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

[mrmage]

Mixed bag. Controls, path planning, and simulation are poor, so no loss there. Only perception has made good progress, so Steedly is probably a loss. Mr Havlak's group now has a total ignoramus (about self driving) in charge, so it will continue to make little to no progress until Baglino moves to another job or is fired.

Did I say it would take at least 5 years? I'll revise that to at least 10 years -- they are not hiring people who understand the nature of the problems, apart from Karpathy. This is probably Musk's fault because Musk is delusional and does not want to admit the nature of the problems.


Yep. You can bet money on that. The total failure of full self driving should not harm the company too much though; it is still the best driver assist system out there. I do want to know how much in FSD deposits they have taken, because they will have to refund a lot of it.

Looking at this from a macro perspective, EM has a pretty good track record after he becomes personnally involved in a project and makes it a priority.

Although I only give FSD a 15% chance in 5 years on my financial projections, I think they have at least a 30% chance in 5 years. However, the odds are very good that someone will drive coast to coast on AP without intervention, including local traffic, in the next year or two.

When videos are posted and people talk about it, market perception of robo will completely change. It will still take years to make it safe enough for robos in the general case, but once people can see something coming, it will start getting priced in.

We might be saying the same thing. The glass being 30% full vs 70% empty.

 

[Singuy]

Market doesn't handle externalities well. We got rid of slavery, smog & lead paint and asbestos and DDT … because of govt action.

Problem is not waiting for government, problem is when everyone gives up on government. Then, only big lobbies get a say in what govt does.

Slavery will probably be around if the north needed them as much as the south.

Getting the masses to jump onboard the renewable bandwagon flow lobbying money to the top which will stop misinformation or at least get freaken Wisconsin to allow Tesla sales.

 

[Stars aligned]

I am with both of you on this. We'll see Model Y sooner than promised and in proper quality and quantity.

It seems also if not obvious so quite likely to me. Makes me wonder why there is no word in usual press (teslarati & co) what might be produced with the additional capacity. But maybe better so.

 

[LytHka]

Total dumbassery! Wow. I am glad that it does not really hurt the company for this to fail. Musk needs some sense slapped into him. He's already two years past his original two-year prediction and he doesn't yet realize that it's actually a much harder problem than people think it is.


I am beginning to think that the stock price is low not because of Q1, but because Autonomy Day demonstrated that Musk is actually delusional. And many investors don't want to invest in a stock with a delusional CEO.

Personally, I don't think these delusions hurt the company's financial future that much.

They'll have to refund a lot of the "full self driving" payments -- but apparently the first round of "full self driving" sales didn't amount to much since very few people bought it. I don't know how many people are buying the current package, but since the package has some features now, and a lot of disclaimers, they'd probably only have to refund *part* of the money -- any lawyers have a guess as to what percentage?

If they do something idiotic like self-insuring driverless robotaxis (no third party insurance), that actually could kill the company, so if they do that, I'm out. But so far it doesn't look like they're considering that, thank goodness.

So to me, it just doesn't matter that full self-driving is not coming any time soon. But Musk's looney-tunes ravings about it could be spooking a lot of other investors and holding the stock price down.

The funny thing is that you're just as sure it's 10 years away as Musk is that it's 1-1.5 years away. I think it will probably happen sooner rather than later, given the advancements in silicon that have happened enabling more complex neural net models. Maybe instead of 1.5 years it takes 2x longer at 3 years. But far less than 10.

 

[neroden]

I would argue the single hardest AND most important software problem of our time would be an AI that could label fake news as such. FSD is known inputs and outputs to mimic very basic human intelligence. Comparatively FSD is a cake walk. Labeling fake news as fake is something the vast majority of humanity itself struggles to do.

Fair enough, the problem with fake news is that you have an adversary who is specifically trying to defraud you. The problem of self-driving is effectively a full AI problem *without* an adversary (unless pranksters are out on the roads setting up fake road signs and painting fake road lines), while the problem of identifying fake news is a full AI problem *with* an adversary, so it's harder.

I would also argue curing many genetic diseases are harder problems.

OT:

Nah, I'd say curing each properly identified genetic disease is easier. Biology is very tricky idiosyncratic stuff but you'd be amazed. The hard part is just figuring out what's actually going on -- learning what the diseases actually *are*, biochemically speaking. That actually is extremely difficult and has to be done individually for each disease. Once you've pinpointed what gene isn't working right and what allele should be present to work right, we already effectively can replace genes with other genes, though the delivery system for rewriting the cells without getting hit by autoimmunity remains tricky.

 

[Pezpunk]

(stock looking rather BAF...)

 

[KSilver2000]

If they do something idiotic like self-insuring driverless robotaxis (no third party insurance), that actually could kill the company, so if they do that, I'm out. But so far it doesn't look like they're considering that, thank goodness.

On a related note:
2+ months late and counting on Musk’s original estimate of rolling out a Tesla insurance. Not sure if he completely thought through how complex getting into insurance business would be.

