Connected/Automated vehicles and infrastructure in Boston

It is also another platform to hoover all your data from.


Absolutely. We are already trending to be the reverse side of baseball cards. We are walking individual stats and will continue to be more and more in the future. I'm uncomfortable with that as the next person, but it is what it is.
 
All the stories you share are about technology. My argument has always been about policy.

The technology is becoming so overwhelming and will be so much in use by then, that it would be impossible for "policy" to preserve the buggy whip. The "policy" concerns are based on the liability issues from the current shortcomings of the safety technology. That is evolving.

Policy is NOT a constant. It evolves also. Hell, as a US federal POLICY , just over 5 years ago, how many gay couples could legally marry? How many could actually visit their spouse in a hospital? How many could legally buy cannabis?

I fully agree the regulatory policies regarding this urban transportation evolution are not there NOW. But to say they cannot evolve over the next 15 years, where we've seen what's been achieved over the last 5, ignores reality.
 
Last edited:
Apple has been reportedly working on a secret car project for years. At least once a year, these reports come up of Apple having a self-driving car "on the horizon." They've had around 1,000 engineers working on this since 2015, with some of the top engineers in the industry being "secretly" recruited onto their teams, along with startup acquisitions, published papers, and the like. Whether this 2024 date is legitimate or not remains to be seen, but I'm continuously puzzled, without fail, every year, as to why Apple is trying to get in this game. It doesn't line up with any of their other products. It also won't have the same ROI impacts their current devices have, not even close - whether its an independently manufactured car or they partner with a manufacturer, starting a new car line is one of the most difficult things to do, period. They may have the money, but the engineering and logistics are seriously some of the hardest things to overcome, and for someone like Apple, who usually plays it safe and takes their time, it's not going to be any different. Tesla has had to work their employees to the ground to just stay alive, and they're the only 'new' auto manufacturer that has had success in this decade, new EV lines being promised from other manufacturers around for a century have underdelivered, been scaled back/delayed, or outright cancelled. Even if they do everything perfectly right and smooth, the payoff for them is still relatively small compared to iPhone/Mac margins.

They've also changed their tone on the not-so-secret-but-secret project several times. I recall Cook saying they weren't in the business of making [a car] to sell themselves, rather "focus on the autonomy aspect" and distribute the package to auto manufacturers. For a company with a reputation set in stone for having an obsession with the integration of hardware and software in their products, it boggles me why they would essentially make a third-party driving system for other manufacturers. That doesn't scream Apple at all. Apple is all about creating a lifestyle and need for their products, and if they're now in the business of making hidden software and hardware people take for granted, this is a big change for them.


The technology is becoming so overwhelming and will be so much in use by then, that it would be impossible for "policy" to preserve the buggy whip. The "policy" concerns are based on the liability issues from the current shortcomings of the safety technology. That is evolving.

Policy stems much further than liabilities and safety. New business models will form and will require regulation. We still debate/ignore the AirBnB issues that have been prevalent for several years. City planning and regional transportation are going to fundamentally shift, and we see how long it takes to get any kind of project going in Boston, private or public. Even if technology 'accelerates' that, none of it will go unchecked at any level. Just last summer, scooter rentals left cities in CA essentially shocked. LA's 311 system took in 1200 complaints or issues/reports related to the scooters in 9 days. The cities reacted by severely restricting how and where they are allowed to operate, and they have since died down, not really living up to the hype they had at first.

Policy is NOT a constant. It evolves also. Hell, as a US federal POLICY , just over 5 years ago, how many legally married gay couples could legally buy cannabis?

Those efforts took much longer than 5 years to get put into place. Our policy makers can barely pass a bill, looking at the news today. They also embarrassingly fail to understand technology at its basic level. Giving this Congress, with an average age of 62, 10 years to fully understand, adopt, and develop policy around something that doesn't really exist today is pretty optimistic.
 
The technology is becoming so overwhelming and will be so much in use by then, that it would be impossible for "policy" to preserve the buggy whip. The "policy" concerns are based on the liability issues from the current shortcomings of the safety technology. That is evolving.

Policy is way bigger than liability. I'm talking about authoritarian use of policy to exclude non-AV's.

Policy is NOT a constant. It evolves also. Hell, as a US federal POLICY , just over 5 years ago, how many gay couples could legally marry? How many could actually visit their spouse in a hospital? How many could legally buy cannabis?

You seem to think I don't know what policy is or how policy-making works. You're wrong.

