Towards the end of last year, Meta released SAM 3, which stands for the third version of its Segment Anything Model. The way it generally works is that it allows you to detect, edit, and experiment with things in images and videos. For example, if you were looking at a video of a street, you could ask it to find all the scooters (which I did below), count the number of pedestrians wearing black pants, blur all the license plates on the cars, and so on.

This is immediately useful for a company like Meta because it allows for object-level modifications across its content creation platforms. So if you took a video of someone dancing and you desperately wanted to give them a bobblehead, SAM 3, I'm told, would allow you to quickly do that. Other AI models, such as Gemini, can also segment, but supposedly the SAM models are better and more precise at this specific task.

Beyond bobblehead videos, the potential of this model seems enormous for real estate, cities, and, of course, many other things. Using the above image as an example, you can quickly imagine SAM 3 being used to count and track modal splits across a city, and then make planning decisions based on real-time data.

People are also using it for real estate purposes. Pair the model with satellite images, and you can ask it to tell you how many houses have a pool, which houses recently had their roof replaced (and have solar panels), how many cars are parked on a street, how many cars are parked at Canadian Tire, and the average building lot coverage in an area.

You could also use it to swap out finishes in a real estate listing (including in videos), and get material/area takeoffs ahead of a construction project. I don't know for sure, but I would also imagine that this model would make a great building condition inspector. Come to think of it, I'd love a SAM 3 that could walk our construction sites and document every little detail!

Of course, a lot of these use cases are already being tackled. But the models are getting that much better. And that will lead to even more innovation.

Cover photo by Above Horizon on Unsplash

Sometime last year, Instagram changed its bottom menu bar to the following:

Bookended by the home button and the user profile button are now video reels, DMs, and the explore page. The create a new post button, which was formerly here in the center, was moved up to the top of the screen in a far less conspicuous place. These changes felt weird at first, but they were, of course, based on real user data. What people do on Instagram these days is watch reels and then share them with their friends. The era of posting beautiful square photos with nice filter edits died a long time ago.

But even today's world of video reels and TikTok videos is in massive flux. AI is flooding the system, and it's impossible to know what is "real" anymore. The name of the game with social media used to be authenticity. This is how individuals gained distribution control from institutions and large brands; they were more real and authentic. But today, we are in a world where AI-generated content can be entirely indistinguishable from "real" or captured content.

I have felt this change myself. As someone who has been a hobby photographer since undergrad some 20+ years ago, I have noticed myself grabbing my Fujifilm camera a lot less over the last year. Instead, I've just been using my phone and spending more time playing around with AI. And, of course, it's not just me. I see my architect and real estate friends using AI to test concepts, create presentation renderings, and more. So, where does all of this leave a platform like Instagram that was designed around individuals creating and sharing their own content?

A few days ago, Adam Mosseri, the head of Instagram, published these twenty slides about how the company sees the world as we head into 2026. They're an interesting read because they mark a shift in messaging. Previously, the narrative was all about connecting the world and empowering creators. Now it's about labeling, mediating, and controlling this new world. In the words of Silicon Valley journalist and entrepreneur Om Malik, "deep down, Instagram is frightened."

But there is a path forward (excerpt also from Malik):

It starts by verifying who is behind an account, embedding provenance in media, and rewarding trust signals. Over time, Meta may tighten control and aim to be an identity broker for everyone. Instagrams want [sic] you to be prepared for this new era of tighter control over identity, authenticity, and content provenance.

One of the most important slides in Mosseri's post for me is this one here:

I've been arguing for years that crypto has an important role to play in a world filled with AI. When nobody knows what is "real" anymore, there's value in being able to say with finality that, hey, this thing over here is authentic and comes from this source. Social media (web2) showed us that people would rather tie something back to an individual instead of a large faceless brand. AI is disrupting this chain of provenance, but I think crypto will bring us back to it, somehow. Whether Instagram will be a part of it, of course, remains to be seen.

Cover photo by Jakob Owens on Unsplash

The best part about making predictions for a year ahead is that at the end of the year you get to look back with humility on what you were thinking at the time and realize how much you missed and how different things turned out.

So, what might happen in 2026?

