BOY! As a former planning commissioner who served my community for nine years, Frankie was very impressive. There were a lot of great ideas and I feel Frankie combined some very good and important ideas and concepts. Overall, I think as a large general plan of that specific area, Frankie's ideas should be taken seriously. There is hope for AI.
A lot of the ideas mentioned are things that we have been talking about in academia for years. It was still rather vague, and I would say there were some outdated thinking with the high prioritization of cost saving especially touching on micro units (I would also truthfully love for there to be more directed conversation with how higher upfront cost investment into improved & sustainable building systems, ie. gray water storage or green rooftops, would have incremental savings in the long term), but it was mostly hitting on the points that younger designers like myself have been stressing for a while now. Now let's hope private developers would actually listen to this.
@@AdambYates Could you please do this concept for Los Angeles? Our Metro is expanding so I'd like to see it plan new routes and stations, find streets and areas where we could apply the biking/pedestrian only concept, and make LA a 15 minute city with sustainability and walkability in mind. LA has a lot of things that would make this a really interesting challenge. Thanks!
The fact that the AI does this so well, is becouse so many humans have already written about this in a multitude of ways, hoping for such cities to improve peoples live. The problem is therefore not directly the people designing, but rather the factors which set guidlines and boundaries for the designers. We have to see why these plans don't get followed through and publicly adress them to make a difference.
I think if anything this shows just how impactful Strong Towns and detailed city plans from Copenhagen and Amsterdam and similar cities are to our overall development..I hope this pushes planners to be even more thorough in how they label and write up their plans, as it could be used by the latest gpt for city planning.
Did this myself in Cities Skylines. Went from city wide traffic jam at 50,000 population without public transport or bike paths to 150,000 population and less traffic than in my real life small town. Bike paths though parks surrounded by residential and a ring of business with a tram to pass by them all is the way to go (and a subway for district to district for anything bigger than a square mile)
One thing you didn't point out is that Delve is an AI tool developed by Alphabet and it's original use was for that specific area in Toronto you want redesigned. That means that Delve is optimized for that area and that ChatGPT can access sidewalk labs public plans for the area and bing can access the public images for the plans which looks to be the case because the ai generated images are very similar to sidewalks labs images for Toronto. It would be interesting to see it tried out on a place that doesn't have such a well covered development.
I thought the same thing. I like the video idea and the overall outcome but videos like these just add to the whole AI bubble. And realy showcase the lack of understanding the public has for such complex fields as AI.
@@johannes1826 Not to mention one of ChatGPTs shortcomings is planning. Granted that planning can mean different things and it's "plans" are very solid but it's not like it's actually thinking of the plans itself.
Thanks for the comment and good point. With that said, I have applied Delve on other master planned sites in the Greater Toronto Area and found it to be quite good. May be a different story outside of Toronto. I also think there are a number of really good programs out there that also do a good job at optimizing site design based on all the different constraints. I absolutely love a program called RatioCity (I think this is only available in Canada at the moment). In contrast to Delve it actually brings in municipal planning policy, so it works for infill sites in downtown areas.
The most amazing thing is that its not even an AI created for this task. And even this one is less than 7 months old. Imagine what a mature and dedicated city planning AI could do.
Amazing. Using multiple AIs with different purposes can encompass such vast knowledge and experience made by so many humans combined! The degree of comprehension by the AI for its tasks keeps blowing my mind. Thank you Adam, you need to be seen by SO many more people. Especially all community developers and planners in North America! 🤓
An AI like the ones we have, does not have any comprehension. It doesnt actually understand anything. It is a relatively static model of a certain task, like human language. There is a misconception, that AI is dangerous, because it is so smart. AI is not smart. The danger is in using AI for tasks, that need a deep understanding of what the world is and how it works and blindly believing in it.
It doesn't "agree" - it's just recognizing that that's the most common type of content. It should not be surprising that most content is made by content creators.
Now the government should just let AI build the cities to be honest. If AI would make that city happen, that would be better than anything we have right now
The problem with Ai is: It wasn't programmed to actually care about people's comforts, over the collective wants [if you understand what I mean]. Sure, it wants to protect bicyclers & that's a good thing but when it comes to people's living quarters it thinks like a communist encampment [shared kitchens & living space⁉️ NO way is that people friendly, we need PRIVACY... And where is the human green-space (like open to the air balconies). All Ai did was get the information from the NWdisOrder websites |WEF/UN/etcetera] - Ai created nothing‼️
This is fantastic, and exactly what we should be using Artificial Intelligence (or Augmented/Assistive Intelligence) for. People should be making the final decisions, but these new tools that we have at our disposal will generate new ideas and connections that we may not have otherwise considered. Adam, I hope that you shared this video with the Toronto City Planning Commission (or its equivalent)!
To answer your last question, no, AI cannot design a city better than a human, because all the AI did was take best practices from cities that nearly everyone already knows have the best land use policies. In other words, if North American planners just copied what they do over there, we'd be in a much better state. Ultimately though, this was a great exercise and I think taught some good lessons. One thing I'd note about affordability though is that unfortunately, I think this portlands neighborhood won't end up being affordable because it's built so well. The problem with our current development pattern is that we build so few of these good neighborhoods that when they pop up they are immensely popular and the price of everything skyrockets. What's needed is to transform the entire city into this type of development so that the uniqueness of it doesn't drive up the cost.
I agree with you, I think you are spot on. And I agree with your second paragraph as well. Definitely need City-wide change to drive changes in affordability. Hope this happens!
I've been trying to do something like this but you knocked it out of the park. There is definitely potential here. of course, anything made by AI would still be edited and changed by real human beings to actually be adaptable in the real world, but there's definitely potential.
At first I was sceptical on that chat GPT actually wrote this, it seemed like it had more of an opinion than usual but for the most part this was just way too good to be true. This is EXACTLY what needs to be done. It would fix soo many problems and crisis. I think it is as close to perfection as it gets. If we could just implement this, and if everything would be made with the same goals in mind (that being sustainability and everything being made for the well-being of the people) I am super optimistic for the future
I am just pulling together this very project! Using AI to assist in developing a presentation of my ideas in urban-regional planning. Thanks for the hints in how to go about this. I've go full time on this project on the 5th!
