News for hackers. Stuff that matters.
Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by ArtTimeInvestor on Wednesday November 29, 2023 @02:52AM (#38456700)
Robotaxis.

Car rides are a big market.

There are 2 billion cars on the road.

If 10% of car rides are done in a robotaxi, that is over 200 million rides a day, over 70 billion rides a year.

Sell a ride for $10 and that's a 700 billion dollar market.

Capture 10% of that and you make $70B a year.

One taxi can do over 100 rides a month. That is $1000 in revenue. Operating cost should be less than half of that.

So you have 50% margin: $35B profit per year.

At a p/e of 30, that is over $1T market cap.

Teslas current market cap is $700B.

Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by bagels on Wednesday November 29, 2023 @02:57AM (#38456729)
Robotaxi is not a foregone conclusion. Maybe they'll pull it off. I don't know if there's going to be a giant leap like GPT, or a continued slow indefinite grind like we've seen so far with self driving. I'd be skeptical of anyone who claims to know unless they've already accomplished it.
Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by ArtTimeInvestor on Wednesday November 29, 2023 @03:16AM (#38456847)
I think it is.

The current generation of neural networks already seem to do pattern recognition better than humans. And that is all that is needed to press the brake in a dangerous situation. All that is needed is to train them on enough examples of different situations that occur on the road.

Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by bagels on Wednesday November 29, 2023 @03:22AM (#38456900)
I think, as Cruise found out, you can't have priors for everything (eg. don't drag a pedestrian you ran over down the road).
Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by ArtTimeInvestor on Wednesday November 29, 2023 @03:44AM (#38457024)
You only need more priors than humans have in their mind.

As I said, neural nets are already good at pattern recognition. That is all that is needed to recognise you ran over someone.

Humans are still better at long reasoning sessions. Aka "thinking something through". But that is not involved in recognizing you ran over someone.

Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by alienicecream on Wednesday November 29, 2023 @08:24AM (#38459026)
You think a human wouldn't drag another human under their car because they've pattern recognized that from a prior example?
Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by ArtTimeInvestor on Wednesday November 29, 2023 @09:41AM (#38459853)
Not just one example. They learned from many examples that dragging someone or something along is not a good thing.
Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by alienicecream on Wednesday November 29, 2023 @11:29AM (#38461495)
I've seen examples of humans beings shot out of cannons where they were perfectly fine afterwards and the crowd applauded. I haven't seen any examples of people being shot out of cannons where they were not fine afterwards. Should I shoot people out of cannons?
Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by ArtTimeInvestor on Wednesday November 29, 2023 @11:44AM (#38461701)
What's your point?
Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by alienicecream on Wednesday November 29, 2023 @01:11PM (#38462899)
That understanding is not the same as pattern matching and if that's all you rely on to drive your car you're going to end up dead or seriously injured.
Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by ArtTimeInvestor on Wednesday November 29, 2023 @02:17PM (#38463869)
Even when pattern matching is your only tool and you have seen only positive outcomes of humans beings shot out of cannons, you will still not want to shoot people out of cannons. Because it matches so many negative patterns. You learned that explosions are bad, cannons are bad, weapons in general are bad, loud bangs are bad, people flying through the air are bad etc.

So this is not an example of pattern matching being insufficient to avoid danger.

Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by ActorNightly on Wednesday November 29, 2023 @04:12AM (#38457208)
>All that is needed is to train them on enough examples of different situations that occur on the road.

The only real way this would work is if you build a pipeline where you have a simulated world with realistic physics, rendered through some pipeline to represent camera images (in relatively high fidelity) and/or lidar outputs, and teach the network to predict the evolution of these scenes, and do this with a shitload of random data

This is probably orders of magnitude more expensive to do than training GPT models, since for vision, you would have to render it with ratyracing.

Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by ArtTimeInvestor on Wednesday November 29, 2023 @04:18AM (#38457248)
Human drivers are proof that this is not needed.
Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by ActorNightly on Wednesday November 29, 2023 @05:12AM (#38457615)
I hate that companies in the vision self driving space like Tesla and Comma AI have made the act of driving seem like its trivial.

Human drivers drive with EXCEPTIONALLY more "software" under the hood. We simulate the world around us using learned rules across multiple domains. For example, if someone has never driven at night, but understand that the cars have tail lights that are red, we can deduce that pairs of red lights are cars on the road without ever being subjected to the visual of driving at night. Its this kind of processing that lets us be safe.

