Hacker News Reader

Comment by vgchh | Hacker News Reader

Serious question: How is Waymo’s tech better?

I know Tesla’s FSD is used by people quite successfully. Don’t know if people have experiences with Waymo’s

Tesla is at the level Waymo was when they started running Waymo service in Chandler with safety drivers. In Tesla's case, you are the safety driver. That's a huge achievement, but Waymo is now running with no safety driver at all. People will say "but it's a geofenced area". Sure, but Tesla is a ways off from doing even that.

Are they a "ways off"? It's not clear to me that the two are in a linear progression.

Is Tesla even attempting tests with no safety driver? They're selling personal auto mobiles, not ride sharing cabs.

Depends on metrics you choose to evaluate. Tesla is leading on deaths occurring with FSD (many), Uber 1, Waymo 0.

Curious on the down votes, these are verifiable numbers and related deaths seems like a reasonable metric.

Probably because you ignored the fact that Tesla is the first to deploy, so obviously there have been more crashes. but on a per mile basis maybe not. It's such an oversight that it comes off as dishonest.

I can see your point and it makes sense. Counterpoint, if I had answered "depends on the metrics you choose but Tesla was first to market" or "depends on the metrics you choose but Tesla has the greatest miles driven" do you think the reaction would be the same?

Not op, but people here don't generally take to clearly misleading data that was presented that way to make a point. The population here is probably sufficiently enriched for data peeps.

And I mean this in the least hostile way, though my choice of language above may sound sarcastic.

This is the type of disingenuous analysis that greatly skews reality; breakdown for distance driven, % of road types driven on, driver neglect, etc.

If Waymo is at 0 then normalizing by miles driven doesn't really matter...

It does if they’ve driven zero miles

It matters because a low enough miles driven would show insufficient data.

Agree on distance driven as an important metrics, road type too. However, is driver neglect really a variable for self driving cars? If yes, then is it a self-driving car?

Reply on news.ycombinator.com