I think it is a valuable tool to take as a normative starting point for planners. This saves time and energy and upfront work for a typical planner. Long gone are the days where planners spent time developing these plans and design ideas, I think the future will be more about planners engaging critically with the implementation of these plans.
Interesting for sure, though I'd think this would have to be from a ground up deal. AI's probably not accounting for all the people who'd likely oppose it for some kind of NIMBY reason, or another...that and/or developer greed. Still, I like the possibilities of those concepts.
Where’s the plan? Could you, please, provide some kind of link to see details. I heard lots of buzz words and some interesting ideas, however there were no details and no references to validate any numbers. As you probably know, AI makes some facts from time to time.
I would worry about the fire protection for tall, timber framed buildings, and I would like to see waste collection addressed, and, if it's successful, how to deal with the influx of tourists.
@@AdambYates and fire. It's only a matter of time before there's a really spectacular multistorey timber framed building fire. They dry out, then move, and all the fire barriers become invalidated. I can't bear the thought of the lives that are on countdown.
I appreciate AI's ability to draw attention to general planning concepts as a scaffold for actual design. That said, Frankie's presentation is very general and obviously actual design is more granular with much more detail. The devil, they say, is in those details. I wonder if AI can continue to be used for the actions involved in the heavy lift of specificity.
I’m a Brit living in Amsterdam, and it really is an amazing city. I’m not even talking an about the historic centre. The post war communities are carefully planned and as we move away from the car they are getting even better every year
In Amsterdam they just closed a major road that goes straight in to the city for 6 weeks. No roadworks going on, but just to see what would happen. And off course it aims on removing this road all together at some point (except for emergency services and public transport).
Frankie gave us the best designs available based on the best numbers from what has already been proven to work. So every city planner should run a.i. just to make sure they thought of everything.
As an architect and aficionado of urban design, the issue as I understand it is auto centric zoning codes. And elected policy makers swayed by special interests.
could have been 3 mins shorter, but i enjoyed it and more to the point, it wowed me architecturally, and opened up concepts of modern living I had never thought about. I would live there, if they had a portion of the unitits designed for seniors. It showed Ai technology off, demonstrating user friendly apps.
Very interesting! Good work and coordination. Would AI be able to suggest a plan for areas outside of downtown (like suburbs) that would hit those 5 points as well? The city core stuff has been done and re-done ad nauseum. I would like to see a suburb that doesn't function like a highway strip mall. Thanks!
Nice work, Frankie and Adam. Beautifully constructed and visualised. And based on best practice and evidence from around the world. I would love to see a review of our synthesis by some seasoned and indepedent experts. I love your work, and I'm easily swayed towards the beauty you show and the type of urbanism you propose. I think we might read the same blogs! Is expert review an option to help build credibility for AI generated content? Peer-review even? Or does peer-review in the AI context boil down to Model Intercomparison Projects?
you can really see the limitations of Chatgpt, it's sounds just like an autofill tool that fluffs up anything you want, packing anything you wrant with cliches and coporate terminology
This is not a limitation, you can tell it to write the same thing at a 3 year old level and it will simplify everything, or you can ask it to be more technical. or you can ask it to write it with humour, , Flirty, or dreed, etc...
This reminds me so much of the Computer from Star Trek: The Next Generation. The Computer was AI that generated the best information possible given prompts by users. The last time I saw ST:TNG imagine what the future could be like, Starfleet officers used PADDs (personal access display devices) for everything from reading books to writing reports.
Wow, that’s a great comparison! I agree that it’s amazing how technology has advanced so much since then. I wonder what other gadgets from the show will become reality in the future. Maybe holodecks?
AI takes pre-existing information and forms responses based on your prompts, so while it gave you a nice plan to redevelop Toronto, you still need humans to come up with these great ideas in the first place. Great video mate, well done.
Yeah, I definitely agree. I do think it’s interesting that it’s able to pull practices from around the world, ie what people have written about from around the world. Like I had no idea about the Copenhagen waterfront baths - those are so cool!
The problem with city planning is implementing it. The worst part, comes when you go against those who are benefited from the current status of the city. Finally common kitchens are a horrible idea. Who wants to share his/her fridge? Into a company people have problems using a small kitchen, scale it up in a house complex.
Haha I totally agree about the common kitchen. Don’t think Frankie thought that through haha And your right about going against those that have benefited from the status quo - makes it almost impossible to drive change.
@@AdambYates yeah i del with these issues constantly here i live (Greece) and i bet everywhere is the same... otherwise, what the AI mention is almost a common knowledge and for years in European architecture schools we learn about these principles. But we are not able to implement them...
Gpt is not completely multimodal. It barely understands city plans. It has some visual understanding but for city plan ing you need visual modality and a specific Llm for city planning .In one year or less you can do it easily. Even now making a good quality llm is not that energy or time consuming.
I'm not sure it can design a better city than a human, but it definitely came up with a better design than is likely to be approved. This was a great example of urban planning, yet how often do we actually see these principles implemented?
The best cities are not designed, they evolve. I am in a fairly normal suburb of London. The main road nearby is 1000 years old, maybe several times that. Buildings range from new to 300 years old, many 100-150 (with some redesign by the Luftwaffe). My local pub has had an inn on the site for 600 years. I can walk in 5 minutes to 3 pubs, 6 cafes, 3 restaurants, 4 food shops, a cinema, a park, an underground train station and 6 bus stops. many times that for a 15 minute walk (including 2 cinemas and 4 parks). I have a car but use it once a week on average. I walk to most shops, bus for local (2-4km), tube for further (3-15km) and drive or fast train for further afield.
Designing a city isn't the problem. The problem is "special interests", ignorance, stupidity, greed, and corrupt bureaucrats, and politicians. In other words there is nothing that people can't screw up !
Not all of Europe. Still a lot to to but somehow the car Industrs don't like it. Public transportation in Germany exists, but it could be sooo much better with a little efford and a lot of money. Instead our government pays India 11 billion € for "local environmental projects". And at the same time heavily subsidizing the Car Industry here. It's all about corrupt politicians.
Task it with incorporating the existing industrial uses with transit & neighborhoods to minimize the distance a worker has to travel! Or how would it zone a neighborhood to ensure the people in the lowest income jobs can live in the same or adjacent neighborhood they serve?