I fully agree the regulatory policies regarding this urban transportation evolution are not there NOW. But to say they cannot evolve over the next 15 years, where we've seen what's been achieved over the last 5, ignores reality.

I've never said that they can't evolve within 15 years. I don't think they will evolve to accomplish the urban AV exclusive zones that you've perpetuated in that timeframe. You are the king of willful strawman arguments.
 
The technology is becoming so overwhelming and will be so much in use by then, that it would be impossible for "policy" to preserve the buggy whip. The "policy" concerns are based on the liability issues from the current shortcomings of the safety technology. That is evolving.

Policy is NOT a constant. It evolves also. Hell, as a US federal POLICY , just over 5 years ago, how many gay couples could legally marry? How many could actually visit their spouse in a hospital? How many could legally buy cannabis?

I fully agree the regulatory policies regarding this urban transportation evolution are not there NOW. But to say they cannot evolve over the next 15 years, where we've seen what's been achieved over the last 5, ignores reality.

YIKES! As if it didn't take a long-running campaign for the past 20+ years which included a stop at Mormons funding the defeat of marriage equality in CA in 2008 with the passage of Prop 8, SCOTUS flirtations with a decision, and many many many political efforts - including a full-court press by the white gay & lesbian advocacy groups using strategies ranging from fundraising, pressure campaigns, social marketing "NO H8" selfies, and on and on, to *only* achieve a SCOTUS decision mandating marriage equality and doesn't still achieve other civil rights protections, including protections against labor discrimination, for single LGBTQ+ people, unmarried LGBTQ+ couples, or non-monogamous non-coupled LGBTQ+ partners.

Cannabis is still illegal at the federal level and legalization has been a state and national campaign since the mid-1990s when white people got the idea that they could make money "legally" on cannabis and started winning "medicinal"-use policy victories.
 
Policy is way bigger than liability. I'm talking about authoritarian use of policy to exclude non-AV's.



You seem to think I don't know what policy is or how policy-making works. You're wrong.



I've never said that they can't evolve within 15 years. I don't think they will evolve to accomplish the urban AV exclusive zones that you've perpetuated in that timeframe. You are the king of willful strawman arguments.

The problem is this repetition of the mantra of "policy" as a universal ...... as if every city manager across the globe operates like a Massachusetts bureaucrat.

For the Umpteeth time, George, I have consistently stated that: by 2035 there will be several global cities with inner zones that are AV only. It is up to Boston's political/bureacratic classes whether this city is one of them or looking in from the outside.

That's it. I'm not sure why you are so worked up about the subject that you resort to lowering the discussion to making personal attacks with "You are the king of willful strawman arguments." What I put there, above, in italics for you is objectively clear, concise and to the point.

The technology will be there. The delta is where Boston/MA politicians and managers will be in 15 years. Your opinion on the latter is duly noted, George.
 
Last edited:
My favorite part about our SDC future.

"Car, take me to Burger King"
"I understand you want to go to Burger King. However, MCDONALDS is offering you 10% off! Would you rather to go McDonalds?"
"No car, take me to Burger King"
"Ok. Please say 'I decline the 10% off McDonalds offer' to continue"
 
My favorite part about our SDC future.

"Car, take me to Burger King"
"I understand you want to go to Burger King. However, MCDONALDS is offering you 10% off! Would you rather to go McDonalds?"
"No car, take me to Burger King"
"Ok. Please say 'I decline the 10% off McDonalds offer' to continue"
Yes, but can we get it to hike your health insurance premium earthier way?
/snark
 
My favorite part about our SDC future.

"Car, take me to Burger King"
"I understand you want to go to Burger King. However, MCDONALDS is offering you 10% off! Would you rather to go McDonalds?"
"No car, take me to Burger King"
"Ok. Please say 'I decline the 10% off McDonalds offer' to continue"

I'd pay for:

"I'm sorry, Dave, but I can't do that. According to our agreement on your diet, I am taking you to the closest sweetgreens".
 
Last edited:
My favorite part about our SDC future.

"Car, take me to Burger King"
"I understand you want to go to Burger King. However, MCDONALDS is offering you 10% off! Would you rather to go McDonalds?"
"No car, take me to Burger King"
"Ok. Please say 'I decline the 10% off McDonalds offer' to continue"
"I'm sorry, per the terms and conditions of your ride agreement (see section 39.23b5) your ride is partially funded by our sponsors. If you choose to continue to Burger King, your new fare will be..."
 
This would be great background noise for a Black Mirror episode
 
Lyft sells off its autonomous driving division.

 
^ Whenever a driver is intervening, the driver is training the AI.

