Hacker News Reader

Comment by ArtTimeInvestor | Hacker News Reader

Only one (dismissive) side note about self-driving in the article.

I have yet to hear a solid argument against Tesla achieving full self-driving within the next ten years.

Self-driving will automate billions of daily man hours.

Capturing even just a very small fraction of that value results in enough earnings to justify Tesla's current market cap. Even if you have to discount the earnings for another ten years, in case it takes so long to start selling FSD software.

Talking rationally about FSD always results in reflexive downvoting here on HN. Which I take as a sign, that the thinking about Tesla's future is muddied by emotions among many people. Maybe because people can't imagine fundamental change and also fear change.

View on HN

The only argument I have is, well not so much an argument, but more an observation of a fundamental. Will we be able to get enough compute power without using too much power into a car to do the task? A few years back I would have said "No", but progress is still stumbling along and I am now in the maybe pile.

If we can get the compute power, I cannot see a reason why it cannot happen but it might still be longer than 10 years away.

I think the issue with FSD and Tesla is just how many bold claims from Musk have not come to pass, it is a boy who cried wolf situation. It is my usual criticism is Musk and most of the companies he runs. They are awesome companies in what they have actually done. Anyone would be proud of even half the things Telsa or Space X have done. So why does he always how to go the snake oil sales man route and way over promise? It kind of drags down the excitement for these things.

I'm not talking about excitement.

I'm talking about valuation.

Tesla is the market leader in a market that will have room for multiple trillion dollar companies.

> I have yet to hear a solid argument against Tesla achieving full self-driving within the next ten years.

You haven't heard a solid argument against me personally achieving full self-driving either, but that doesn't make it realistic

You are also assuming they are able to sell their FSD for a profit, and aren't out-competed by e.g. Waymo. Maybe Tesla does make full self driving a reality, but only after Waymo has licensed a better version of self driving to Toyota who sell millions of automated cars before Tesla

Mercedes-Benz is actually in some ways the leader: "first automaker to receive government approval in the US for a Level 3 driving feature. "

https://www.theverge.com/2023/1/27/23572942/mercedes-drive-p...

But yes _if_ it's possible and inevitable then it will also soon be a commodity.

No. They’re not. How people keep falling for this bullshit is beyond me https://www.reddit.com/r/SelfDrivingCars/comments/19e8hjf/wh...

Not denying your point, but the people who drive those Mercedes, are not exactly the same people as the ones buying Teslas and posting their successful rides all over the place.

Yes of course. Not a single fan of self driving among buyers and nobody decided to rent one for lulz and/or youtube clout.

Please.

Well, the arguments against goes the same way - if the system was no good, you can be sure that Tesla promoters would have posted 100s of videos of it already.

The fact that they didn't, tells me it's not that bad.

Thats a WILD derivative you came up with lol I salute your perseverance sir

I do not think that the argument "If the Mercedes-Benz system is so good, then where are all the youtube videos?" is a stronger argument than "If the Tesla system is so good, then where is all the regulatory approval and the full legal liability?"

The latter seems much harder to game, to me.

Also, since you've put the goalposts here, I just looked for these videos and other reviews of the MB-system. What do you know, they do exist.

You think all the Tesla videos are gamed?

I don't think your question shows understanding of what I wrote, or good faith.

The point is: The videos of Tesla's FSD are pretty good evidence of their progress.

MB being allowed to keep the lane in certain situations is not much evidence of them progressing towards FSD.

But not in the same category as regulatory approval and full legal liability.

IDK, this is pretty much what I said above already, just in slightly different words. But whatever, you put your full trust and faith in youtube instead if you want.

Videos of cars handling complex situations are a much better indicator of progress than regulatory approval to do very simple things.

Then it's odd that the people with the cool videos didn't manage to get the "easier" regulatory approval. I already said that in different words in a comment above, maybe you missed it?

A video on Youtube is not a "better indicator" for anything serious or requiring certification. It is an entertainment platform at best, spammed full of disinformation and bias. Preferring it over anything else that has legal weight is some kind of credulous joke. But this just restates my above comment on which is easier to game, edit or cherry-pick.

    Then it's odd that the people with the cool videos didn't
    manage to get the "easier" regulatory approval.

https://news.ycombinator.com/item?id=39754238

    entertainment platform

It's not the platform that is the indicator. It's the videos.

Also, when third or fourth player enters the market. Would FSD not be commodity tech? Just buy a license and then every automaker will have it. At that point it is race to bottom, any self-driving car likely only making small margin over lifetime...