I agree, pretty good plan by new software that was just released 6 months ago. Really wonder what it will be able to do in 10 years. Thanks for the feedback, if I make another video about this topic, I’ll include more ChatGPT scenes!
Water front development areas need to be build with sea level rise in mind. Instead of building up and trying to keep the water out. Build down and into/under the water and use the shore as a transport point to the above ground part of the city. No climate change worry. No need for insulation. Easy to upgrade/build around. Don't try to stop water rise. Enjoy it!
Were the renders made by bing image creator or chat gpt? What is Delve and what is it used for? Shockingly impressed by what you've asked AI to do in your city
Thanks for the comment. I asked ChatGPT to provide the prompts for Bing Image Creator, then I copied and pasted those prompts into Bing Image Creator and it produced the renderings. Delve is a pretty cool program. This youtube video explains it best: ruclips.net/video/J8QzRw9rQMk/видео.html. Basically, Delve uses AI to create millions of design possibilities for urban developments. You input the project information, location, size, priorities, etc., and it generates and ranks options that meet your criteria. It finds the optimal way to layout the roads, buildings, etc.
I don't see any actual design here. Just a presentation saying "I would do X". I high school student could do that. This is just regurgitating best practices that have been identified by humans and talked about at length online. The AI has just absorbed that and applied it to a few geo-specific things (like the Don River mouth), but I don't see any masterplan. I don't see any renders of an actual design. We are still quite a way off from AI doing the serious grunt work.
I’ll see if I can export the delve plan or may do another video that more clearly shows the layout, massing and statistics. Fair enough, I guess it depends where you live. In Toronto we still have a number thought leaders that think that we need to remove bicycle lanes on main streets and increase the number of vehicles lanes to improve traffic congestion.
Heyy I have a huge land lot around 2.5m sq ft, What is the best AI Program to help me get ideas for a masterplan of the lot? any recommendations would be incredibly helpful. Thanks in advance.
Ai is a powerful analytical tool, so it can design a city that is perfect when it comes to efficiency. I wonder what it would be like to live in such a city?
ChatGPT keeps saying "incorporating street designs from Amsterdam" but nowhere in the pictures it does the street design resemble anywhere in Amsterdam. Not even future Amsterdam developments (of which planning documents are of course available but not in English). Even in the Lower Don Lands (which looks really good by the way) it looks distinctly Canadian. That's of course a good thing, but it also means that the infrastructure still isn't on par with Amsterdam. I can see cycle friendly streets, but I'm missing a cycle friendly network. I can see bike "provisions" but I am not seeing dedicated well-marked wide cycle paths that are continuous. I'm seeing a lot of glass facades but not a lot of non-glass facades. Nothing wrong with big sglass storefronts, but glass facades' overuse do have a downside: glass reflects and the outside light and turns entire walls into mirrors. Making a place feel cold and impersonal (large hard smooth textureless surfaces). Having actual walls, colums and depth in facades is part of good architecture, and is more sustainable anyway.
Interestingly this provides insight into how the algorithm works and its limitations. It builds on/copies/combines the contents of two popular RUclips channels that gave sparked a movement, Not Just Bikes and Strongtowns. So there's hope for us humans yet!
I'm glad to hear we need AI to reckon what we've always felt being harmony, beauty and efficiency is the good approach to building up cities. Wait for a couple of years, and AI will tell us that morality and decency are better than individualism to live together in peace.
I love this. Certainly would like to like in a city like this... couple things could be better. Maybe separate the pedestrian walking network vertically and have the cars on the lowest level if say 3. Also would having houses with only bedrooms reduce population demographics. One br for the parents one for the child?. We have 3 kids its tight in a 3 br place already. But the main thing even to get thus working. Maybe for a follow up video. Whats Frankie's plan to get all the stakeholders on board. How is Frankie going to convince the politicians, and executives of say car companies to do this. How is is it Frankie going to convince the people of the city to embrace this. I'd love to see this type of city built but certain people don't want to change. I think that would be a cool video. Thanks
@@AdambYates I would say Copenhagen has many cool neighborhoods that work well together, all making the city good in general. We still need some improvement though🇩🇰! If the whole of Toronto was changed to Frankie's plan, I would definitely considering moving there🇨🇦!
I'm slightly bothered that there was only one winter image - of people skating. One of the major problems in cities where one wants to bike or walk year round is ensuring that those activities are practical during inclement weather. I'm surprised that there was not more about infrastructure and maintenance thereof. If central heating/cooling is used, there needs to be redundancy - I worked in downtown Detroit during a week-long central heating outage when the high temperatures were about -20C. Not fun.
This is a very valid criticism of AI’s plan. Definitely a important consideration for whether this plan could ultimately be successful in the real world!
The ideas it came up with were better than anything that already exists in the US/Canada, but its impossible to actually do because there are too many barriers that can't be broken through
My main question is how innovative is AI? Right now it’s pulling from the existing knowledge base, but can it generate new models, paradigms or configurations for improvement?
How do we change the system? By uniting, organizing and building new cities from scratch. As Buckminster Fuller said: "You don't change things by fighting the existing reality. To change something you need to build a new model tham makes the old one obsolete." So to "change" the (economic) system we currently use it is more effective to build a new economic system, than it is to fight against the current one. And the most efficient way to build a new economic system is to build new cities from scratch, because it is in cities that economic systems materialize. For more on this see: 1 - Zeitgeist Moving Forward: ruclips.net/video/4Z9WVZddH9w/видео.html 2 - Sotsgorod: ruclips.net/video/k1fFd4S9IBc/видео.html 3 - Marinaleda: ruclips.net/video/Wgudj5Rw10I/видео.html 4 - Stafford Beer - Viable System: ruclips.net/video/gPnWVg7CSIg/видео.html 5 - Buckminster Fuller: ruclips.net/video/3ZB2La-oCVI/видео.html 6 - Egypt's new capital currently being built by China: ruclips.net/video/9-ThusbaRW8/видео.html 7 - Socialism: ruclips.net/video/OUig0Qwnc4I/видео.html
Women fought against the laws that kept them from vpting and basic rights line signing contracts. Prooves you are wrong. Why build something new when you can use old infrasteucture? Your idea just waistes ressozrces that are very limited.
The recommendations are all based on existing solutions which are not acceptable to all types of likely inhabitants eg those who have to commute to work or schools or meet elderly relatives who need care and support. It just moves existing (traffic) problems to other areas exacerbating probs like emergency service access to ow traffic neighbourhoods.
