 'naked short selling' today is almost exactly the same as an unusually large trade in the After-hrs session at 7:07 PM (EST).

This is compeling evidence of MMs covering their naked short sales from today, but look at the premium price they paid After-hrs: $674.00 vs. todays VWAP $646.79

They must have really needed those shares to spend $200M on them, hmm?
...

Consider the market making needs of the option writers Friday: The stock price ranged from $611 to $690, trimming that a bit to options in the $620 to $680 strike range assuming that outside that strike range both puts and calls were already fully hedged before the day's trading and delta didn't change much, we find (from Tesla, TSLA & the Investment World: the 2019-2020 Investors' Roundtable) that there were 15,516 open put contracts in that range, representing 1,551,600 shares of put hedging. On the call side, there were 3,228 open contracts in the strike range of $620 to $680, representing 322,800 shares of call hedging.

So -1,102,200 x (avg put delta) + 322,800 x (avg call delta) was dynamically hedged throughout the day. If we assume the same average delta for calls and puts (not a great idea but I can't think of a better approach) that means they would short 779,400 x delta shares. If the 301,960 shares exempt shorting represents those hedges, that would produce an average delta of 0.387 which seems possible.

Every call contract $660 and below closed ITM and the MM will need shares to deliver to those call holders (1,688 contracts or 168,800 shares). All the puts $670 strike and above ended ITM, and the MM will be expecting delivery of that stock (4494 contracts, or 449,400 shares) when the puts are exercised. That nets 280,600 shares expected to be delivered to them, which doesn't quite fully cover their 301,960 shares they shorted over the day, so they could be expected to buy 21,360 shares to close their short. That the net stock delivered to them from the options with strikes over the day's trading range matches pretty close to the net exempt shorting is additional evidence that the day's exempt shorting may well be mostly legitimate market making activity.

Additionally because the after-hours price crept up, it is possible many of the $670 call holders could request exercise even though it was OTM at the close. That represents up to 44,400 additional shares to purchase. At the same time, some $670 put holders could request that their puts not be exercised, but since exercise is automatic for them unless they take action I think this would be fewer. Still this would not bring the net after-hours covering anywhere close to the after hours buy that you identified.

In summary, on the one side, shorting 301,960 shares throughout the day seems potentially plausible as a legitimate market making activity. On the other hand, if hedged correctly you would expect little need to re-balance after the close (21,360 shares + up to 44,400 more bought by my estimate) which does not match up with the after hours purchase. So the two share sizes could just be a coincidence.