How would Waymo with their sensors bolted onto vehicles they gotta purchase and modify (currently costing them over 200k) outcompete Tesla’s extremely lean manufacturing prowess where they make entire vehicles with all the hardware needed for $20-30k?

Because the sensor package isn’t good enough or else they’d be much closer to Waymo’s capabilities.

1) We’re talking about a scenario in which Tesla does achieve FSD

2) says who? They seem quite close in my view. Certainly beyond ehere waymo was when they started test rides with safety drivers in Phoenix

https://m.youtube.com/watch?v=7Fd9cEEHCzY

https://twitter.com/TeslaPodcast/status/1769500530739421414

> 1) We’re talking about a scenario in which Tesla does achieve FSD

I'm talking about the scenario where Waymo achieves full self driving at a cheaper per vehicle cost than Tesla (once their self-driving is good enough they will pivot to reducing the sensor costs)

Why Tesla? Why not Mercedes, who have shown real self-driving (e.g. https://abcnews.go.com/Business/future-driving-hands-off-eye...) in specific environments, unlike FSD, which cannot self-drive? Or Waymo with their self-driving taxis?

Being the first at claiming you are able to (or will be able to) do something doesn't make you a relevant contender in a space.

Fake ass PR fluff and you fell for it. https://www.reddit.com/r/SelfDrivingCars/comments/19e8hjf/wh...

And the US (and other countries) authorities which allowed Mercedes Self-Driving Level 3 did too? Sounds more like reddit bullshit to me.

Well first of all, you can’t purchase it in the US. It doesn’t exist. Apparently you can order it in Germany, but even then who cares? It’s a traffic assist system, below 30mph or so, needs to closely follow a lead car, and it would probably still shit the bed if something fell out of said car. Mercedes made the conditions so tight they’re just willing to take the risk for the PR fluff they received. Complete garbage.

And with all that, not a single fucking video on YouTube from an owner casually using it

All I see in that highly edited video is a car staying in lane and making a lane change.

While there are many long unedited videos from Tesla owners in which the car does way more.

If Tesla’s cars are more capable, why doesn’t Tesla market them as Level 3?

That's like asking a startup founder "If you are so good at business, why are you earning less than me as an employee?".

The path to complex, high value outcomes is different than the path to simple, low value outcomes.

No need. A 1 minute google makes it extremely obvious which one is more capable.

I didn't link a video, I linked an article.

Why would Tesla be the only one with FSD, they are currently behind in abilities compared to (currently hideously expensive) models form tradional manufacturers.

That’s a hilariously ignorant statement. Please do show me anything remotely similar to this from a traditional manufacturer:

https://twitter.com/TeslaPodcast/status/1769500530739421414

Hint: yea the headlines you saw were misleading

That's why I mentioned that "Capturing even just a very small fraction" is enough.

There is room for multiple trillion dollar companies in this space.

And yet, no other company that is also working on self-driving technology has such a valuation.

Something in your rationale does not add up.

Which other company has as good FSD as Tesla?

"as good as" is doing a lot of heavy lifting.

No company on this area have reliable FSD.

I'm not sure I follow your logic.

Let's assume that you're correct: Tesla achieves FSD in the next decade, and it automates billions of daily man hours. That's only really relevant to Tesla's current market cap if you also assume that other car makers don't achieve FSD in the next decade.

If everyone has FSD, then in terms of relative company value it's the same as no one having it; it's just another feature, like A/C or cruise control.

If self-driving works it's inevitable that the market will transform from mostly selling cars to mostly providing robotaxi service.

Robotaxi service has "winner takes most" dynamics, just like regular human-taxi service. Just look at how stable Uber and Lyft marketshare was over the years.

Imagine company X drops 10 thousand robotaxis in San Francisco area. Anyone who cares to can get a taxi for half the price (or less) of regular taxi.

Company Y develops self-driving technology 6 months later. There's just no way for company Y to re-take significant market share in San Francisco if their costs are similar to company X.

They can drop 10 thousand cars but most San Franciscan who want such service already have app for service X installed and a habit of using it.

Company Y might decide on (expensive) marketing campaign or price war but company X can do that as well. Company X can finance price war from profit but company Y must borrow. And there's only so much you can borrow.

In US we saw this with Uber vs Lyft vs everyone else. Early market share of Uber proved to be bullet proof regardless of what Lyft tried to do.

In China Uber was in a trench price warfare with Didi and when they didn't see a way to win, they decided to loose (and sold their business to Didi).

So the winner of robotaxi market will be the first company that has a working self-driving software AND can produce millions of cars per year to capture the market.

