Prolly not the right thread for a breakout sidebar. . .
TLDR version: There are known chaos factors that AI...any presently kown form of AI...simply cannot anticipate for traffic on city streets. It's chaos that exceeds limited-access roadways or results reproducible in a lab setting, and the tech completist enough to tackle that isn't yet the stuff of whitepapers let alone production environment. "80/20" rule isn't good enough..."99/1" isn't even good enough.
Second...the production AI for self-driving today is so error-prone it trailing its own self-predicted rate of improvement by years. And nobody is totally sure why because the best researchers in the land can't fully quantify where it's whiffing. That's...um...bad for future prognostication. You shouldn't for example, have closed-course self-driving demo races where a Jetsons Shit demonstrator car decides to up and crash itself into a wall on Lap 1, like that viral blooper from last month. That was *supposed* to be a layup right now in 2020; it wasn't.
YMMV on if it ever hits that threshold. But 2035, 15 years from now? Not a chance in hell...way too over-optimistic. The best money and minds are already behind it...and it's running way way far behind that pace. 15 years is not a long time; advancement is falling too far behind a pace that'll net results anywhere close to that timeframe. That much you can predict with fast-increasing reliability now.
Key words in that post
"presently known form....".
I say this on a day in which the SECOND (in 8 days) mRNA vaccine ever know to humankind is being unveiled to the world. Where was that capability even 5 years ago? Where was Space-X? How many people paid by credit card (let alone cellphones) in grocery stores 15 years ago? How much cheaper was coal than solar back in 2005?????? Robert Murray died last week waiting for personal government financial aid. How powerful was he in 2005?
What if these automated pods are made of exteriors 3 or 4 times thicker and just as pliable as 2 liter coke bottles? If this is indeed just for urban core centers, the speed limit on these vehicles could be maxed at 15-20 mph. How dangerous is that?
Btw, everyone is focused merely on the automated vehicle. What about the OTHER SIDE of the safety equation? What about also transforming the physical aspect of the sidewalk/crosswalk with virtual and changeable barriers that can rise or fall with traffic signalling? How dangerous are today's evolving and multiplying bike lanes with the flex posts for increased safety? Too many people are thinking within concrete enforced boxes right now. The future will deem those barriers quaint.
As Fattony wrote regarding the evolution of smartphones from 2000-2015 - - 15 years is a long time these days. To paraphrase the well-worn phrase "Necessity is the mother of invention", I'd say
trillions of dollars in real estate and urban economics benefits is the mother of invention. Money talks and there are trillions at stake. It might not happen in Boston by 2035, but I'm pretty damn sure it will in places like the city centers of Singapore or Tokyo or London or Portland, Oregon. Boston, most probably, would be a middle-to-late adopter. Places like Cambridge/Somerville might go earlier. It's only an opinion, and everyone is entitled to their own, but I'm looking forward to watching this play out.
Once again, just LOOK at the picture in the OP:
Find local businesses, view maps and get driving directions in Google Maps.
www.google.com