• Condominium development in Toronto: I think 2026 will be an important turning point year. If I keep saying this, at some point I'll be right, right? 2026 is the first year where we will start to see new condominium completions from the last cycle fall off significantly. Last year (2025), we were projecting nearly 32,000 condominium home completions. This year, it's projected to drop to ~17,487, with 2027 falling off even further as we head to almost no new supply (based on the current pipeline). What I think this means is that the first half of 2026 will still be painful as the market absorbs new inventory and the inventory from 2025 (including unsold units, units in default, and other scenarios), but that things will start to stabilize and feel better toward the end of 2026 and into 2027. New supply will now be delivering below the 10-year average for the first time in many years.

• Purpose-built rental development in Toronto: The story since the condominium market turned in 2022 has been the flip to rental. But not all developers and sites can make this switch and, as I have argued before, the numbers suggest that it won't be enough to offset our dwindling new condominium supply. That said, I think rental rates will remain soft throughout 2026. The supply crunch we're headed toward will need a bit more time to be felt by the market. In the meantime, we will see the highly-amenitized purpose-built rental model fail. The strategy of using over-the-top amenities to drive high rents will finally fall apart in the current market environment. In its place will be a flight to value: boring rental models that offer a quality housing experience at reasonable prices.

• Boutique end-user projects: In markets like Toronto and Vancouver, where the development landscape remains unfavourable, we will see a continued focus on smaller projects and projects catering exclusively to end-users. This demand segment is the most resilient and this re-orientation will help the next development cycle start on more solid footing.

• Foreign buyer ban: The Canadian federal government will relax the foreign buyer ban (which is set to expire on January 1, 2027) and allow foreigners to buy pre-construction homes. There are already rumblings about this so I acknowledge this isn't that bold a prediction. But beyond just relaxing the ban, I think government will start actively courting foreign capital to help solve our housing needs.

• AI bubble: 2026 will be the year that the AI bubble bursts. Not because AI isn't powerful tech that will continue to change the world, but because we are, in the words of investor Howard Marks, in an "inflection bubble." This is different from a fake bubble like Tulip Mania where there was ultimately no underlying reason for tulips to be valued so highly. An inflection bubble is where we get the direction right (AI is a big deal), but the magnitude wrong (shit, we overspent on CapEx). Not every AI company can and will survive. There will only be a select few once the dust settles. And since AI seems to be what's driving the market these days, I think the market will close the end of this year down (measured as the performance of the S&P 500).

• Continued AI adoption: That said, AI will continue to change the way we all live and work. While this is going to put some people out of a job, my bias is an optimistic one in that new technologies tend to create new opportunities and generally grow the overall economy. However, I think that at least two enormous internet-type shifts are underway. One, AI is creating a massive productivity leverage for the people and firms that know how to harness it and, two, the backend of the global financial market is moving "onchain." These are profound shifts that I, unfortunately, think will lead to even more social and political division in the short term. A government somewhere in the world will respond with a universal basic income.

• AI bubble impact on real estate: An AI bubble bursting will generally help the real estate market as investors look for returns somewhere else, with the exception of the data center market. It will also create downward pressure on interest rates (which, in the US, remain the highest they have been since the Great Recession in 2008). As we know, lower rates help boost the values of highly-levered assets like real estate.

• AR/VR/AI for design and construction coordination: I was blown away the first time I tried Apple Vision Pro. It's a magical experience. But it has failed as a consumer product and who knows what Apple will launch next. Regardless, this year we will see clear use cases emerge for AR, VR, and smart glasses. I'd like to see the problems of design and construction coordination get immediately solved because they're massive and costly and they have yet to be solved.

• Mainstream tokenization: In yesterday's post, I spoke about the lack of a breakout consumer-facing web3 app in 2025 (with honourable mention going to the Base app). But perhaps one of the big stories of last year was stablecoins entering the mainstream. Most people now agree they have achieved product-market fit. This is crypto solving real problems (cheap/fast cross-border remittances, payments, etc) with users not needing to think or care about the underlying blockchain technology. In 2026, we will see a noteworthy office building or apartment building get tokenized on the Ethereum blockchain.

• Autonomous vehicles: Last year, I predicted that autonomous vehicles were going to have a year, and it certainly felt that way. This year will be the first year that I ride in one. I came close on a layover in San Francisco in December. I considered leaving the airport and taking one to Apple Park. But I would have been cutting it too close. In 2026, we will see an insurer refuse to cover a human driver for the first time, marking a clear global shift toward autonomy. Already, none of us should be driving cars anymore looking at current safety data.