In the current form, ML models are way subpar to this. A superhuman driving agent should be able to drive through a construction site with cones and debris, or a busy parking lot, or a grass field full of logs. No model out there can do this. This is why you absolutely need simulations if you want to do it traditional way. Comma AI has gained quite a bit of success in training in a simulated environment that simulates deviations from straight line driving on real world video.

Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by ArtTimeInvestor on Wednesday November 29, 2023 @05:32AM (#38457746)

    we can deduce that pairs of red lights are cars on the road
Generalization, abstraction, understanding. That is what neural networks are about. Current NN's are already very good at that. FSD constantly has to understand lighting situations it has not seen before.
Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by ActorNightly on Wednesday November 29, 2023 @03:01PM (#38464465)
You still need training data.

Even if you had a simulation set up for training, and enough compute on the vehicle to run something like MuZero real time inference for self driving, you would still run into the problem of humans setting up all the possible driving scenarios, which will likely leave some out and the model will never learn the right actions for those - meanwhile, to any human, the action would be very obvious. Over time you could probably get close enough to a very very small error rate, but you would still be hesitant to trust the system.

A true superhuman driving agent would most likely be able to take a picture of a scene, and then give a prediction of evolution of object position in 3 dimensions for a given time window, and give a confidence score on that. For example, if the single image is from few on a highway, it should be able to predict that cars are in fact moving, because the chance of cars standing still on a highway is very low.

And to train a model like that, you most likely need a base model that can "understand" physics.

Additional images from the past or the future would generate prediction that is more accurate, and also improve the model self assessed confidence. And then you would have some heuristic algorithm like MCTS based on confidence levels on the best course of action.

Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by bagels on Wednesday November 29, 2023 @05:29AM (#38457730)
It's not just objects, but it has to predict what people are doing, going to do, or are trying to communicate. For example, police redirecting traffic, a passenger having a medical emergency, other drivers driving erratically, etc.
Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by chx on Wednesday November 29, 2023 @04:27AM (#38457304)
And current generative AI has jack all to do with humans.
Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by amenhotep on Wednesday November 29, 2023 @05:31AM (#38457737)
Have you not seen the results that have been shown recently with Gaussian splatting? Photorealistic rendering looks like it'll become a hell of a lot cheaper than raytracing soon enough, and this might actually be a realistic idea.
Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by ActorNightly on Wednesday November 29, 2023 @03:05PM (#38464530)
I dunno if statistical methods can be applied to generating photorealistic images from simulation, where you don't have enough information in a scene, since you are starting with 3d models.
Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by aylmao on Wednesday November 29, 2023 @05:06AM (#38457576)
Even if this were a given (and it isn’t), I think it’s unlikely Tesla robotaxis capture the transport market like that.

There’s a lot of cars to replace. Capturing 10% of yearly car sales, worldwide, is already a very tall order.

Capturing 10% of all cars on the road would require this level of gargantuan sales for many years, and that’s just so 10% of cars in the world are your brand, not so 10% of rides are on cars owned by the fraction of people who bought your brand and decided to let it drive strangers while they’re not using it.

Moreover, beyond the safety concerns there’s policy concerns— knowing how to safely drive in Indonesia doesn’t mean being allowed to drive in Indonesia. Resource constraints— where is all the lithium going to come from, and how will the people around the world feel about this?

And there already are alternatives beyond the car industry. Public transit, for example. Demand for not driving or not owning a car can’t only be safely fulfilled with robotaxis.

Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by ArtTimeInvestor on Wednesday November 29, 2023 @05:28AM (#38457725)
I said 10% of rides will be robotaxis and 10% of that will be in Teslas.

That is 1% of all rides taking place in Tesla robotaxis, not 10%.

Also it is about rides, not about cars. Less than 1% of cars being Tesla robotaxis is enough for 1% of rides to be Tesla robotaxi rides. A robotaxi will do multiple times more rides per day that the average car.

Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by ben_w on Wednesday November 29, 2023 @03:01AM (#38456760)
Given what's happening with all the rest of AI, I suspect the demand for commuting is going to go down by the time robo-taxis are viable.

Will robots that lead to less commuting be simply balanced with more fun trips for similar revenue potential, or will they lead to a complete shift of world economics that means the money stops being meaningfully comparable to today's?

Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by ArtTimeInvestor on Wednesday November 29, 2023 @03:18AM (#38456863)
Robotaxis will bring down the price of rides. That means the number of rides will go up.

