Our daily life seems to increasingly revolve around smart phones, and soon smart houses and smart cars, but what about the towns and cities we live in? Some months back, I was at an event where a lot of discussion was taking place about urban planning and the various technologies we might incorporate into it; everything from vehicle charging stations to phone charging stations, wifi coverage, and many points in between. Ironically, for someone who dislikes living in cities personally, preferring my farm here in Plymouth, I always find city management and planning fascinating, or at least I have since playing SimCity on the Super Nintendo a few decades back. So the conversations on urban planning at that event caught my interest and I heard the term smart city getting used there, which I’d heard in-passing elsewhere in recent years but had never really heard it defined. I asked some folks for more info on the general concept and I got a lot of fascinating
but also often vague details and explanations, some nt matching others. Needless to say, our goal in today’s episode isn’t really to try to better-define a term that’s still rather embryonic, but rather to discuss the concepts under its umbrella and what we’ll see coming of it in both the near and far future. The term “Smart City” has been growing in popularity in recent years, doubtlessly due to the influence of the word smartphone, and while many of the concepts involved are older than the widespread use of cellphones, it seems that the new term is here to stay. A quick check of Wikipedia, which currently lists only 4 US cities including Columbus, here in Ohio, as Smart cities, gives us a rather complicated opening definition, seemingly heavy in technospeak that I will boil down as saying “Smart cities are those which collect and collate data and then use it in their management”.
Which, of course, pretty much defines any city or village or town that actually has someone managing it, who is not utterly incompetent or lazy. Of course, you collect data about your town and use it to make decisions from. So, obviously it’s a matter of degrees here and as for the decisions, we might throw in some automation too. As a fairly simple example, conceptually, traffic lights that monitored flow rates based on the time of day and week and were able to adjust themselves to optimize traffic flow, would be an example of a smart city device. Taking that the next step up, we could have a timer on each that nearby cars could detect and set up on their console or HUD automatically, giving the remaining countdown before it turns, and we could also have ones that we could inform of upcoming roadwork or big events, like a festival, which would alter those ideal timings. But the back end of that is that you can be using all those traffic counts and times to be predicting when you’ll need to repair roads or expand existing segments. And of course, that’s something civil engineers have been doing
for years, but the more improvements in making data-collection up-to-date and cheaply-collected, the better. This is also where we start seeing AI having a role, as, while AI is now finally in the public awareness for its chatbot ability as a real thing that exists, not technology of twenty years from now, it is generally much better for applications like real-time traffic-monitoring and learning how to predict and optimize traffic patterns, than it is at writing a song for you. Right now, to avoid smog, for instance, we optimize around minimum time spent just idling, which is all for the good, but it's less of a concern for electric cars or those which automatically shut off when not moving. Situations change a lot if we decide to start using more automation in cars. I don’t like the term self-driving car, it’s not really accurate anyway,
plus it bugs people, but having an auto-pilot that can keep you stay inside the lines doesn’t really seem to bug folks any more than cruise control, and our real objective is to allow maximum safe and convenient driving, which, at least for the immediate future, is always going to be at its best when any onboard computer is merely assisting a human, as opposed to being in direct control. I think most folks want the robot that’s able to slam the break for the dog or kid running into the road, and that notices when your eyes are off the road and can beep at you, or just keep things rolling-on, not the machine equivalent of a chauffeur, though many will like that too, and in time, that might become the norm. For now though, it’s all about how much automation we actually have in terms of capability and price, and which ways it can safely, and non-irritatingly, be incorporated into our driving and public transportation, or even our walking and biking. And to communicate to the other devices around it, so that your car can see a kid running out into the road, but the kid’s smartwatch also warns the car and so do the motion sensors on the streetlight. Redundancy and alternative avenues are critical to safety and automation can really help with that. That’s even more the case for cities, so, as we contemplate smart cities, and here on SFIA at least we are going to contemplate actual AI potentially being incorporated as a city, we need to ask how much autonomy is available, preferable, and agreeable to both the city’s administrators and the citizens in it. In an
ultimate sense we like the idea of a city where strips of grass, trees and flowers are all over the place and carefully tended by robots, so it presumably costs less and can be done at night when it won’t interfere with traffic. In the nearer term though, we could imagine the monitors on a stretch of nicely tended green noticing how often folks slowed or paused there compared to other places and what they most took interest in. A slightly more advanced version might also anonymously query the smartphones and fitbits of passersbys to see if the passage or pause there had lowered their stress or anxiety. And in more complex terms we might even see data sufficiently utilized to determine if a given greenspace or piece of public art had contributed to lower crime rates, or stress rates, or even heart attack and divorce rates. Needless to say, that does raise some ethical concerns about data-use and privacy, and I don’t want to spend too much time on that today, not because it’s not a major deal, but because it’s an issue in basically every part of life these days and cities aren’t special in this regard compared to other entities. Your local city government either is or isn’t tyrannical already. Such technology, like many others, is merely a tool for enhancing what can be done, which includes brutal police-state oppression, if that’s on the agenda. There’s a tech-tree cutscene for the Self-Aware
