Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Mercedes becomes the first automaker to sell autonomous cars in the U.S.

fortune.com

Post reply on HN
Print view View original

131–140 of 141 posts

Re: Mercedes becomes the first automaker to sell autonomous cars in the U.S.

#131

Post by mlyle »

kgwgk wrote: post #29
mlyle wrote: post #23

Earlier quoted context omitted.

Sure, everyone knows Waymo's schedule has slipped. I read the comment as saying that Waymo's product itself is not living up to expectations. It looks like it's slowly but steadily expanding, at least in sunnier climes, and offering a decent user experience.

Ok, maybe the as-of-2024 product is living up to the 2018-horizon expectations from 2017 (as long as your expectations were not at the high end of the range at the time).

Yah, the question is whether we are getting there. No one in 2030 will care whether it was delivered in 2026 or 2020.

Re: Mercedes becomes the first automaker to sell autonomous cars in the U.S.

#132

Post by mlyle »

SoftTalker wrote: post #88
qznc wrote: post #46

Earlier quoted context omitted.

Waymo does not sell cars. They are in a different business. I don’t really know the numbers but Waymo adds something like 50k worth of hardware into the cars. That is not a viable approach for consumer cars (Mercedes, Tesla, etc).

That hardware cost will drop eventually as it becomes commoditized and competing providers emerge. We're still pretty early in the game.

Maybe. Not everything ever manages to make it to low cost, high volume- either because there is no reasonable path to making it that cheap, or because there are not sufficiently large markets in the in-between prices.

Re: Mercedes becomes the first automaker to sell autonomous cars in the U.S.

#133

Post by bhauer »

llamaimperative wrote: post #86
bhauer wrote: post #81

Earlier quoted context omitted.

The limitations of the Mercedes system are only a little higher than the hypothetical parking lot Level 3 system. The point is to emphasize that a narrow Level 3 system is not very interesting compared to a broad Level 2 system. Technical capability > appeasing lawyers.

Uhh but a narrow Level 3 system will also in general do broad Level 2. You think these Mercedes can't do adaptive cruise control at 60mph wherever you want to activate it? I would take a car with reliable, narrow L3 and reliable broad L2 (especially when the L3 is reliable in the most frustrating driving conditions to encounter ) over broad pseudo-L3 any day. Waymos are incredible, but Teslas are a much worse tradeof…

> Uhh but a narrow Level 3 system will also in general do broad Level 2. You think these Mercedes can't do adaptive cruise control at 60mph wherever you want to activate it?

The Mercedes system will absolutely not follow navigation routes, change lanes, execute right and left turns, execute U-turns, stop at traffic controls, and so on. All of which FSD will do.

A broadly capable Level 2 system that is suitable for 99% of driving scenarios (with supervision) is way more interesting than a tightly constrained Level 3 system that is only suitable for <1% of driving scenarios.

Re: Mercedes becomes the first automaker to sell autonomous cars in the U.S.

#134

Post by adrianN »

petermcneeley wrote: post #123
adrianN wrote: post #96

Earlier quoted context omitted.

Unlikely given the computation required for this feature.

Highways are well defined spaces. You wouldnt need much computation at all.

It's the edge cases that are difficult.

Re: Mercedes becomes the first automaker to sell autonomous cars in the U.S.

#135

Post by tim333 »

harles wrote: post #8

> As of April 11, there were 65 Mercedes autonomous vehicles available for sale in California ... One of those has since been sold, which marks the first sale of an autonomous Mercedes in California ... Those are pretty small numbers and strike me more as them testing the waters. It's neat to see some potential in the space after what felt like a total collapse with Waymo and Tesla especially not living up to expecta…

Given that Mercedes is taking responsibility for accidents caused by the system I can see why they would launch cautiously. Less of an issue if it goes wrong.

Re: Mercedes becomes the first automaker to sell autonomous cars in the U.S.

#136

Post by kgwgk »

mlyle wrote: post #131
kgwgk wrote: post #29

Earlier quoted context omitted.

Ok, maybe the as-of-2024 product is living up to the 2018-horizon expectations from 2017 (as long as your expectations were not at the high end of the range at the time).

Yah, the question is whether we are getting there. No one in 2030 will care whether it was delivered in 2026 or 2020.

The people who didn't get a driving license in 2017 because what was the point if cars would drive themselves in a couple of years may care. Even more so if they finally have to wait to 2050.

Anyway, the question was about expectations - and those came with a timeline.

Re: Mercedes becomes the first automaker to sell autonomous cars in the U.S.

#137

Post by matsemann »

ec109685 wrote: post #121
ra7 wrote: post #98

Earlier quoted context omitted.

Then Tesla should take liability in those conditions, if they are confident.

I think it would be too confusing to have a third state: - manual driving - supervised autopilot - unsupervised autopilot

Agreed. Remove the useless middle one. It only adds danger when people aren't paying attention.

Re: Mercedes becomes the first automaker to sell autonomous cars in the U.S.

#138

Post by petermcneeley »

adrianN wrote: post #134
petermcneeley wrote: post #123

Earlier quoted context omitted.

Highways are well defined spaces. You wouldnt need much computation at all.

It's the edge cases that are difficult.

That is why it is highway driving only. This is no different than aircraft autopilot of the 90s.

Re: Mercedes becomes the first automaker to sell autonomous cars in the U.S.

#139

Post by freetinker »

pavlov wrote: post #63
ra7 wrote: post #57

It operates under such specific conditions, it’s not very useful right now. Unless you spend a large part of your commute on clogged California highways, there’s not a big use case. But Mercedes would be hoping to expand that operational design domain (ODD). Ultimately, I don’t think their L3 technology is the groundbreaking part (I think they use Nvidia Drive). It’s that they are willing to take liability inside the…

> “commute on clogged California highways” It’s embarrassing for Tesla because they pretended they had this solved and practically ready to ship eight years ago, started taking customer money; still take customer money on promises of “Real Soon Now”; and yet it’s a bunch of staid old Germans who actually ship it.

I don’t quite understand Tesla’s insistence on vision-only. I don’t believe vision-only “FSD” will ever be good enough for mass adoption – even if it performs at human-level. Autonomous systems need to be better than humans to gain acceptance, and that’s simply not possible with a vision-only let’s-mimic-humans approach.

Re: Mercedes becomes the first automaker to sell autonomous cars in the U.S.

#140

Post by mlyle »

kgwgk wrote: post #136
mlyle wrote: post #131

Earlier quoted context omitted.

Yah, the question is whether we are getting there. No one in 2030 will care whether it was delivered in 2026 or 2020.

The people who didn't get a driving license in 2017 because what was the point if cars would drive themselves in a couple of years may care. Even more so if they finally have to wait to 2050. Anyway, the question was about expectations - and those came with a timeline.

> Anyway, the question was about expectations - and those came with a timeline.

The mention was of a "collapse" in the space, too-- which can't be true if we're still getting there.

I expect the timeframe of 2030 is when we get most of the way there. Even exponential growth takes time to reach a big world. Most of what Waymo has left to do is rollout.

Post reply on HN