Less people will commute to work. But more people will meet for a coffee and order stuff online which then gets delivered in 30 minutes by a robotaxi.

Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by the_other on Wednesday November 29, 2023 @03:33AM (#38456968)
> Less people will commute to work. But more people will meet for a coffee and order stuff online which then gets delivered in 30 minutes by a robotaxi.

We havn’t ever seen a situation where a reduction in the cost of work lead to less work being done.

What happens is that employers (and the market) demand more work, and only labour law holds back the demand that employees work all their waking hours.

You’ll probably respond that UBI will give everyone all the free time they want… but capital owners will have to fund UBI through tax. And capital HATES paying tax.

This feels like stuck gears to me. What am I missing?

Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by ben_w on Wednesday November 29, 2023 @03:40AM (#38457003)
> You’ll probably respond that UBI will give everyone all the free time they want… but capital owners will have to fund UBI through tax. And capital HATES paying tax.

> This feels like stuck gears to me. What am I missing?

Communism.

Or at least, some variant of it. If your government owns the full automation and operates it for the citizens, that's much the same as the upper limiting case of a corporation owning the same full automation for the benefit of the shareholders.

(I'd say "and this time nobody needs to go to the gulags", but having the government aligned with the citizens is a separate problem: any given state might get it right, or history might rhyme).

Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by u320 on Wednesday November 29, 2023 @03:50AM (#38457071)
It trivial to project big profits if you assume a monopoly. The problem is that it's true for any sector, you have to justify why Tesla is the one to achieve it or else your numbers are meaningless.
Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by dkrich on Wednesday November 29, 2023 @03:00AM (#38456752)
I can’t tell if this is sincere or a joke.
Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by lvncelot on Wednesday November 29, 2023 @03:04AM (#38456784)
It kind of reads like some sort of VC version of the Drake Equation.
Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by adhesive_wombat on Wednesday November 29, 2023 @03:51AM (#38457077)
I've heard the same flavour of equations from everyone I've ever heard talking about starting a company. One guy was calculating that if he sold his gadget to "just" 10% of people in Africa, he'd make so much money.

Spoiler: he did not sell a thing to 10% of people in Africa.

Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by consp on Wednesday November 29, 2023 @03:13AM (#38456835)
This is simply "calculating yourself into richness", and in the mean time forgetting a lot of cost and failure modes. If you can get money from a VC that way the joke is on the VC.
Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by NhanH on Wednesday November 29, 2023 @02:58AM (#38456743)
You are claiming thay current human driven taxi/ car rides market are making 700 billion dollar profit per year (70B / 10%)

They don’t

Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by ArtTimeInvestor on Wednesday November 29, 2023 @03:05AM (#38456793)
Not profit. Cost. Using robotaxis is going to cost less than operating your own car.

The cost of operating the current car fleet is way higher than $700B. Buying the cars alone is higher. There are 70M new cars on the road each year. At $10,000 per car, that would already be $700B.

Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by bagels on Wednesday November 29, 2023 @03:44AM (#38457030)
Why would it cost less?

What is the difference in utility of a robotaxi and a meat taxi besides maybe a reduction in cost per ride (how much is labor vs. non-labor cost), and why isn't everyone using Uber to get everywhere now?

The robotaxi vehicle still needs to be manufactured, purchased, financed, maintained, fueled, insured, with room for profit for the robotaxi company. None of these things are free.

Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by ArtTimeInvestor on Wednesday November 29, 2023 @03:51AM (#38457081)
Capacity utilization will make using a robotaxi multiple times cheaper than owning a car. You car stands still most of the time. A robotaxi can drive most of the time.

Automation and economies of scale (maintenance, repair, cleaning, charging) will also contribute to the lower cost of a robotaxi fleet compared to privately owned cars.

Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by bagels on Wednesday November 29, 2023 @03:58AM (#38457128)
Why will people choose robotaxi who are not currently using Uber?
Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by ArtTimeInvestor on Wednesday November 29, 2023 @04:09AM (#38457188)
Because the robotaxi arrives faster, is cheaper, safer and more convenient.
Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by hef19898 on Wednesday November 29, 2023 @04:24AM (#38457285)
Besides driver breaks, robotaxis are bound by the same constarintsbas normal ones. So no, it won't arrive faster, might (and that is already some claim) be a tad cheaper if you ignore CAPEX, have yet to prove to be safer than humans (so far self druving cars are not, and by a long shot, safer than humans) and convenience doesn't change except for the most, how to put it politely, asocial beings that cannot stand being anywhere near to other human beings, in case of taxis a driver.
Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by ArtTimeInvestor on Wednesday November 29, 2023 @04:39AM (#38457394)
It arrives faster because there are more of them. Every time the density doubles, wait times roughly halve.