Colony in Sid Meier’s classic video game ‘Alpha Centauri’, where someone is spraying graffiti around, about needing to dissent, presumably to the government, where the city basically stalks the guy in the dark of night, neutralizes him, and erases the graffiti, all without witnesses. That always stuck with me when I first read 1984 some years later. That really isn’t all that impressive on the Big Brother scale of what an AI running a city with full decision-making ability, with cameras and drones all-over, might be able to do either. So basically, if your city council, mayor or dictator is the sort of person who read 1984 like an instruction manual, you’ve already got problems, but if not, then AI with lots of public transparency, carefully-considered use, and healthy caution, can be an amazing tool to make cities safer and more pleasant places to live, and with cost-savings to boot. Nor are these applications limited to large cities, many will work well in towns and villages. Going back to the more pleasant aspects,
like adding green spaces, and other beautification options, this is where we have the concept of “Third Places”, which include public-operated space, like libraries and parks, as well as private businesses like cafes, churches or gyms. The notion here is that it is a social surrounding that’s separate from either your home or workplace, those generally being considered the first and second place you spend time at and socialize in. The term is relatively recent, having been coined by urban sociologist Ray Oldenburg in his 1989 book “The Great Good Place”, who passed away shortly before I decided to do this episode, at the age of 90. Nonetheless it can still feel a touch dated with so many of us working from home and using virtual places for socializing, as well as the growing use of third places for doing work. Going to study in the library is nothing new for instance, but by and large, most jobs of the past did not really have an option to go sit in the sun and relax on a bench in a garden park while working. Even just taking your cellphone to the park to make phone calls, which was new and uncommon tech when the term emerged, could be difficult due to road and crowd noise. Now we aim to make third places, publicly or privately owned, the sort of place where you
can spend tons of time while socializing or working or both, and they can help a person feel more like a part of their community as well as being a vent from home and work. Needless to say, they tend to attract people and businesses to come to such communities, when well-done anyway, though obviously at a cost. Balanced right though, it’s the option that lets you decrease office tensions by encouraging employees to take working breaks to the park where there’s wifi, and lets parents have some place to take their kids to play and socialize, and a thousand other things including everything from local theater groups to monthly flea markets.
I don’t want to overly focus on greenspace here, or even third places in general, but greenspaces tend to be reliable, cheap, and universal options for third places, at least when the weather is nice. Cheap is relative of course, devoting a few acres to a football or baseball field in the middle of a city is obviously not cheap, and even if you’re just removing a vacant and low-interest lot, like an old broken-down warehouse, you still have to mow the thing and it’s not generating you much direct tax revenue like a strip mall or some apartment buildings would. It can indirectly of course, by raising local property values and encouraging folks to move nearby. But both are much more expensive than, say, a basketball court, which take up way less space and are easier for a city to find room for, and which also don’t need to be mowed, or sprayed with pesticides or fungicides. Of course, robotic lawn mowers and hedge trimmers – and yeah those
are a thing now too – alter that paradigm. In a way this isn’t new either, the village green everyone could use and which festivals and games could be held on is older than dirt. No doubt we’ll also see them in early space colonies too, especially when a dome on a planet or garden spot on a space habitat is still incredibly expensive and so needs to be a shared space. It is really easy to talk about adding parks to towns or rooftop gardens but it’s generally a lot more complex than folks tend to think, and a lot of that is maintenance and administration, and keeping up with changing tastes and moving populations, not to mention managing resistance to doing so. Having good data as to what folks like in a given park, and what attracts them to other parks, is handy for improving your parks but equally handy for convincing local community members that your planned overhaul is justified. If you want to replace the old baseball field with a skatepark and lots of cool marble benches and herb gardens it helps to have some actual data to show the folks who have lived there their whole life and grew up playing baseball on that field, and aren't anxious to see it changed. Same, if you’re trying to repurpose some
old site with neighborhood garden plots. Not only does that data help decide if that’s warranted and remains so, but technology and data-collection make it easier to set up soil-monitoring and automatic watering and mowing, and having a shared tool shed that doesn’t turn into a cluttered mess, full of everything but the tools it was supposed to have which have all been stolen or borrowed indefinitely. Again the modern notion of smart city isn’t really about robots going around fixing stuff, it’s about gathering data to make better decisions with. But that is likely to be a path we go down with further automation, and as an example, most cities love the idea of repurposing some derelict, ruined building into some greenspace, like community vegetable gardens, but you have a big problem of maintaining plots not in use so that they either look nice, or at least not be a giant overgrown mess of weeds which some skunk decided to claim as its home, and enforcing reasonable standards or having some robot that tills patches and sprays them with some nice pollinator-friendly seed mix of flowers while vacant is definitely a plus. Especially in terms of headaches since your overseers can just shrug and say “I’m sorry ma’am, the robot is just set to turn a garden plot over if no one has tended it in a year… didn’t you get its automated texts and emails?” Greenspaces for play are hardly the only options of course, and they’re also not really very vertical. A park might have some up and down elements but is generally just one level, not several stories like the buildings near it.
