[TheOtherHobbes]

Tesla spent almost years scraping by, almost went bankrupt in 2018, had three years of "exceeds expectation" from 2020 to 2023 and is now being murdered by competitors.

A lot of "income" comes from selling regulatory credits, not cars.

FSD is up to v14 and there is no sense in which it's "Full", or is likely to be "Full" any time soon.

The early computer industry created a good few companies which achieved profitability quickly and posted solid YoY expansion for decades.

The histories of SpaceX and Tesla show no evidence of that pattern.

[scottyah]

> murdered by competitors

Is just a downright lie. The Tesla Model Y maintained its title as the world's best-selling vehicle in 2025 across all powertrains, marking its third consecutive year at number one. Even the CT is the best selling EV pickup, the last bestseller could not find a way to sell it profitably.

[vel0city]

> The Tesla Model Y maintained its title as the world's best-selling vehicle in 2025 across all powertrains

No, the RAV-4 won that title. Its also pretty easy to make that claim when you're only making practically two models of vehicles compared to most other companies with several models. A single car selling well isn't necessarily telling the whole story.

> Even the CT is the best selling EV pickup

Easy to do when 20%+ are being bought by the ownership's other companies.

The CT is a massive failure measured by their own stated expectations. They stated they'd sell somewhere between 250,000-500,000 units a year. They ended up selling only a bit over 20,000 last year.

FWIW, the F-150 Lightning outsold the CT for 2025.

[wnmurphy]

> FSD is up to v14 and there is no sense in which it's "Full", or is likely to be "Full" any time soon.

I think the SpaceX IPO is way overvalued, but have you actually tried FSD v14 yourself?

My car has 8-year old hardware running v12, and it handles like 90+% of my driving. When I test drove 14 it blew my mind how good it had gotten in only 8 months of development. In my opinion, there's question that it's "when," not "if."

[veqz]

Yes, I hear lots of people saying Tesla's "FSD" product can drive by itself already.

Yet if the car crashes while in FSD mode, it is still you who are legally the driver and you who will be responsible. You are the driver, regardless of what the name of the car's software is.

This can be contrasted with e.g. Mercedes' offering, which is actually self-driving, and where Mercedes is responsible if the car crashes (The limitation in that case, of course, is the very limited roads and conditions where that software will actually be enabled.).

So, yes: Until Tesla (or someone else) actually takes responsibility while the car drives itself, and the "driver" is legally just a passenger, it is not a fully self-driving car. If you ever take your attention away from the road while using FSD you are breaking the law, which says that as the driver, you must always pay attention.

[wnmurphy]

> Until Tesla (or someone else) actually takes responsibility while the car drives itself

Were you aware they're already doing exactly this in Austin and Dallas?

[vel0city]

Yeah, there's ~40 of them in Dallas, and they're highly geofenced.

Seems pretty small for a platform they're so confident about.

[soperj]

They've also had multiple crashes in very few miles driven.

[stetrain]

> it's "when," not "if."

That has always been the question since Elon started making self-driving claims back in 2016. If you believed him any time up until now you'd have been wrong and potentially own a car that will never drive without a responsible human in the driver's seat.

Another question is how many resources each "self-driving" car needs to complete its task. All of the major self-driving taxi providers do have monitoring staff who need to intervene occasionally to get the car unstuck from some situations.

If you need to pay one monitor per 10 cars you can run a taxi service in specific markets but it's pretty hard to roll that out to a fleet of millions of personally-owned cars.

[vel0city]

Its so good that's why they've got tons of robotaxis all over the place.

Oh, they only have a tiny fleet in a few cities which are still mostly monitored? Strange.

[soperj]

I hope you've filled out your will.

[criddell]

Do you think full self-driving (level 5) is near? Driving a car is something we teach children to do. Surely FSD has to be getting close, no?

Is this something Anthropic, Google, OpenAI, Meta, or any of the other big AI investors are close to solving?

[soperj]

Google is the farthest down the line.

[criddell]

It seems crazy that AIs are doing all these impressive things yet struggles with something we teach children to do.

[soperj]

Not particularly. Deep Blue beat Kasparov in 97. I don't doubt that these impressive things are very very similar to that.

[golem14]

Not a fanboy, but the latest generation Tesla self driving is really quite good. I still don't understand why they still aren't adding lidars now since prices have come down a lot.

[wat10000]

You’re talking about a company that removed the rain sensor and made their automatic wipers use neural nets and cameras instead.

[golem14]

Well, it it works well enough, I don't care too deeply. But it's clear to me that the self driving on the older hardware changes it's personalty depending on something I cannot figure out (software updates? Light conditions?). From confident formula-1 driver to nervous granny. This unpredictability is very unnerving to me. Like having a teenager drive my car.

I rented a tesla with the new hardware for a week and it was awesome. But they could just update the model overnight and give me a nasty surprise. That's ... not a very happy thought.