• Polycentric world: Some have argued that 2025 marked the end of globalization. I'm not sure that is accurate. I think it marked the end of the US-led post-war world order and the acceleration of a more polycentric world order. It was the start of greater US insularity. In 2026, Canada will start to see the benefits of this shift. What it is doing is shaking us out of complacency and forcing us to look east to Europe and west to Asia, as opposed to just south to the US.

What are your predictions for the year ahead?

Towards the end of last year, Meta released SAM 3, which stands for the third version of its Segment Anything Model. The way it generally works is that it allows you to detect, edit, and experiment with things in images and videos. For example, if you were looking at a video of a street, you could ask it to find all the scooters (which I did below), count the number of pedestrians wearing black pants, blur all the license plates on the cars, and so on.

This is immediately useful for a company like Meta because it allows for object-level modifications across its content creation platforms. So if you took a video of someone dancing and you desperately wanted to give them a bobblehead, SAM 3, I'm told, would allow you to quickly do that. Other AI models, such as Gemini, can also segment, but supposedly the SAM models are better and more precise at this specific task.

Beyond bobblehead videos, the potential of this model seems enormous for real estate, cities, and, of course, many other things. Using the above image as an example, you can quickly imagine SAM 3 being used to count and track modal splits across a city, and then make planning decisions based on real-time data.

People are also using it for real estate purposes. Pair the model with satellite images, and you can ask it to tell you how many houses have a pool, which houses recently had their roof replaced (and have solar panels), how many cars are parked on a street, how many cars are parked at Canadian Tire, and the average building lot coverage in an area.

You could also use it to swap out finishes in a real estate listing (including in videos), and get material/area takeoffs ahead of a construction project. I don't know for sure, but I would also imagine that this model would make a great building condition inspector. Come to think of it, I'd love a SAM 3 that could walk our construction sites and document every little detail!

Of course, a lot of these use cases are already being tackled. But the models are getting that much better. And that will lead to even more innovation.

Cover photo by Above Horizon on Unsplash

Sometime last year, Instagram changed its bottom menu bar to the following:

Bookended by the home button and the user profile button are now video reels, DMs, and the explore page. The create a new post button, which was formerly here in the center, was moved up to the top of the screen in a far less conspicuous place. These changes felt weird at first, but they were, of course, based on real user data. What people do on Instagram these days is watch reels and then share them with their friends. The era of posting beautiful square photos with nice filter edits died a long time ago.

But even today's world of video reels and TikTok videos is in massive flux. AI is flooding the system, and it's impossible to know what is "real" anymore. The name of the game with social media used to be authenticity. This is how individuals gained distribution control from institutions and large brands; they were more real and authentic. But today, we are in a world where AI-generated content can be entirely indistinguishable from "real" or captured content.

I have felt this change myself. As someone who has been a hobby photographer since undergrad some 20+ years ago, I have noticed myself grabbing my Fujifilm camera a lot less over the last year. Instead, I've just been using my phone and spending more time playing around with AI. And, of course, it's not just me. I see my architect and real estate friends using AI to test concepts, create presentation renderings, and more. So, where does all of this leave a platform like Instagram that was designed around individuals creating and sharing their own content?

A few days ago, Adam Mosseri, the head of Instagram, published these twenty slides about how the company sees the world as we head into 2026. They're an interesting read because they mark a shift in messaging. Previously, the narrative was all about connecting the world and empowering creators. Now it's about labeling, mediating, and controlling this new world. In the words of Silicon Valley journalist and entrepreneur Om Malik, "deep down, Instagram is frightened."

But there is a path forward (excerpt also from Malik):

It starts by verifying who is behind an account, embedding provenance in media, and rewarding trust signals. Over time, Meta may tighten control and aim to be an identity broker for everyone. Instagrams want [sic] you to be prepared for this new era of tighter control over identity, authenticity, and content provenance.

One of the most important slides in Mosseri's post for me is this one here:

I've been arguing for years that crypto has an important role to play in a world filled with AI. When nobody knows what is "real" anymore, there's value in being able to say with finality that, hey, this thing over here is authentic and comes from this source. Social media (web2) showed us that people would rather tie something back to an individual instead of a large faceless brand. AI is disrupting this chain of provenance, but I think crypto will bring us back to it, somehow. Whether Instagram will be a part of it, of course, remains to be seen.

