Then it detaches and becomes a drone that watches over your car when you park.
Google has had LIDAR scanners on their streetview cars for 8 years.
1 Like
Any idea what they are doing with the data?
Ground truth - streetview’s pretty photos are not the main thing. They are indexing the world. Ground truth!
1 Like
And currently fighting over LIDAR technology:
“Waymo, the self-driving car business of Google parent Alphabet, accused Uber in a court filing on Thursday of stealing its lidar designs. Former Google engineer Anthony Levandowski downloaded 9.7 gigabytes of files before he left to found a startup that was subsequently acquired by Uber, the company said in the suit.”
Apple’s car top LIDAR setup is still kick ass.
1 Like
Urm… Think bigger. Transportation is a huge industry, and even if self-driving vehicles only shave off 1/5, that’s still millions of jobs that no sector is prepared to soak up; almost everyone is shedding jobs right now, and any assault on the big remaining sectors (transport, health care) is going to just start jacking up the unemployment indices.
Don’t fall for the trap of “well it seems hard so it’ll never happen”. Driving is 99.9% easy, with 0.1% of complex edge cases. The first 99.9% is pretty well solved at this point; we’re just trying to make autonomous vehicles a little bit more reliable than humans for that last 0.1%. And if you’ve ever driven in a big city, you’d know that beating humans at split-second, high-stakes decision-making behind the wheel isn’t actually that hard.
So, yeah. There’s a lot of really, really smart people working on this problem, because these companies are all drooling at the massive amount of dollars waiting for them in the transport sector by removing a chunk of humans from the equation.
Example: A 99% autonomous semi truck would replace a “team”-driven semi (where drivers take turns to keep the truck in more-or-less continuous operation) that averages $100-150k per year. Let’s say that you’re right, and that occasionally human intervention in edge cases is needed. Well, now you can put a decently paid guy in by telepresence and have him intervene on any vehicle that sends him an alert (think like how a drone is remote-piloted, but in this case it’s only for five minutes like every 50,000 miles). So even if this guy is paid well, say $125k a year, he’s replacing dozens or hundreds of trucker teams, reducing labor overhead to a single-digit percentage of what it used to be… And putting most long-haul truckers out of work in the process, because any company that doesn’t do this will get underbid and run out of business by the ones that do.
2 Likes
When I’m testing an autonomous driving system to ensure it doesn’t end in a wildly destructive fiery crash, my first thought is: “Let’s strap it to a lexus… no need to waste a perfectly good kia.”
Except for mysterious holes like Francisco Street in San Francisco.
Hmm, I think @tgarretteaton is actually thinking bigger than you here. You seem to consider the status quo and how it will evolve once autonomous driving will take off. @tgarretteaton (like me) instead argues that self-driving cars (btw. do we need to call them auto-auto-mobiles?) will not solve the main problems we have with the status quo of transportation, namely congestion and resource wastefulness, so we urgently need a system change towards proper public transport.
You are very right to point out that money currently flows in another direction, and that stakes are high especially in the goods transport sector. I hope that all that research and development at some point will also help to get better public transit systems.
I always wonder if that telepresence intervention will really work out. My impression is that in driving, the 99.9% “easy” conditions can turn into the 0.1% “complex” conditions in any moment, within a second, making a split-second reaction necessary. Telepresence won’t help for these kind of situations.
1 Like
Sorry, I didn’t think that I would need to point out that self-driving cars will almost inevitably shift vehicle ownership away from individuals and towards companies or what basically will amount to local cooperatives.
After all, we don’t need our cars 90-95% of the day, except for a few very heavy drivers (sales, delivery) that are mostly oriented towards business applications. And some overcommitted soccer moms, I suppose.
Anyway, if you can just schedule a car pickup for your normal work times and add in pickups as needed, with premium prices for “on demand” driving at peak times and so on, you allow the market to load-balance and determine a more optimal number of cars on the road, massively reducing waste, and at least in theory reducing cost to the consumer as well.
As for the telepresence thing - it’s not for split-second stuff, it’s more for high-urgency “Help, I’m stuck in some weird situation that I don’t recognize and I had to come to a stop or enter a holding pattern, I need a tie-breaking human vote in the next minute or so” kind of things.
1 Like
This topic was automatically closed after 5 days. New replies are no longer allowed.