This is also much of the appeal to arcology towers, the giant super-skyscraper concept where the inside might have whole levels of parks and gardens and even farms. For further discussion of those and adding verticality to cities, see our Arcologies and Arcology Design episodes. So, what are some of the goals of smart cities? These are inevitably going to be a bit vague, and variations on improving the wellbeing of citizens and the economic health of a town, with data and communication, as it really is just how to better-use computers and modern technology to help in planning and running a city. This is going to involve a lot of the Internet of Things, because it’s about integrating a million and one types of gadgets into the wide grid of that city. I’m not sure if there’s a specific term for something like a blend of noosphere and internet at a city level, maybe noopolis, but the keystone on a lot of this is making those integrations fast and seamless. I’m going to guess the big one in another decade
or so will be how to create a shared augmented reality at the city level, probably with several map layers, that is useful to citizens, administrators, businesses, and visitors. This strikes me as something that on the standardized software side of things, might be on the Microsoft Office, Google Search Engine, Amazon Marketplace, Facebook level of cultural impact for any software engineers looking for long-range, potentially super-profitable projects, especially given that the clients are likely to be public entities. Though I suppose it’s worth noting one way that data and AI might help run cities smartly is to help them find waste and fraud, including some piece of software they have an expensive license for that barely gets used. We might instead ask what some of the examples are of cities doing this and how those might be realized. As I mentioned near the beginning, one of the few examples of Smart Cities I could
find was my state’s capital of Columbus, and if you’ve been there, it is a pretty-well-laid and modern town, but the articles discussing why and how it is a smart city seem to mostly focus on attempts to best locate 300 charging stations for electric cars, principally on a grant from the US Department of Transportation, and some forward thinking about how autonomous cars might play a role in the near future. My own bet is that the main role autonomous cars, ones intended to drive themselves entirely and possibly without someone in them, in the near future, is to get themselves banned very nearly everywhere. I think it’s a big lift to get people okay with their kids’ school bus being driven by a robot, and if you can make that work you can probably sell people on automatic freight trucks and deliveries very easily. Many places have a heck of time finding enough school bus drivers these days, and I suppose if parents could access the cameras such a vehicle would presumably have, they might feel better.
But that whole area has a lot of room for automation and AI, be it bussing routes or snow plows, and I would not be surprised if cameras which parents could view on their baby monitor app started showing up on buses and at the fronts of schools for dropoff, complete with facial recognition software to let you know your kid got dropped off. Facial recognition software at schools in general, and at parks, is probably a this-decade technology, though depending on its rollout it might be beloved or spark outrage, suspicion, or all of the above. I’m actually betting that a lot of smart city tech is going to be stuff the public initially hates and demand gets removed, and has to get a decade of PR rehab and rebranding before being gradually and systematically reintroduced. Notice all those traffic cameras we don’t have in most American cities that can save a lot of money, make the roads safer, and free police up for other tasks. I remember over a decade back, the nearest city to me spent a ton of money
putting them in, and a citizen-driven petition and election promptly saw them all removed. As another example of that, one of the other ‘smart cities’ I saw listed was Santa Cruz, specifically for its use of data for predictive policing, to figure out how many police should be on shift and where police officers should be waiting when not responding to crimes. The basic method was to generate by analysis some spots where property crimes were more likely and have a handful each day an officer would be at. It was discontinued back in 2018.