Frankie would maybe not better as the best city planners, but definitely this sounds like the best city planner most cities would get to. Just ask for best practices combined with a guiding back story and get a planner that tries to meet your goals with no ego or profit motive in the way.
AI really is human because all the text generated came from learning from human inputs. The problem isn't that we dont know how to make more desirable urban areas in the USA and Canada, its that we lack the political will.
Impressive, sadly Frankie couldn't take into consideration local politicians stuck in the 60s, corrupt developers, car manufacturers lobbyists, NIMBYs, etc . But at this rate Skynet is coming in about 10 years so...........................
Damn thats interesting. I didn't quite understand what AI (?) you used to visualize the texts given by the AI. Or did i just misunderstand your intro and you used your own programm to do that?
There's a free program call Bing Image Creator (www.bing.com/create). I asked ChatGPT to make the prompts for Bing Image Creator, and Bing Image Creator created the images.
All Chat GPT did was regurgitate, without understanding, what humans from its training data are saying are best practices. It doesn't understand. That said, solver functions can go a long way to help plan thinga. Still needs something that understands the problem and processes, which AI's don't. Still takes a human for good results. Best results are human with AI and solver functions to model things out
Everything she said was cheaper on money and cheaper on carbon emissions! We should just always let A.I. do our homework for us... I want to live in her city!
Hi Adam, is there a way to utilize this for a different city? I work for a city in the US and would to change the parameters for that city to present this to people that I know here. Thank you! - Ian
I like Frankie's pitch, prefigured as it was by you with terms like "world's best practice" but then I'm one of those progressive cosmopolitan sorts. It would be interesting to ask Frankie to alter her presentation in a way that would appeal to a different kind of person, one who is wary of globalization and state-certified expertise, but who would nonetheless benefit from a more localized community-oriented way of life.
Adam, impressed by your AI in city planning. Wish to discuss EcoVision Velddrif, a 2,700ha sustainable urban development in SA, fully renewable, aiming for a transformative community model. Open for a call to explore collaboration?
![Lethal Company Music Video 🎵 "Follow The Orders" [VERSION A]](http://i.ytimg.com/vi/QeooZB4Z7RA/mqdefault.jpg)
BOY! As a former planning commissioner who served my community for nine years, Frankie was very impressive. There were a lot of great ideas and I feel Frankie combined some very good and important ideas and concepts. Overall, I think as a large general plan of that specific area, Frankie's ideas should be taken seriously. There is hope for AI.
It’s great to know from a former planning commissioner that Frankie is on the right track!!
A lot of the ideas mentioned are things that we have been talking about in academia for years. It was still rather vague, and I would say there were some outdated thinking with the high prioritization of cost saving especially touching on micro units (I would also truthfully love for there to be more directed conversation with how higher upfront cost investment into improved & sustainable building systems, ie. gray water storage or green rooftops, would have incremental savings in the long term), but it was mostly hitting on the points that younger designers like myself have been stressing for a while now. Now let's hope private developers would actually listen to this.
I agree. I really hope councillors and the public start to use this for educational purposes.
@@AdambYates Could you please do this concept for Los Angeles? Our Metro is expanding so I'd like to see it plan new routes and stations, find streets and areas where we could apply the biking/pedestrian only concept, and make LA a 15 minute city with sustainability and walkability in mind. LA has a lot of things that would make this a really interesting challenge. Thanks!
@@AdambYatesEducational Purposes? Yes. Work "with" AI to improve our lives. Make us better humans if it's possible.
The fact that the AI does this so well, is becouse so many humans have already written about this in a multitude of ways, hoping for such cities to improve peoples live.
The problem is therefore not directly the people designing, but rather the factors which set guidlines and boundaries for the designers.
We have to see why these plans don't get followed through and publicly adress them to make a difference.
💯 agree with this!
Because people don't want to walk/bike when the weather sucks?
The oil industry is still too big of a political sponsor
@@cyberslim7955 I bike when snow and rain are bearable. When they are unbearable even people in cars avoid going out. So there is that.
@@cyberslim7955What? Just grab and umbrella and get on the bus, it’s not difficult or much different from taking a car
I think if anything this shows just how impactful Strong Towns and detailed city plans from Copenhagen and Amsterdam and similar cities are to our overall development..I hope this pushes planners to be even more thorough in how they label and write up their plans, as it could be used by the latest gpt for city planning.
I agree!
@@AdambYatessoviet microdistrict vs USA suburb
Haha
@@carkawalakhatulistiwaore like : US single use zoning stupidity vs mixed use common sense ('' most of the rest of the world ' ) .
This is a sick plan. Good job Frankie.
Haha, thanks Kyle for the comment, I am sure Frankie appreciates it!
Great job, Frankie! 😊
Well done!
Did this myself in Cities Skylines.
Went from city wide traffic jam at 50,000 population without public transport or bike paths to 150,000 population and less traffic than in my real life small town.
Bike paths though parks surrounded by residential and a ring of business with a tram to pass by them all is the way to go (and a subway for district to district for anything bigger than a square mile)
Very cool!! Sounds like an interesting way to model cities
You mean "Sick" as in "Disturbed", "Sadistic", "Creating a living hell"?
One thing you didn't point out is that Delve is an AI tool developed by Alphabet and it's original use was for that specific area in Toronto you want redesigned. That means that Delve is optimized for that area and that ChatGPT can access sidewalk labs public plans for the area and bing can access the public images for the plans which looks to be the case because the ai generated images are very similar to sidewalks labs images for Toronto. It would be interesting to see it tried out on a place that doesn't have such a well covered development.
I thought the same thing. I like the video idea and the overall outcome but videos like these just add to the whole AI bubble. And realy showcase the lack of understanding the public has for such complex fields as AI.
@@johannes1826 Not to mention one of ChatGPTs shortcomings is planning. Granted that planning can mean different things and it's "plans" are very solid but it's not like it's actually thinking of the plans itself.
Thanks for the comment and good point. With that said, I have applied Delve on other master planned sites in the Greater Toronto Area and found it to be quite good. May be a different story outside of Toronto. I also think there are a number of really good programs out there that also do a good job at optimizing site design based on all the different constraints. I absolutely love a program called RatioCity (I think this is only available in Canada at the moment). In contrast to Delve it actually brings in municipal planning policy, so it works for infill sites in downtown areas.