 

[Stars aligned]

Total dumbassery! Wow. I am glad that it does not really hurt the company for this to fail. Musk needs some sense slapped into him. He's already two years past his original two-year prediction and he doesn't yet realize that it's actually a much harder problem than people think it is.


I am beginning to think that the stock price is low not because of Q1, but because Autonomy Day demonstrated that Musk is actually delusional. And many investors don't want to invest in a stock with a delusional CEO.

Personally, I don't think these delusions hurt the company's financial future that much.

They'll have to refund a lot of the "full self driving" payments -- but apparently the first round of "full self driving" sales didn't amount to much since very few people bought it. I don't know how many people are buying the current package, but since the package has some features now, and a lot of disclaimers, they'd probably only have to refund *part* of the money -- any lawyers have a guess as to what percentage?

If they do something idiotic like self-insuring driverless robotaxis (no third party insurance), that actually could kill the company, so if they do that, I'm out. But so far it doesn't look like they're considering that, thank goodness.

So to me, it just doesn't matter that full self-driving is not coming any time soon. But Musk's looney-tunes ravings about it could be spooking a lot of other investors and holding the stock price down.

Have you paid attention to the fact that the SP is up quite a bit lately?

 

[M3Rider]

Google "Yevgeny Prigozhin". He runs the only proven-to-exist paid troll farm.

Btw, that is the least lethal part of his "empire".
He also is responsible for:
- attacking American troops position in Syria with his "private military contractors" drawn from regular/retired troops ranks. The whole notion of private military is illegal in Russia, yet he has no trouble starting a little private war for the purpose of attempting to privatize that Syrian oil extraction site and no consequences from the government side.
- hiring semi-criminal elements to assault opposition figures/bloggers, some of them killed by injection of chemicals simulating heart attack, which you can guess are not sold in a grocery store/pharmacy.
- organizing assassination of 3 journalists in Cental African Republic for attempting investigation into his private military contractors company Wagner.

Russian government is not interested in investigating any of these.

FYI, per the troll that reports on the factory, some of their "work assignments" seem to be commercial, i.e. not only do they work for the government, but also serve private commercial interests, such as bashing competition or products on the internet.

With deep fakes coming up and AI being able to generate readable articles on a specified subject, pretty soon you can't trust anything you read on the internet.

Maybe only reviews for the verified purchases on Amazon? Sad times.

 

[neroden]

The funny thing is that you're just as sure it's 10 years away as Musk is that it's 1-1.5 years away. I think it will probably happen sooner rather than later, given the advancements in silicon that have happened enabling more complex neural net models. Maybe instead of 1.5 years it takes 2x longer at 3 years. But far less than 10.

I think if Tesla's team, including Musk, really understood the nature of the problem and were parallelizing their work appropriately, 2 years might be possible. Appointing random engineers who are more used to working on hardware to manage parts of the team is going to delay progress by several more years, because they have zero experience and the wrong attitude -- any habits learned from hardware development are wrong.

 

[NakedLong]

Total dumbassery! Wow. I am glad that it does not really hurt the company for this to fail. Musk needs some sense slapped into him. He's already two years past his original two-year prediction and he doesn't yet realize that it's actually a much harder problem than people think it is.

Tesla's/Elon's understanding of the problem since bringing it in house after MobileEye in 2016 has no doubt improved drastically. I would say that from 2017-2018 Elon would have had a solid understanding of the issues and problems after 1-2 years of working on the problem directly. I vaguely remember Elon admitting some mistakes in how they approached FSD... can't remember what it was though - using Lidar maybe? I would therefore put a lot more weight on Elon's later predictions than the earlier one's before the massive amount of work that has now been done.

Considering Elon is surrounded by talent in the field like Karpathy, and has done a lot of work around all aspects of self driving, I would argue that Elon currently is one of the most qualified people to judge when FSD will arrive, despite previously over optimistic estimates and a history of aggressive deadlines.*

I think you can take Elon's estimate of 1 year, add a year or 2 for real world issues, and you will have FSD in 2-3 years in the first markets, with incremental rollout to the world taking an additional 5 years. Anybody thinking it will take longer than that probably needs some sense slapped into them

* I think it is important to know who knows more than you in specific fields. I had a discussion with a physicist who told me that most likely the universe is non-deterministic (because of Bell's Theorem). My understanding of Quantum Mechanics is far less than his, so I should adopt his conclusion. I apply the same logic to Elon and FSD. Even if their predictions turn out to be wrong, there are usually good reasons for making wrong predictions. There is no point in chasing the rabbit, as they say in poker. Historically in the field of AI the early predictions were not borne out. Now however that we have advances in NN and ML, a lot of breakthroughs are happening, often faster than people are predicting. What is clear is that it is a very hard field to estimate timelines in, and being pedantic about a few years does not do justice to the complexity of the problem.