Again, I expect that the job of making folks feel safe and comfortable with a given new bit of smart city tech, especially the ‘data collection’ part of that, is going to be as big of one as the actual engineering of the technology. New York City is unsurprisingly a place trying its hand at smart city initiatives, and the big one there is LinkNYC, which actually builds off an effort to replace the 10,000 or so remaining payphones… I assume they used to have way more too… and it’s weird to think back to a time when these things were everywhere and also to realize that payphones, collect phone calls, and the battle between AT&T and MCI and various other less-known long distance providers belongs to another century now. But, the Big Apple started converting those into wifi hotpsots and USB charging stations with big screens almost a decade ago, and we have certainly seen variations of that in many other places. In the city, being able to stroll along the sidewalk while maintaining high-speed wifi access is getting to be as important a bit of infrastructure as those actual sidewalks, much as last mile broadband rural internet has come to be viewed in recent years as akin to getting telephone and power lines out to everyone. Even ignoring Starlink and parallel options, my guess is that we’ll also see state initiatives to get not just every city with public wifi, but maintain some degree of it along the major roadways and highways. It’s just too critical to commerce, or will be, and that sort of easy and everywhere wifi is critical to any smart city design too. If wireless power transmission makes much progress,
it may also develop along those lines, people able to have their devices powered or recharge in their pocket all the time, no plugs, which might require some usage monitoring, unless we have energy abundance. Pondering how that affects the local economy gets tricky, as it also starts opening the door to drones that can hang around all the time, smartly controlled to avoid collision by the local grid, which also runs power into them. Options continue for things like determining how much power you need and how much production capacity you need for that, or for water and sewerage, or what your optimal garbage and recycling pickup paths and frequency are.
And also where you might be needing two different systems, rather than something one-size-fits-all for different parts of the city with different layouts and densities, which is easier to have if an AI is doing a lot of the basic administrating, so you don’t have the extra administrative burden of using multiple methods. On the administrative note, and as a last near-modern topic before we jump ahead to the more distant future, I think we should assume an increasing amount of virtual spaces for public offices like the BMV. Why stand in line to do some paperwork that’s mostly filling out your name and address when a mix of modern tech and augmented reality and maybe ChatGPT-like interfaces can handle almost all basic clerk-at-the-counter interactions, so that one person can handle many times the basic clerk work.
I had to get a minivan recently to accommodate the trio of new kids we adopted, and the various nieces and nephews who often come over, and I was at the BMV registering it and have no complaints about the process. It’s faster and easier than when I was 16, but I was struck by how easily that could have been done from my home PC or phone if we had a fairly seamless and trusted way to handle the whole “Need to see some ID, fill this out” part, and the inevitable “Can you explain this bit here on the form?” part. In some ways, speed isn’t a virtue on upgrading these things, there are hardworking folks who have been clerks their whole career and the cost for retraining and disruption to their lives is often cutting heavily into your apparent savings for some new tech, but I think this era is definitely coming, where an awful lot of government offices close down physically and shift more and more virtually. This might end in something like the local 24-hour county community center as the place to go if you have something very atypical or just can’t figure out the tech remotely, and most every other office is done from home, or third places, or whichever. It also means you can share specialized personnel easier with neighboring towns. Needless to say that alters your city planning a lot, when the office of city planning is no longer at city hall and city hall is more like a museum.
