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

[EVNow]

Basically, Tesla is betting that with NN will be able to crack FSD - and quickly because of exponential advancement.

The first big step would be when they release NOA in city streets. If that works well, I think market will react.

 

[mblakele]

Brad Templeton is impressed: Tesla reveals impressive new details on autonomy plans

 

[Cosmacelf]

One simple answer might be they will just upload a crap ton of data, without doing any manual labeling. And they will just assume the driver's input is the True condition.

E.g. the driver's input is the label. Even though in some cases the driver will do the wrong thing, this is such a small % of the data that they expect to get most of the way to good accuracy without worrying about that.

BTW I surmise the reason they are going to do this is they are still using the NNs mainly for perception. Which may be good enough. But I would guess the next gen chip is expected to do more end-to-end deep learning for driving - this requires much more data and compute power (both training and inference).

That's their next-level goal.

Yes, it was implied that right now, the NN just does perception and the actual driving is hard coded software. But they are starting to work on a system where the NN would do all the driving. Definitely next level.

 

[oneday]

Was I the only one cringing when Elon felt he had to re-state that all cars made today will be able to FSD? Wasn't AP1 supposed to be that, then 2, then 2.5? And even then, which member of the audience wasn't up to speed with this yet?

AP1 was never sold as self driving. AP2 & 2.5 was sold as FSD capable but might need a computer upgrade. That computer is here.

 

[carterm2]

I don't believe that any car, any vehicle, except perhaps a jetliner, has ever had significant parts like the motor or batteries to be rated for a million miles lifetime. Unprecedented. It sold me on Tesla when I heard about the motor... EM was thinking way past 5 year timeline.

(This is where you smarter people point out the outliers.)

 

[EVNow]

Triggers. By the sound of it they can remotely configure cars to watch for "similar images to X" and upload the results.

Tangentially: shadow mode is real — but it sounds like it's used to pre-flight specific features rather than always running.

So, their ability to collect edge data will be limited to how well the car can catch "similar images". May be the ability gets better as NN gets better and can label more stuff … not sure.

 

[capster]

That's a key point: learning takes place offline. Neural nets don't change while they're running.

That said, Lex Fridman's latest interviewee Ian Goodfellow mentioned one possible reason to change this: security. He sees it as a potential security problem that a given NN always returns the same result for the same input. A neural net that does some form of online learning might be more secure. However I think that's some years off.

Supervised neural networks don't change when running in the field. An unsupervised network can be left to continue to learn. They can learn a new category in as little as one exemplar.

However, there are perverse training regimens that can conceivably mess up an unsupervised classifier.

Basically, you can cause categories to drift. You do this by presenting exemplars that match a category vector but do so in a sequence that slowly migrates that category vector.

 

[Ulmo]

Elon: "If you need a geofenced area you don't have self-driving!"

Who was asking that? That's a stupid question. Of course Elon is right, but why would he even need to say it?

 

[Cosmacelf]

Was I the only one cringing when Elon felt he had to re-state that all cars made today will be able to FSD? Wasn't AP1 supposed to be that, then 2, then 2.5? And even then, which member of the audience wasn't up to speed with this yet?

Wrong AP1 was never sold as FSD. Freeway only, which it does.

 

[bdy0627]

that's irrelevant. As Karpathy said in his presentation, the important thing is to get to the tail 9s. you are not gonna experience all those 9s in a test drive.

Maybe, but as they say, a picture is worth a thousand words. What's an actual FSD demo drive worth, particularly compared with all of the engineering stuff they presented today on paper? If impressive, it can make all the difference in the perception of where Tesla is at and how close it is to conquering FSD. If it is really impressive, I think they should have done the drives first and then showed the nuts and bolts of the engineering behind it. It's more impressive to see a magic trick first before learning how it was done than the other way around.

 

[Right_Said_Fred]

One thing that struck me about today's event: how different Jobs and Musk are.

Think back to Jobs announcing the iPhone in 2007. It was a monumental/historical event, a game-changing milestone that fundamentally changed the industry and launched "mobility." Sure there had been hints it was coming with Palm Treos and Blackberries, but the iPhone blew everyone away and set Apple on a path that would take it to a trillion-dollar market cap.

Now think of Elon's presentation today. I'd argue the event, in hindsight, has the potential to be seen as a game-changing milestone that fundamentally changed an industry and launched "robotaxis." Sure there have been hints it's coming, with Waymo et al. But what Tesla showed off today was pretty mind-blowing and sets Tesla on a path that, maybe?, takes it to a trillion-dollar market cap.

Now think if Jobs had announced the iPhone in 2006, or 2005, a year or two years before it was out and purchasable. And imagine how Wall St would have reacted. They probably would have been "meh" and "show me." Jobs waited until the iPhone was imminently available before announcing it. People realized this was gonna be obtainable in days or weeks. Musk is announcing a robotaxi network that's still a ways out, obviously to alert the market about the difference between Lyft/Uber (both IPOs) and Tesla's competing offering. But Wall Street and the media are at peak skepticism/FUD when it comes to Tesla. Elon could have announced that oh yeah, the our HW3 proprietary chip's spare trillions of compute operations available beyond FSD in the cars will be used for a cure for cancer, which will be done within 18 months of rollout. I don't think even that would move the stock. I'm not sure anything from today is going to be a big boost to the stock any time soon Tesla shows it is executing brilliantly. No more fumbles, like announcing full closure of stores one day and then walking back that announcement 48 hours later. The market is still reeling from that, for good reason.