The most amazing thing is that its not even an AI created for this task. And even this one is less than 7 months old. Imagine what a mature and dedicated city planning AI could do.
I agree, this is truly impressive.
Amazing. Using multiple AIs with different purposes can encompass such vast knowledge and experience made by so many humans combined! The degree of comprehension by the AI for its tasks keeps blowing my mind.
Thank you Adam, you need to be seen by SO many more people. Especially all community developers and planners in North America! 🤓
Thanks! I completely agree, AI is blowing me away!!
I could see how AI being used for higher education could bring a updated set of best practices, could accelerate adopting them.
An AI like the ones we have, does not have any comprehension. It doesnt actually understand anything. It is a relatively static model of a certain task, like human language. There is a misconception, that AI is dangerous, because it is so smart. AI is not smart. The danger is in using AI for tasks, that need a deep understanding of what the world is and how it works and blindly believing in it.
Yep!
It does seem to understand basic logic though… like if you ask it what happens if you cut the line to balloons, it knows that the balloons fly away.
Basically AI agrees with Strong Towns, Not Just Bikes, City Beautiful, Oh the Urbanity, and most of the other urbanist channels.
Yeah pretty interesting!
It doesn't "agree" - it's just recognizing that that's the most common type of content. It should not be surprising that most content is made by content creators.
@@manzell Is it really? I can guarantee that there is also a lot of written text that explicitly advocates for more car usage and wider lanes.
@@iwiffitthitotonacc4673 Not a chance.
@@manzell Or maybe it was trying to reduce traffic deaths or increase energy efficiency 🤣
Cars are insane.
Now the government should just let AI build the cities to be honest. If AI would make that city happen, that would be better than anything we have right now
Well human planners may well design good cities, but cities aren't dictatorships. There's a lot of compromise that results from the political process.
Yeah, no old fashioned outdated or lobbying to corrupt the design either.
Perfection
The problem with Ai is:
It wasn't programmed to actually care about people's comforts, over the collective wants [if you understand what I mean]. Sure, it wants to protect bicyclers & that's a good thing but when it comes to people's living quarters it thinks like a communist encampment [shared kitchens & living space⁉️ NO way is that people friendly, we need PRIVACY... And where is the human green-space (like open to the air balconies).
All Ai did was get the information from the NWdisOrder websites |WEF/UN/etcetera] - Ai created nothing‼️
@@matthewboyd8689 Ya, foolish Matthew - just a govt/Ai dictatorship to deal with...
@@kellikelli4413 that's why it says it would implement a *variety* of housing types...
This just described my home city Groningen. Love it.
Haha sounds like you live in a great place!
I am impressed, and unsurprised. Good job, you two!
Thanks Emma! Appreciate the comment :)
I would like to see Frankie do many other cities, too:)
Any other suggestions?
@@AdambYates DENVER. Frankie needs to fix the affordable housing crisis there, along with the lack of decent transit, parks, and bike lanes.
Hilo, Hawaii! Would be interesting to see what Frankie could do with a smaller city (pop. 45k)
Quincy washington pop 5,000 (when i was last there over a decade ago)
Great, I’ll check those out! I want to try with ChatGPT 4 and Google’s Bard.
This is fantastic, and exactly what we should be using Artificial Intelligence (or Augmented/Assistive Intelligence) for. People should be making the final decisions, but these new tools that we have at our disposal will generate new ideas and connections that we may not have otherwise considered. Adam, I hope that you shared this video with the Toronto City Planning Commission (or its equivalent)!
Welcome to Utrecht, Frankie! Quite sad you only mentioned Amsterdam after that, but I guess A'dam is more popular internationally.
Haha
To answer your last question, no, AI cannot design a city better than a human, because all the AI did was take best practices from cities that nearly everyone already knows have the best land use policies. In other words, if North American planners just copied what they do over there, we'd be in a much better state.
Ultimately though, this was a great exercise and I think taught some good lessons.
One thing I'd note about affordability though is that unfortunately, I think this portlands neighborhood won't end up being affordable because it's built so well. The problem with our current development pattern is that we build so few of these good neighborhoods that when they pop up they are immensely popular and the price of everything skyrockets. What's needed is to transform the entire city into this type of development so that the uniqueness of it doesn't drive up the cost.
I agree with you, I think you are spot on. And I agree with your second paragraph as well. Definitely need City-wide change to drive changes in affordability. Hope this happens!
I've been trying to do something like this but you knocked it out of the park. There is definitely potential here. of course, anything made by AI would still be edited and changed by real human beings to actually be adaptable in the real world, but there's definitely potential.
Thanks appreciate the comment! And I definitely agree that it would still need a human touch!
Great video! I haven't seen anyone using Ai to help design a part of the city before. Also great graphics and flow!
Thanks Kevin! I appreciate the comment :) I’m definitely surprised how well it did!
Basically a hybrid of Amsterdam and Copenhagen - both my favourite cities. Well done Frankie!
Mr. Yates, you are the man. I really, really love how you reply to comments. That’s so nice of you.
haha thanks!! If someone is nice enough to watch my video, I'm definitely going to try my best to respond to their comment!
i'm only 3.5 minutes in and my brain has already stopped working it's crazy how well these ais work together
Pretty wild, I agree!
It was good actually to see some numbers. The only thing I didn't like was tall building walls looking at each other
At first I was sceptical on that chat GPT actually wrote this, it seemed like it had more of an opinion than usual but for the most part this was just way too good to be true. This is EXACTLY what needs to be done. It would fix soo many problems and crisis. I think it is as close to perfection as it gets. If we could just implement this, and if everything would be made with the same goals in mind (that being sustainability and everything being made for the well-being of the people) I am super optimistic for the future
Thanks for the comment and I completely agree!
I am just pulling together this very project! Using AI to assist in developing a presentation of my ideas in urban-regional planning.
Thanks for the hints in how to go about this. I've go full time on this project on the 5th!
That sounds awesome! Good luck!!
Low buildings with gardens. Parks with Benches.
I think it is a valuable tool to take as a normative starting point for planners. This saves time and energy and upfront work for a typical planner. Long gone are the days where planners spent time developing these plans and design ideas, I think the future will be more about planners engaging critically with the implementation of these plans.
great video! I think the only thing it was missing for me was good music to go along with your topics. Keep up the great content.