 

[neroden]

OT

OT
Got a solution for emissions from concrete, steel, agriculture? Govt has to get involved, or the Keeling Curve keeps climbing, even with every ICEV replaced with EV and every coal plant replaced with solar/storage.
One hope is that govt are inspired by Tesla and having seen that solutions are possible, get off their butts and try doing something.

Solutions for concrete exist, the problem is that old-fashioned concrete is *slightly* cheaper and already has plants built. A targeted carbon tax on concrete, designed to make magnesium concrete and iron concrete cheaper than quicklime concrete, would do the trick.

Solutions for steel exist and are being built right now in Sweden. The problem is that it's *slightly* cheaper to make direct reduced iron using coal or methane than to make it using hydrogen, and it's cheaper to make hydrogen from methane than from renewables, and there are a bunch of old blast furnaces operating (although electric arc furnaces are already cheaper to operate). Again, a carbon tax of the correct price would do the trick.

The main solution for agriculture is to electrify the agricultural equipment (tractors etc.)

 

[neroden]

Bloomberg - Are you a robot?
My guess:
1) Tesla have bought the MY robots
2) They will build the line to temporarily make M3
3) They will make changes in 1 year to make the line 100% MY

Nah. I think new lines ares coming up immediately as MY.

I think what's happening now is optimizations in the rest of the factory: both in the shared parts of the factory (like the paint shop), and perhaps "back-porting" Model Y design improvements (like the "1 piece instead of 20" thing) to the Model 3.

 

[Tslynk67]

(stock looking rather BAF...)

It was, but seems to have fallen into an open manhole cover right now. Any reason for this?

 

[StarFoxisDown!]

It was, but seems to have fallen into an open manhole cover right now. Any reason for this?

At this point with all the shenanigans we've seen with this stock.....there really shouldn't be a question of "reason" anymore

When the stock started to rise out of the of 236 range, I literally said to myself, this is going to be capped by a series of sell orders...….the manipulation never fails

 

[Dutchie]

Looking at this from a macro perspective, EM has a pretty good track record after he becomes personnally involved in a project and makes it a priority.

Although I only give FSD a 15% chance in 5 years on my financial projections, I think they have at least a 30% chance in 5 years. However, the odds are very good that someone will drive coast to coast on AP without intervention, including local traffic, in the next year or two.

When videos are posted and people talk about it, market perception of robo will completely change. It will still take years to make it safe enough for robos in the general case, but once people can see something coming, it will start getting priced in.

We might be saying the same thing. The glass being 30% full vs 70% empty.

Idk. They have a couple of hundred - well educated - people working on the project. I trust Karpathy in what he is doing. Again, Elon's timeline will probably be off but i trust it will come.

 

[Tslynk67]

At this point with all the shenanigans we've seen with this stock.....there really shouldn't be a question of "reason" anymore

When the stock started to rise out of the of 236 range, I literally said to myself, this is going to be capped by a series of sell orders...….the manipulation never fails

You're right - after more than three years of owning this stock, I still seek reason and logic

 

[Pezpunk]

It was, but seems to have fallen into an open manhole cover right now. Any reason for this?

clearly my fault. apologies!

 

[neroden]

I think Tesla was serious about the stretch goal of 500K production by the end of the year or — more realistically — the backup goal of “very likely” 500K production in the 12 mos through 6/30/2020 stated in the Q1 letter:

“If our Gigafactory Shanghai is able to reach volume production early in Q4 this year, we may be able to produce as many as 500,000 vehicles globally in 2019. This is an aggressive schedule, but it is what we are targeting. However, based on what we know today, being able to produce over 500,000 vehicles globally in the 12-month period ending June 30, 2020 does appear very likely.”​

I think many dismissed that as pie in the sky, especially in the midst of Q1 doom and gloom, but depending on the speed of GF3 ramp 500K in the next 12 months is a stretch but not impossible.

There's no particular known reason why they wouldn't be able to get to the number for the June 2019-June 2020 period. Remember, 500K/year PLUS Model S/X was the rate planned for several years ago when Model 3 reservations were opened. The Model 3 line didn't live up to expectations and seems to be topping out between 300K-350K/year; Model S/X adds 80K-100K, for a total of 380K-450K. Shanghai is going to a rate of 150K/year, but only starting at a rate of 1 in 2020, so probably around 40K total by June 30 -- that's 420K-490K. Model Y will also be starting in 2020, targeting 300K-350K rate... so it might have produced 70K-80K by June 30 -- that's 490K - 570K.

If ramp-up goes well at either Shanghai or for Model Y, numbers will be higher earlier. If it goes badly, they will be later. Still, it's all in the right ballpark.

 

[Cherry Wine]

At this point with all the shenanigans we've seen with this stock.....there really shouldn't be a question of "reason" anymore

When the stock started to rise out of the of 236 range, I literally said to myself, this is going to be capped by a series of sell orders...….the manipulation never fails

The weekly chart is starting to look pretty solid though - six straight weeks of higher closes if this week holds. Are we officially out of that downtrend yet?