Hacker News Reader

Hacker News

Comment by 40acres | Hacker News Reader

Parent Full thread

40acres

What an incredible post. I just love this guys blog, I wonder how long it took to research all of this.

As a side note this gives me great confidence that Waymo will come out of the self driving car race in pole position.

View on HN

Yes !

Each time justin writes a blog post, I feel like a cartographer for a couple of minutes.

I love the attention he puts to the evolution of mapping services.

The only better thing would be the same blog as a post mortem from Google where they discuss the reasoning, design and tech behind these changes.

Up to a point. At that point an autonomous system needs to be able to interpret and react to the physical world as it actually is at the moment as opposed to how it's supposed to be.

This is why efforts based on exhaustive mapping make me nervous. Things change, sometimes rapidly. The vehicle should be using its maps as a general guide, not as some kind of ground truth.

The vehicle should be using its maps as a general guide, not as some kind of ground truth.

Not to sound like a jerk, but I would be shocked if there were anyone working on this that was not pursuing things in this way. It just seems kind of obvious that you can't have some big hunk of metal rolling around, following some abstracted track from a map, without "looking" where it is going.

And yet we have Waymo building 2cm-resolution maps and not operating outside of mapped areas, and (more scarily) Tesla geotagging false positives where Autopilot misidentifies some roadside feature and panic brakes.

Are those approaches mutually incompatible? It seems to me that it would be easier to build a system which can drive fast when what it's seeing matches what it's expecting to see (most of the time), and which falls back to a much more conservative stance in the case of surprisal (i.e. expectation not matching observation).

If all the cars are always mapping and collecting data will maps ever be out of date?

Think of all the many times a software developer says "X will never happen."

Then think of what percentage of the time they're right.

Yes. Cast your mind to lightly-traveled rural roads. If you are driving one of the many unpaved roads in Arizona during flash flood season, and the last car through was a couple hours ago, will you bet your life that the road is still there?

But hey, that's not a realistic danger in the city, right? Well, look up the sinkhole named "Steve" that opened up in the middle of an Oakland, CA freeway.

Then there's Highway 1, which is known for landslides. Not to mention that little incident with the Bay Bridge back in '89...

By definition, any car that makes an update to a map has experienced an out of date map.

Cars obviously have sensors for real-time data capture. However, the question is how well they handle new signage/lights, road changes, construction, flooded lanes, etc. when those aren't in the database. It's fair to say that autonomous driving today is some point between "we run 100% off maps" and "we can handle things like a human could if we lost access to our maps." (And much closer to the former.)

Good point, the map is not the territory and the cars are engaging with the territory.

Another interesting thing to consider, should there be widespread use of autonomous vehicles is the absolutely massive deployment of digital sensors that would entail, and who would use this data and to what ends? It would basically be realtime Streetview.

There's a fairly steady stream of posts here about data/privacy/etc. It's easy to pillory companies for violating privacy rules in various ways or mishandling data. However, given that data can't be reliably anonymized and that a lot of things that happen in public have simply never been systematically recorded in the past, there are a lot of fundamental questions over what sorts of data should be recorded when they provide legitimate value to users.

As a hypothetical, what if autonomous driving is seen to depend on what many would see as invasive data collection and monitoring? Should it be permitted? That's a hypothetical, but I don't think there's an obvious answer to the question as posed.

I feel that in the long run the answer is going to be heavily biased towards automated driving - unautomated driving kills. Every year more than a million of people worldwide die for no good reason, just because we suck as drivers; currently automated driving isn't yet superior, but in the long run it will clearly be so.

I don't think that we'll ever justify that massive loss of privacy is more important than massive loss of life and health; especially if the bad consequences of loss of privacy is generally hypothetical and in the future; and the bad consequences of traffic accidents are obvious and immediate.

If it's the type of autonomous cars that are prototyped today, they are definitely going to be dependent on invasive data collection.

Does it get sent back to the mothership? How is it processed? Is it discarded? Who has access?

When data is the new oil, these questions answer themselves if there is not legislation to prevent it.

Such data would be hugely attractive for commercial interests.

Hedge-funds and the like, use aerial imagery to get data like the amount of cars in parking lots at shopping areas.

With a fleet of cars with 360 degree cameras driving around, not only could the number of people on the sidewalks be counted, but it would be technically possible to identify most of them.

Reply on news.ycombinator.com