It is way cheaper because majority of Uber cost is the driver.

Safety: https://news.ycombinator.com/item?id=38456847

It is more convenient because it has more space. No drivers seat, steering wheel etc.

Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by hef19898 on Wednesday November 29, 2023 @04:48AM (#38457454)
More taxis mean more CAPEx, mean utilization goes down off-peak, bothe of which result in lower profitability.

Robot driver or not, basic financials and operational constraints still apply. They can only be ignored for quite a while if a company has endless VC money, unit economics ultimately catch up so, always.

Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by aylmao on Wednesday November 29, 2023 @05:24AM (#38457689)
> It arrives faster because there are more of them

I think you’re overestimating how many people will buy Teslas.

Also, how many people who own Teslas would robotaxi it. A company might do it as a business, but it’d just take waking up one morning to discover a drunk passenger puked all over your back seat, or fluid remains of someone having sex in there [1] for individual owners to pull out of the program.

[1] https://sfstandard.com/2023/08/11/san-francisco-robotaxi-cru...

Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by dagw on Wednesday November 29, 2023 @05:45AM (#38457828)
It is way cheaper because majority of Uber cost is the driver.

That is only true because Uber doesn't have to pay for buying and maintaining the cars in their 'fleet', having offloaded that cost to their drivers. If Uber were to move to owning robotaxis then they have to have to move the car ownership costs back onto their books.

Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by ArtTimeInvestor on Wednesday November 29, 2023 @06:11AM (#38458016)
The total monthly cost of owning and maintaining a car is just a few hundred dollars. The total monthly revenue of a taxi is several thousand dollars.

When you take the driver out of the equation, the cost of providing rides plummets massively.

Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by bagels on Wednesday November 29, 2023 @06:38AM (#38458202)
Who is paying for fuel and insurance?
Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by ArtTimeInvestor on Wednesday November 29, 2023 @07:33AM (#38458601)
That's in the "few hundred dollars" monthly total cost.
Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by hef19898 on Wednesday November 29, 2023 @09:58AM (#38460118)
Ever heard of deprecation of assets?

But sure, as long as other take all the costs and most of risks away from the new liberatrian tech nobility we are suppossed to not care about those little things, right?

Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by hef19898 on Wednesday November 29, 2023 @03:05AM (#38456788)
Which is exactly ehat Tesla did trying to maintain a "tech" market cap instead of defaulting back to a more realistic car maker one. Among other things.
Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by poisonborz on Wednesday November 29, 2023 @06:05AM (#38457967)
If you are truly an investor as per your username and the above is your serious reasoning for buying Tesla stock...
Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by crq-yml on Wednesday November 29, 2023 @08:43AM (#38459218)
The players deploying robotaxi right now are entering a market that resembles shared micromobility(scooter, e-bike) - what they need is not just vehicles, but to get on the side of the appropriate regulatory bodies and to get eyeballs on the product. It's a fight won one city at a time, and it favors partnering with companies like Uber that have built experience in both of those aspects.

This is a fight that Tesla doesn't seem to be engaged with - it's staying on the manufacturing side of the issue. Its opponents are going to see this as a gun they can fire to shut out Tesla's supply-centric economics: install regulations on robotaxi - vehicle form factor, signals, testing - that make it impossible for anyone but a well-capitalized player to enter. Then it doesn't matter how cheap the vehicle gets if the other costs required to operate legally keep small players out.

Tesla's recourse in this market is to either sell a commodity product that passes regulation, which is terrible for their bottom line, or to become a robotaxi company themselves, which is not unfeasible(they have the Supercharger network, they can grow more vertically) but which they aren't currently aimed for. If they did that the fast path would be to find a ridehail partner and acquire them.

I believe in the robotaxi concept, but I don't think it's settled as to what form it ultimately takes.

Re:It's Still Funny That 1k People Gave Tesla $250k for a Ro... by ArtTimeInvestor on Wednesday November 29, 2023 @09:39AM (#38459832)
I'm sure that if Tesla creates the best robotaxi software, it will find its way onto the road one way or another. My feeling is that they will operate the robotaxi fleet themselves. But they might also just sell or rent the cars. Or even just the software.