I’m not expecting these upgrades to come very fast, but they’re too valuable -if you can make them work- not to try to implement, and as improvements come along, they should get better, cheaper and easier to make people comfortable with. Where that shifts over to becoming a city hall that is actually an artificial intelligence gets a bit blurry, but I suspect folks will find letting a computer run their road crew dispatches is going to come easier than letting them run your tanks and fighter jets. I would not expect android police to be a thing, especially for beat cops and crowd control, but the AI detective might be a thing. Probably not the way we see with Eljiah Bailey and his humanoid android partner R. Daneel Olivaw in Isaac Asimov’s classic robot novels, but AI is already
making its way into detective work and of course, computers have been used in that role since before there were home PCs, let alone smartphones. Would you have an actual AI as the incorporated mind of a city? Maybe, something like a Genus Loci, the spirit of a place, that we often see in some religions or mythologies, and which we have discussed as an option for maintaining space habitats, especially intentionally unpopulated ones like nature preserves. More likely, I think, than an actual AI that thought of itself as New York City or London, you would probably have something more like Department Deputies, the AI that helps the engineer’s office for a town and might slowly evolve to become the actual chief engineer. In a lot of counties in Ohio and other states, it’s not unusual for there to be an elected department head and a deputy who does most of the work but avoids most of the credit or blame, and I could see this role shifting to something an AI is filling. Indeed, I could imagine pathways to futures where the AI is regarded as a person and has been doing the job for decades and just inherits the role, or even gets elected to it. The notion of an AI who ran
the road department in a small town deciding it wanted to move up and run for office in a bigger city is an interesting one to contemplate. One could imagine a dozen or more different department AI who could be under a human city council or mayor, who may or may not be entirely human themselves anymore, lots of cyborg and transhuman options on the table unrelated to all this that are likely to play a big role anyway. Maybe they can engage in a temporary hive mind, like the Unimind we see with the Eternals in Marvel comics, when a big decision needs making. Those might be changing too, where you annex some ward of neighboring city that had its
own AI that now moves into the mix, or instead of a city hive mind, the twenty AI in the region who all handle power distribution or water or sewers have to bump heads, or merge them, to plot out where to build the new water treatment facility or shift the load while one is being upgraded. Smart cities, or smartly-run cities at least, will grow and others will probably not, or even shrink, and so I think we will see ourselves moving more this way because, again, fundamentally it is just too useful to use AI, automation and data collection to run things and get better results, and to offer more options to citizens, and by and large, it’s results and options that people care about when picking their community, and when picking who leads or manages it. One thing seems sure though, whatever the role of AI and the internet is in cities moving forward, it is going to be a big one, whether the city is a big one or a small village. So today we were talking about Smart Cities and the idea of making them just generally better places to live and work in, to have your home or your business. A lot of that is about being more
efficient and organized, but just as much is about making communities people actually love being part of. It’s not just about finding a place or job you enjoy, it’s about finding something that makes you feel positive about doing it. Every job has good and bad days, but if you love your work and feel it makes a difference in the world, even hard days can be energizing, because you know it matters.
Most people spend about 80,000 hours of their life in their job, 40 hours a week, 50 weeks a year, for 40 years, and would prefer that time helped the world. 80,000 hours is a lot of time, so it’s a good idea to do some research on how best to use it. And that’s where our friends at 80,000 Hours come in, they’re a non-profit that spent a decade alongside academics at Oxford University researching which careers have the most impact, and all the research is on their website, for free. 80,000 hours website includes a lot of tools
and resources to help you find your path. Problem profiles, aiming to answer the question ‘what are the most pressing problems facing humanity today?’ Career reviews, which discuss which career paths they recommend if you want to have an impact, and lots of decision-making tools, such as their 8-week career planning course. They also have a jobs board, that’s curated & constantly updated and list of hundreds of active job openings, that they think might help you have a high impact, and you can filter those by job location, role type, job requirements, and what problem area they work on. Scanning around, there’s everything up there from Nuclear Physics Jobs at MIT or AI research in Silicon Valley to city management or ecology roles, from Intern to senior directorships. Potential careers for a wide variety of different skill sets and career stages. And again, 80,000 hours is free, completely, if you looking for a career that works on one of the world’s most pressing problems, sign up now at 80000hours.org/isaacarthur
So that will wrap us up for April, and I’m heading off for an anniversary trip with my wife this weekend so apologies in advance if I’m delayed responding to comments on the episode. We’ll back next Thursday though for two of our favorite topics, the Fermi Paradox and Megastructures, as we contemplate Dysonian SETI, and how we can search for Dyson Spheres. Then we’ll look at a lot of the common misconceptions about Space, Life, the Universe, and Everything, on May 11th.
After that we’ll have our Scifi Sunday episode, on May 14th, as we explore the grim realities of super-urbanized Hive Worlds, then we’ll have its companion episode, Hungry Aliens, on May 18th. If you’d like to get alerts when those and other episodes come out, make sure to hit the like, subscribe, and notification buttons. You can also help support the show on Patreon, and if you want to donate and help in other ways, you can see those options by visiting our website, IsaacArthur.net. You can also catch all of SFIA’s episodes early and ad free on our streaming service, Nebula, at go.nebula.tv/isaacarthur. As always, thanks for watching, and have a Great Week!