Thanks for the feedback, I’ll try that for the next video!
Passive Solar design techniques can also be used to calculate the sun angles, so there's less demand for heating/AC
A core shopping-centre for groceries and scholing elementary are different side than colleges.
I really enjoyed this. AI gets a lot of bad PR, but Franky shows what a valuable asset AI can be. Newly subscribed.
Interesting for sure, though I'd think this would have to be from a ground up deal. AI's probably not accounting for all the people who'd likely oppose it for some kind of NIMBY reason, or another...that and/or developer greed. Still, I like the possibilities of those concepts.
I agree
Where’s the plan? Could you, please, provide some kind of link to see details. I heard lots of buzz words and some interesting ideas, however there were no details and no references to validate any numbers. As you probably know, AI makes some facts from time to time.
I’ll see if I can export the delve plan or may do another video that more clearly shows the layout, massing and statistics.
Anton, I would also like to see the numbers.
I would worry about the fire protection for tall, timber framed buildings, and I would like to see waste collection addressed, and, if it's successful, how to deal with the influx of tourists.
Waste collection and package delivery is a really good question!!
@@AdambYates and fire. It's only a matter of time before there's a really spectacular multistorey timber framed building fire. They dry out, then move, and all the fire barriers become invalidated. I can't bear the thought of the lives that are on countdown.
Yikes that doesn’t sound good. I was under the impression that they largely solved this issue with gluelam and fire resistant coatings.?
I appreciate AI's ability to draw attention to general planning concepts as a scaffold for actual design. That said, Frankie's presentation is very general and obviously actual design is more granular with much more detail. The devil, they say, is in those details. I wonder if AI can continue to be used for the actions involved in the heavy lift of specificity.
I'd hire Frankie... he seems to know his shit.
I’m sure Frankie would appreciate that! Haha
I’m a Brit living in Amsterdam, and it really is an amazing city. I’m not even talking an about the historic centre. The post war communities are carefully planned and as we move away from the car they are getting even better every year
In Amsterdam they just closed a major road that goes straight in to the city for 6 weeks. No roadworks going on, but just to see what would happen. And off course it aims on removing this road all together at some point (except for emergency services and public transport).
That’s awesome! I’m headed there in August, hopefully they’re still up!
Frankie gave us the best designs available based on the best numbers from what has already been proven to work. So every city planner should run a.i. just to make sure they thought of everything.
As an architect and aficionado of urban design, the issue as I understand it is auto centric zoning codes. And elected policy makers swayed by special interests.
Sums it up well!
@@AdambYates Refreshing to get your response! Thank you. I find your project to have value. Good guidance often gets good results.
could have been 3 mins shorter, but i enjoyed it and more to the point, it wowed me architecturally, and opened up concepts of modern living I had never thought about. I would live there, if they had a portion of the unitits designed for seniors. It showed Ai technology off, demonstrating user friendly apps.
Very interesting! Good work and coordination. Would AI be able to suggest a plan for areas outside of downtown (like suburbs) that would hit those 5 points as well? The city core stuff has been done and re-done ad nauseum. I would like to see a suburb that doesn't function like a highway strip mall. Thanks!
Dude I would love here in a heartbeat! Frankie had some really good ideas!
You and me both! Haha
Frankie did a wonderful job. Tell the robots to start construction!
Haha
i can't believe AI better than humans
Nice work, Frankie and Adam. Beautifully constructed and visualised. And based on best practice and evidence from around the world. I would love to see a review of our synthesis by some seasoned and indepedent experts. I love your work, and I'm easily swayed towards the beauty you show and the type of urbanism you propose. I think we might read the same blogs! Is expert review an option to help build credibility for AI generated content? Peer-review even? Or does peer-review in the AI context boil down to Model Intercomparison Projects?
you can really see the limitations of Chatgpt, it's sounds just like an autofill tool that fluffs up anything you want, packing anything you wrant with cliches and coporate terminology
Yep, ChatGPT definitely has limitations!
This is not a limitation, you can tell it to write the same thing at a 3 year old level and it will simplify everything, or you can ask it to be more technical. or you can ask it to write it with humour, , Flirty, or dreed, etc...
This reminds me so much of the Computer from Star Trek: The Next Generation. The Computer was AI that generated the best information possible given prompts by users. The last time I saw ST:TNG imagine what the future could be like, Starfleet officers used PADDs (personal access display devices) for everything from reading books to writing reports.
Wow, that’s a great comparison! I agree that it’s amazing how technology has advanced so much since then. I wonder what other gadgets from the show will become reality in the future. Maybe holodecks?
Your videos are really high quality for such a small channel. Keep up the good work👍
Thanks!! Appreciate the comment! Means a lot!
AI takes pre-existing information and forms responses based on your prompts, so while it gave you a nice plan to redevelop Toronto, you still need humans to come up with these great ideas in the first place. Great video mate, well done.
Yeah, I definitely agree. I do think it’s interesting that it’s able to pull practices from around the world, ie what people have written about from around the world. Like I had no idea about the Copenhagen waterfront baths - those are so cool!
It's long been said that there's nothing truly new under the sun.
I really appreciate this video to help me with my passion project! ❤
The problem with city planning is implementing it. The worst part, comes when you go against those who are benefited from the current status of the city.
Finally common kitchens are a horrible idea. Who wants to share his/her fridge? Into a company people have problems using a small kitchen, scale it up in a house complex.
Haha I totally agree about the common kitchen. Don’t think Frankie thought that through haha
And your right about going against those that have benefited from the status quo - makes it almost impossible to drive change.
@@AdambYates yeah i del with these issues constantly here i live (Greece) and i bet everywhere is the same...
otherwise, what the AI mention is almost a common knowledge and for years in European architecture schools we learn about these principles. But we are not able to implement them...
Interesting to hear that this isn’t only a North American problem!
Amazing! I learned so much from this
Thanks Borja!
Gpt is not completely multimodal. It barely understands city plans. It has some visual understanding but for city plan ing you need visual modality and a specific Llm for city planning .In one year or less you can do it easily. Even now making a good quality llm is not that energy or time consuming.
