Hacker News new | ask | show | jobs
by mavhc 1062 days ago
Everyone else's rockets don't land after putting things in orbit.

You can't buy another car that can do even 10% of what FSD beta does

2 comments

> You can't buy another car that can do even 10% of what FSD beta does

That's an odd statement, as "the first self-driving system to be approved for European public roads" and "the first automaker to receive government approval in the US for a Level 3 driving feature" is from Mercedes-Benz.

Tested and approved >> beta hype.

https://europe.autonews.com/automakers/mercedes-opens-sales-...

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

"This geofenced Level 3 system works at up to 40 miles per hour on select highways"

It's just worse cruise control, can't even go at highway speeds, there's about 1% chance you could use it.

Meanwhile FSD Beta is driving around cities the same as Waymo/Cruise are, except every city in North America.

We know a few things about AI, it needs lots of data, and lots of CPU. Tesla has 2 million cars with 8 cameras gathering data, and a top 10 in the world supercomputer.

> Meanwhile FSD Beta is driving around cities

Beta.

We know a few things about Musk: it needs hype, smoke-and-mirrors overpromising. Get back to us when it's out of beta.`

Tested and approved >> beta hype.

But in the end it works out.

"During the second quarter of 2022, SpaceX delivered 158.7 metric tons to low Earth orbit which is four times more than second palace China’s space corporation CASC at 38.8 tons.

Roscosmos at 17.2 tons was third, United Launch Alliance 4th at 13.0 tons, and Arianespace at 9.8 tons was fifth."

Telsa has the best selling car in the world, while Ford loses 200% on each EV they sell (according to Ford)

> But in the end it works out.

With Tesla, I think that this very much remains to be seen, and signs are not good (1)

M-B is at least making verifiable progress

1)

https://www.independent.co.uk/tech/tesla-elon-musk-self-driv...

https://www.motortrend.com/news/tesla-fsd-autopilot-crashes-...

Are there any reports on the results of NHTSA investigations instead of just saying they've been started?

Doesn't seem that bad really, 2 FSD reports, first one FSD braked suddenly, maybe, no results from investigation it seems so far. And people driving incorrectly behind them were too close and couldn't brake on time. If only all the cars had FSD and not humans.

And the other FSD report was about driving into a junction where the car they thought was going to hit them had already stopped to let them go, and beeping to warn it was getting close to a cyclist. That's why it's still beta and warns you every time to be aware and it will do stupid things. So far in 300 million miles driven, it's apparently hit 0 people. So that's way better than humans, who kill someone every 100 million miles.

> You can't buy another car that can do even 10% of what FSD beta does

Right, because Tesla is the only company that sees individual car owners as the target market for self-driving, everyone else in the business sees institutionally-owned robotaxi fleets as the target market. Actually, so does Tesla, fairly overtly—though they are behind at actually having auch a thing—they just see individual vehicle owners as a way to defray costs, especially development costs.

They aren't ahead at self-driving, they are just more creative at how they are financing it.