This assumes Tesla wants to enter the taxi business.

If they're just a supplier of robotaxis then a taxi company can choose to switch vendors.

Plus, if the robotaxi business follows the Uber model where "partners" supply their own cars, there isn't even a primary vendor in the first place.

Either way, Tesla cars will become robotaxis at some point.

If a robotaxi does 10 rides for $15 per day, that's $15x10x365=$54,750 per year. If operating the car costs $20k per year, that's $35k profit per year.

It doesn't make sense to sell a car for $30K if it can make $35k in profit per year.

And yet, Uber is loosing money for pretty much their whole existence.

I think we will get to robotaxis one day, most probably. I'm not sure though about the earnings power this will give to those companies.

That day is already here. Waymo's been providing robotaxi rides in San Francisco and Phoenix for a while, and soon, LA and Austin.

I thought that Waymo is 'almost' a robotaxi - in the sense that there is a remote operator that takes over when the car is stuck or it's getting dangerous. I think Cruise revealed the numbers about the % of time an operator takes over and it was surprisingly high - one intervention every 2.5-5 miles.

it doesn't matter what the back end looks like. even if it's all just a call center in the Philippines driving the cars with playstation controllers, on the customer end, there's an app, and it summons a car with no driver in it, and it takes me to my destination.

Their profitability relies on just how much that call center needs to get paid, so it's in their interests to drive that down, and have it be real, but that's their business. As a customer, all I know is that there's a competitor to Uber/Lyft called Waymo, and it's better and worse in specific ways.

No driver means no having to deal with a person. If I'm sending a girlfriend home late at night, she doesn't need to get into a strangers car. the driver could be drinking or hit on her or whatever. robo taxi isn't going to do that, even if it's a call center in the Philippines.

Ok, that's not what I call a 'real' Robotaxi. If the operator needs to intervene maybe once per 100 miles then I would consider we're close.

Re profitability - yes, but if there is more than one companies, they will just drive down the margins. I don't see how it will generate billions frankly.

Being among the best usually means you can stay in there because you can reinvest your earnings into becoming even better.

Google still makes tens of billions of dollars from search every quarter. The revenue is actually still growing.

Other automakers still have software literally worse than a 2013 Model S. They’re still the blackberries of a newly iPhone automotive world. They can not compete with Tesla’s fleet data, OTA updates and widely deployed uniform self driving hardware. Mobileye provides basic ADAS to automakers but they’re hella behind FSD12 and don’t have the kind of deployment capabilities Tesla has.

When the CEO declared that FSD was a "solved problem" EIGHT years ago, then there is good reason to doubt it will be solved in another ten. People who bought cars eight years ago and paid a premium for the promise that their car would have FSD basically lit thousands of dollars on fire because their car is approaching end-of-life, and it still can't drive itself.

There was a divide between people who thought computers would soon beat humans in chess and people who did not believe that.

Computers made only small progress for many years.

That's the nature of technological progress.

You know that it will take place. But you can't predict very well when.

Nowadays, no human even stands a chance in playing chess against a computer.

But chess is a game of complete information, and one where the entire state can be perfectly and unambiguously represented in less than 64 bytes.

I’ve used their FSD package for 4+ years. I don’t trust them to deliver level 2 much less 4 or 5.

> Talking rationally about FSD always results in reflexive downvoting here on HN.

Its hardly a reflexive downvote when you say you haven't heard any solid arguments, when musk has been promising it "next year" every year since 2014 (i.e. the last 10 years). What solid argument do YOU have that the next 10 years will be any different?

The progress. Tesla's FSD, which is very publicly documented, is steadily getting better. There is no sign that the effect of throwing more compute and data at it will slow down.

> I have yet to hear a solid argument against Tesla achieving full self-driving within the next ten years.

You've already asked the same question a few times. Here's what I posted the last time you did (you didn't respond):

Years of deliberate false statements by Tesla about their self-driving capacity should make any rational person very skeptical of their future claims.

Having a CEO who is obsessed with posting far-right-wing political statements on Xitter over managing his actual businesses is also a red flag. What rational manager goes out of their way to alienate at least half of their potential customers?

I am firmly in Tesla's target audience, but under no circumstances would I buy one, if only to avoid the social embarrassment of being associated with this pathologically mendacious malignant narcissist.

Tesla is progressing along a path of more compute and data resulting in better FSD behavior. There is no sign that this progress is slowing down.

I don't see how your points change that progress. It will continue independent of what Musk claims or what political stance he takes.

Reply on news.ycombinator.com