I'm not sure it can design a better city than a human, but it definitely came up with a better design than is likely to be approved. This was a great example of urban planning, yet how often do we actually see these principles implemented?
Wow ChatGPT should just be hired as the city planner lol.
Haha
12:54 everything always sounds good when planning. for today, this might be the better solution, then 20 years from now it isn't.
Definitely! Things are changing so quickly it’s amazing.
The best cities are not designed, they evolve. I am in a fairly normal suburb of London. The main road nearby is 1000 years old, maybe several times that. Buildings range from new to 300 years old, many 100-150 (with some redesign by the Luftwaffe). My local pub has had an inn on the site for 600 years. I can walk in 5 minutes to 3 pubs, 6 cafes, 3 restaurants, 4 food shops, a cinema, a park, an underground train station and 6 bus stops. many times that for a 15 minute walk (including 2 cinemas and 4 parks). I have a car but use it once a week on average. I walk to most shops, bus for local (2-4km), tube for further (3-15km) and drive or fast train for further afield.
Designing a city isn't the problem. The problem is "special interests", ignorance, stupidity, greed, and corrupt bureaucrats, and politicians. In other words there is nothing that people can't screw up !
It's funny that when it comes to public transport AI is just reinventing Europe :D
Not all of Europe. Still a lot to to but somehow the car Industrs don't like it.
Public transportation in Germany exists, but it could be sooo much better with a little efford and a lot of money.
Instead our government pays India 11 billion € for "local environmental projects".
And at the same time heavily subsidizing the Car Industry here.
It's all about corrupt politicians.
Task it with incorporating the existing industrial uses with transit & neighborhoods to minimize the distance a worker has to travel! Or how would it zone a neighborhood to ensure the people in the lowest income jobs can live in the same or adjacent neighborhood they serve?
I forgot public transport which decreases cars 🚘. Bikes 🚲 busses 🚎 trams 🚃 wirk together with walking 🚶♂️ .
Absolutey amazing. Would have liked to see more detail about the GPT interaction.
I agree, pretty good plan by new software that was just released 6 months ago. Really wonder what it will be able to do in 10 years. Thanks for the feedback, if I make another video about this topic, I’ll include more ChatGPT scenes!
Water front development areas need to be build with sea level rise in mind.
Instead of building up and trying to keep the water out.
Build down and into/under the water and use the shore as a transport point to the above ground part of the city.
No climate change worry.
No need for insulation.
Easy to upgrade/build around.
Don't try to stop water rise. Enjoy it!
Preemptive climate change city design is actually an amazing idea!
Toronto waterfront is about 20 m above sea level
Very cool. I never heard of Metro Micro though, gonna have to google it.
Thanks Paul! Yeah would love to know your thoughts on the service if you try it out!
Were the renders made by bing image creator or chat gpt? What is Delve and what is it used for? Shockingly impressed by what you've asked AI to do in your city
Thanks for the comment. I asked ChatGPT to provide the prompts for Bing Image Creator, then I copied and pasted those prompts into Bing Image Creator and it produced the renderings. Delve is a pretty cool program. This youtube video explains it best: ruclips.net/video/J8QzRw9rQMk/видео.html. Basically, Delve uses AI to create millions of design possibilities for urban developments. You input the project information, location, size, priorities, etc., and it generates and ranks options that meet your criteria. It finds the optimal way to layout the roads, buildings, etc.
I don't see any actual design here. Just a presentation saying "I would do X". I high school student could do that. This is just regurgitating best practices that have been identified by humans and talked about at length online. The AI has just absorbed that and applied it to a few geo-specific things (like the Don River mouth), but I don't see any masterplan. I don't see any renders of an actual design. We are still quite a way off from AI doing the serious grunt work.
I’ll see if I can export the delve plan or may do another video that more clearly shows the layout, massing and statistics.
Fair enough, I guess it depends where you live. In Toronto we still have a number thought leaders that think that we need to remove bicycle lanes on main streets and increase the number of vehicles lanes to improve traffic congestion.
@@AdambYates That's a problem with some humans but not others. Not an AI vs human problem though!
Fair point!
Heyy I have a huge land lot around 2.5m sq ft, What is the best AI Program to help me get ideas for a masterplan of the lot? any recommendations would be incredibly helpful. Thanks in advance.
Ai is a powerful analytical tool, so it can design a city that is perfect when it comes to efficiency. I wonder what it would be like to live in such a city?
Go Frankie go!
ChatGPT keeps saying "incorporating street designs from Amsterdam" but nowhere in the pictures it does the street design resemble anywhere in Amsterdam. Not even future Amsterdam developments (of which planning documents are of course available but not in English). Even in the Lower Don Lands (which looks really good by the way) it looks distinctly Canadian. That's of course a good thing, but it also means that the infrastructure still isn't on par with Amsterdam. I can see cycle friendly streets, but I'm missing a cycle friendly network. I can see bike "provisions" but I am not seeing dedicated well-marked wide cycle paths that are continuous. I'm seeing a lot of glass facades but not a lot of non-glass facades. Nothing wrong with big sglass storefronts, but glass facades' overuse do have a downside: glass reflects and the outside light and turns entire walls into mirrors. Making a place feel cold and impersonal (large hard smooth textureless surfaces). Having actual walls, colums and depth in facades is part of good architecture, and is more sustainable anyway.
I’m surprised Frankie specifically cited Strong Towns
This would be the best city on earth. I'd love to live here.
I agree, would love to see this experiment play out!
I wanna c AI create a map of NYC designed all with natural roads, like the road map before the street grids.
Interestingly this provides insight into how the algorithm works and its limitations. It builds on/copies/combines the contents of two popular RUclips channels that gave sparked a movement, Not Just Bikes and Strongtowns. So there's hope for us humans yet!
Apparently OpenAI trained it’s models on RUclips so makes sense.
I'm glad to hear we need AI to reckon what we've always felt being harmony, beauty and efficiency is the good approach to building up cities.
Wait for a couple of years, and AI will tell us that morality and decency are better than individualism to live together in peace.
Alt title: I asked AI to write a not just bikes video
Haha yes!
I love this. Certainly would like to like in a city like this... couple things could be better.
Maybe separate the pedestrian walking network vertically and have the cars on the lowest level if say 3.
Also would having houses with only bedrooms reduce population demographics. One br for the parents one for the child?. We have 3 kids its tight in a 3 br place already.