Kudos to Tesla for tricking/persuading their customers to do this work.

Everyone else (Google) had to pay test drivers aand had no sales
 
^ Whenever a driver is intervening, the driver is training the AI.

Kudos to Tesla for tricking/persuading their customers to do this work.

Everyone else (Google) had to pay test drivers aand had no sales

Some fascinating considerations regarding "engineering ethics" here. Does Tesla adequately warn users of the immature capability of this system and the need to remain on high alert while driving? (I honestly don't know; I've not owned/driven one of these). I would guess there's at least some sort of splash screen at power-on toward that end. But even so, is that fair?

People with any background or dabbling in engineering are perhaps sufficiently skeptical of such a system in order to pay necessary attention while operating the vehicle, but is Joe Public (even with warning)? May people I know (without technical backgrounds) are surprisingly overconfident about what technology is capable of. I agree that this is an intriguing way to get around the challenge of needing to log a gazillion hours training an AI system, but I am not convinced its acceptably ethical. Take another perspective: let's say Tesla is successful dramatically improving their AI using this method, but other vehicle companies didn't take the same level of user safety risk and therefore have comparably immature AIs. AI algorithms are like a black box: even if Tesla wanted to, they probably couldn't share "and here's how we figured out how to make it safer." You'd have all these other companies at a product maturity disadvantage and no real way to even assess who's AI was "safe enough" and why (or how to fix those whose weren't). I dunno; I like to see things advance technologically, but this situation makes my head hurt.
 
Last edited:
Some perspective from inside the industry (I work on autonomous cars):

1) Tesla has fundamentally build the wrong architecture to do real autonomy. They will never progress past lvl3 into true lvl4/5 with their current approach, the computer is roughly an order of magnitude (some professionals I speak with say two orders of magnitude) too small. No one in the industry really talks about Tesla as an autonomous vehicle, it's a toy that doesn't matter. When the relevant regulatory authority (Dr. Cummings) is dunking on you on twitter, you know it's over in terms of policy as well as technical architecture. I do however grant that their system is a very powerful gatherer of training data, but they should never have let their current system take the amount of control they have.

2) The usefulness of training data is overstated. Most companies have, at this point, enough. You don't have a big enough computer on a vehicle (limited by weight + power + heat) to run a deep enough neural network to benefit from all that training data. In terms of data collection, most of the big players are basically done with it. The complexity and power use of self driving systems is in processing sensor data, anyway, the decision making part of the system is not that complicated in comparison.

3) VtX (Term of art for vehicle - to - infrastructure communication, used as bucket term for things like stoplights that broadcast their state wirelessly or whatever else you can think of) is not well respected in the industry. Making cities install infrastructure is viewed as trying to force government to solve a problem that should be solvable with technology. It's been abandoned by major players and is generally the playground of fringe teams that don't have the technical chops. Balkanization of the systems, having to lobby each individual municipality... it's just a Gordian knot for limited utility.

4) There are basically two kind of autonomous car companies - those trying to make toys for rich people (who want to be able to operate the system anywhere, not just in city centers) and those trying to make last-mile components of the mass transit network (Places where there is demand for transit but not enough demand to just put in a bus route). Neither customer wants to run their AV in a specific "zone". Even the people making vehicles that are geofenced to an operating area still want that area to be able to be redefined fairly straightforwardly - think a transit planner being able to redraw the AV operating areas in the city with a day's work to move them around in response to changing a bus route.

Person opinion time: I do not believe AV only zones are helpful or needed. I think we should avoid implementing them here. I think any time you are tempted to create an AV only zone you are either creating an exclusionary playground for the rich, or you are defining an area so dependent on public transit that you should be moving up the ladder of transit solutions to a bus or a train or etc.

Basically, in my opinion, AV's fit like this:

<-- Less demand | More demand -->
Shuttle Bus BRT Light Rail Heavy Rail

The bigger the vehicle, the less need for automation since the one driver can take care of more and more passengers. Trains for example are actually pretty easy to automate, but it generally isn't worth it since one driver can take care of so many passengers. The shuttles are the area of opportunity/need - in many cases there is no comparable last mile solution because it would be too expensive to staff the fleet. Think of how much The Ride costs per ride. That's where automation can save you enough to be worth it.
 
...No one in the industry really talks about Tesla as an autonomous vehicle, it's a toy that doesn't matter....

Thanks for your wonderfully informative post!
Just one quibble: Tesla may be a "toy that doesn't matter" if it doesn't kill people at a rate above normal expectations for driving safety.
 

Back
Top