Does it really matter what SP does tomorrow? Or next week? Or this year? If Elon is right and they reach full FSD this year and start deploying a fleet of a million robotaxi's next year - by the flick of a button, after regulatory approval - then SP will automatically follow. In a huge way. And it really won't matter what it will have done up to that moment. In fact, if it has stayed low or even gone down more, it will only have created a huge buying opportunity. After seeing today's presentation, those who will have missed the opportunity can only blame themselves.

 

[Todd Burch]

Given the demo rides aren’t broadcast, we’ll have to wait for the investors to write up their reactions for more info.

Wonder how MobilEye (now Intel?) is feeling this evening?

 

[Remus]

And they have ML expertise that far exceeds Tesla. And they have an actual Robotaxi service. And they're legally allowed to operate with an empty driver's seat in CA. And they've driven thousands of miles on public roads without anyone in the driver's seat. And they have enough cash to fund any possible business plan.

Where they fall short is business plan. Tesla customers have for years paid thousands of dollars for a product that didn't even exist. Google/Waymo has nothing like that. They better get their a** in gear, or all the "Ands" listed above won't matter.

I am not sure that's the only missing piece.

if they have an "actual" robotaxi service, they won't have trouble getting finances. There is only one very short promotional video from Waymo showing driverless taxi moving. They won't put human drivers back if they didn't have to. Being allowed to test driverless cars in CA does not mean they are conquering the long tail.

I am not saying Tesla is there. But Waymo is definitely not there yet.

 

[bdy0627]

Wall Street doesn't know how to value the robotaxi promises and is skeptical of the timeframe that it might take for Tesla to execute that strategy. Plus, now you have Panasonic limitations on the number of vehicles that Tesla can even produce.

Hopefully Q2 guidance for a profit is still on track so that we don't see as much short term pain.

Panasonic needs to get their production up there, but let's be honest. Whether you produce 7,000/week or 10,000/week robotaxis doesn't change the valuation much compared with the first few thousand you produce.

 

[Cloxxki]

The bigger issue is customer convenience. Would you rather ride in the back seat of a Model 3 or a minivan? Frumpy BEV econoboxes designed specifically for rideshare can offer a superior rider experience.​

If you need to drive any sort of distance, Model 3 for tall guy like myself would be the very last choice. If I could ride in the front I might live. Compared to a much cheaper (per seat) minivan, though...

Once battery production goes global, econoboxes will naturally start to appear. Already today, electric scooters and folding bikes are silly cheap. We were told by JB and Elon that batteries increase by 5-8% annually for energy density? If that all gets cashed in a few years from now, $50/kWh wholesale will be consider pricey. With proper fast charging (Model 3 peaks over 3C, 2017 tech) say at 5C, batteries need not be that large. Sweden is already doing real life in-road charge as you drive infra. At this rate, large batteries may go the way of the dodo in a decade or so. And that would be a good thing, actually. Because, environment and cost.

 

[StarFoxisDown!]

Ugh the suspense of the demo impressions is killing me. I want to know what it can do on surface streets right now.....does anyone know if there's a roundabout on the tour drive???

 

[Cosmacelf]

Given the demo rides aren’t broadcast, we’ll have to wait for the investors to write up their reactions for more info.

Wonder how MobilEye (now Intel?) is feeling this evening?

They are probably fielding a lot of calls from their OEM automakers...

 

[Johnny Ma]

We’re talking ride sharing valuation. Tesla does not generate any revenue from ride sharing. The market is in a “show me” mode with Tesla, they do not believe any forward looking statements until they come to fruition. Therefore, Uber has a $100b ridesharing valuation and Tesla has $0. I don’t agree with that, but that is what the market is saying

show me mood indeed. tesla seems to be aware. elon stressed how they've accomplished everything they've set out to do during his part of the presentation and tried hard to imply they'd also reach their FSD goals.

don't know if I've announced it on this forum, but I'm one of the (few of us) here who are FSD skeptic. I'd be more than thrilled to be wrong. that said, enjoyed the presentation. andrej and pete are brilliant. missed a little bit of the other guy's part. he was certainly the most polished.

 

[azaz]

Elon says what he thinks.
He is straightforward and genuine as always.
Why so surprised about this?

Because missed deadlines are used to say Tesla has execution problems and depress the stock price and excitement for future products by saying they can’t deliver

 

[powertoold]

The PR problem will be huge with a Tesla robotaxi service. Imagine the car fires x 100. Even if the robotaxis are 10x better than human, there will be news stories every day about a Tesla robotaxi doing something stupid, lol.