But the main thing even to get thus working. Maybe for a follow up video. Whats Frankie's plan to get all the stakeholders on board. How is Frankie going to convince the politicians, and executives of say car companies to do this. How is is it Frankie going to convince the people of the city to embrace this. I'd love to see this type of city built but certain people don't want to change. I think that would be a cool video. Thanks
Sick, now build an AI that can convince the politicians.
Crazy what AI can do! As someone from Copenhagen Iwould say, I would definitely consider moving there if this was real🇩🇰🇨🇦! Nice work:)
I was shocked as well! Although, sounds like Copenhagen is overall probably better designed than Frankie’s plan??
@@AdambYates I would say Copenhagen has many cool neighborhoods that work well together, all making the city good in general. We still need some improvement though🇩🇰!
If the whole of Toronto was changed to Frankie's plan, I would definitely considering moving there🇨🇦!
I'm slightly bothered that there was only one winter image - of people skating.
One of the major problems in cities where one wants to bike or walk year round is ensuring that those activities are practical during inclement weather. I'm surprised that there was not more about infrastructure and maintenance thereof. If central heating/cooling is used, there needs to be redundancy - I worked
in downtown Detroit during a week-long central heating outage when the high temperatures were about -20C. Not fun.
This is a very valid criticism of AI’s plan. Definitely a important consideration for whether this plan could ultimately be successful in the real world!
I would live here... Frankie, when can i move in??? 😂
After spending a weekend in Toronto a couple weeks ago, Biking is way better there than in most of the USA.
Thanks for the comment. I agree, although it gets pretty rough outside of the main city centre.
The ideas it came up with were better than anything that already exists in the US/Canada, but its impossible to actually do because there are too many barriers that can't be broken through
I agree
My main question is how innovative is AI? Right now it’s pulling from the existing knowledge base, but can it generate new models, paradigms or configurations for improvement?
How do we change the system? By uniting, organizing and building new cities from scratch. As Buckminster Fuller said: "You don't change things by fighting the existing reality. To change something you need to build a new model tham makes the old one obsolete." So to "change" the (economic) system we currently use it is more effective to build a new economic system, than it is to fight against the current one. And the most efficient way to build a new economic system is to build new cities from scratch, because it is in cities that economic systems materialize.
For more on this see:
1 - Zeitgeist Moving Forward: ruclips.net/video/4Z9WVZddH9w/видео.html
2 - Sotsgorod: ruclips.net/video/k1fFd4S9IBc/видео.html
3 - Marinaleda: ruclips.net/video/Wgudj5Rw10I/видео.html
4 - Stafford Beer - Viable System: ruclips.net/video/gPnWVg7CSIg/видео.html
5 - Buckminster Fuller: ruclips.net/video/3ZB2La-oCVI/видео.html
6 - Egypt's new capital currently being built by China: ruclips.net/video/9-ThusbaRW8/видео.html
7 - Socialism: ruclips.net/video/OUig0Qwnc4I/видео.html
Women fought against the laws that kept them from vpting and basic rights line signing contracts.
Prooves you are wrong.
Why build something new when you can use old infrasteucture?
Your idea just waistes ressozrces that are very limited.
The recommendations are all based on existing solutions which are not acceptable to all types of likely inhabitants eg those who have to commute to work or schools or meet elderly relatives who need care and support.
It just moves existing (traffic) problems to other areas exacerbating probs like emergency service access to ow traffic neighbourhoods.
Frankie would maybe not better as the best city planners, but definitely this sounds like the best city planner most cities would get to. Just ask for best practices combined with a guiding back story and get a planner that tries to meet your goals with no ego or profit motive in the way.
I totally agree!
Please send this to Toronto municipality because it keeps building sprawl car dependent single family houses zones.
Haha you’re right!
The city planner did do a good job on the structures . But the issues I see are low planning for emergency services and vandalism
Thanks for the comment! I agree there needs to be more consideration for emergency vehicles but why vandalism?
AI cannot design a city better than a human, but it can do it faster and with less work
AI really is human because all the text generated came from learning from human inputs. The problem isn't that we dont know how to make more desirable urban areas in the USA and Canada, its that we lack the political will.
100% agree! Political will is definitely the toughest challenge.
Impressive, sadly Frankie couldn't take into consideration local politicians stuck in the 60s, corrupt developers, car manufacturers lobbyists, NIMBYs, etc . But at this rate Skynet is coming in about 10 years so...........................
Haha, yep, we’ll said!
Please do St. Louis Missouri next. For science.
Wow, just googled Missouri urban planning - first thing that came up was “Missouri is the 4th most car-dependent state in America.” 😬
Damn thats interesting.
I didn't quite understand what AI (?) you used to visualize the texts given by the AI. Or did i just misunderstand your intro and you used your own programm to do that?
There's a free program call Bing Image Creator (www.bing.com/create). I asked ChatGPT to make the prompts for Bing Image Creator, and Bing Image Creator created the images.
I would like to use this to create a concept for a community. Would you be able to help me
All Chat GPT did was regurgitate, without understanding, what humans from its training data are saying are best practices. It doesn't understand.
That said, solver functions can go a long way to help plan thinga.
Still needs something that understands the problem and processes, which AI's don't. Still takes a human for good results. Best results are human with AI and solver functions to model things out
I agree. AI responses can be way off. Doesn’t always understand the logic.
Everything she said was cheaper on money and cheaper on carbon emissions! We should just always let A.I. do our homework for us... I want to live in her city!
Hi Adam, is there a way to utilize this for a different city? I work for a city in the US and would to change the parameters for that city to present this to people that I know here.
Thank you! - Ian
Yes, I think Delve was developed in New York City so I think it would work across the US.
I like Frankie's pitch, prefigured as it was by you with terms like "world's best practice" but then I'm one of those progressive cosmopolitan sorts. It would be interesting to ask Frankie to alter her presentation in a way that would appeal to a different kind of person, one who is wary of globalization and state-certified expertise, but who would nonetheless benefit from a more localized community-oriented way of life.
Frankie seems to really like mass timber buildings.
Adam, impressed by your AI in city planning. Wish to discuss EcoVision Velddrif, a 2,700ha sustainable urban development in SA, fully renewable, aiming for a transformative community model. Open for a call to explore collaboration?