Cover photo by Jakob Owens on Unsplash

The best part about making predictions for a year ahead is that at the end of the year you get to look back with humility on what you were thinking at the time and realize how much you missed and how different things turned out.

So, what might happen in 2026?

• Condominium development in Toronto: I think 2026 will be an important turning point year. If I keep saying this, at some point I'll be right, right? 2026 is the first year where we will start to see new condominium completions from the last cycle fall off significantly. Last year (2025), we were projecting nearly 32,000 condominium home completions. This year, it's projected to drop to ~17,487, with 2027 falling off even further as we head to almost no new supply (based on the current pipeline). What I think this means is that the first half of 2026 will still be painful as the market absorbs new inventory and the inventory from 2025 (including unsold units, units in default, and other scenarios), but that things will start to stabilize and feel better toward the end of 2026 and into 2027. New supply will now be delivering below the 10-year average for the first time in many years.

• Purpose-built rental development in Toronto: The story since the condominium market turned in 2022 has been the flip to rental. But not all developers and sites can make this switch and, as I have argued before, the numbers suggest that it won't be enough to offset our dwindling new condominium supply. That said, I think rental rates will remain soft throughout 2026. The supply crunch we're headed toward will need a bit more time to be felt by the market. In the meantime, we will see the highly-amenitized purpose-built rental model fail. The strategy of using over-the-top amenities to drive high rents will finally fall apart in the current market environment. In its place will be a flight to value: boring rental models that offer a quality housing experience at reasonable prices.

• Boutique end-user projects: In markets like Toronto and Vancouver, where the development landscape remains unfavourable, we will see a continued focus on smaller projects and projects catering exclusively to end-users. This demand segment is the most resilient and this re-orientation will help the next development cycle start on more solid footing.

• Foreign buyer ban: The Canadian federal government will relax the foreign buyer ban (which is set to expire on January 1, 2027) and allow foreigners to buy pre-construction homes. There are already rumblings about this so I acknowledge this isn't that bold a prediction. But beyond just relaxing the ban, I think government will start actively courting foreign capital to help solve our housing needs.

• AI bubble: 2026 will be the year that the AI bubble bursts. Not because AI isn't powerful tech that will continue to change the world, but because we are, in the words of investor Howard Marks, in an "inflection bubble." This is different from a fake bubble like Tulip Mania where there was ultimately no underlying reason for tulips to be valued so highly. An inflection bubble is where we get the direction right (AI is a big deal), but the magnitude wrong (shit, we overspent on CapEx). Not every AI company can and will survive. There will only be a select few once the dust settles. And since AI seems to be what's driving the market these days, I think the market will close the end of this year down (measured as the performance of the S&P 500).

• Continued AI adoption: That said, AI will continue to change the way we all live and work. While this is going to put some people out of a job, my bias is an optimistic one in that new technologies tend to create new opportunities and generally grow the overall economy. However, I think that at least two enormous internet-type shifts are underway. One, AI is creating a massive productivity leverage for the people and firms that know how to harness it and, two, the backend of the global financial market is moving "onchain." These are profound shifts that I, unfortunately, think will lead to even more social and political division in the short term. A government somewhere in the world will respond with a universal basic income.

• AI bubble impact on real estate: An AI bubble bursting will generally help the real estate market as investors look for returns somewhere else, with the exception of the data center market. It will also create downward pressure on interest rates (which, in the US, remain the highest they have been since the Great Recession in 2008). As we know, lower rates help boost the values of highly-levered assets like real estate.

• AR/VR/AI for design and construction coordination: I was blown away the first time I tried Apple Vision Pro. It's a magical experience. But it has failed as a consumer product and who knows what Apple will launch next. Regardless, this year we will see clear use cases emerge for AR, VR, and smart glasses. I'd like to see the problems of design and construction coordination get immediately solved because they're massive and costly and they have yet to be solved.

• Mainstream tokenization: In yesterday's post, I spoke about the lack of a breakout consumer-facing web3 app in 2025 (with honourable mention going to the Base app). But perhaps one of the big stories of last year was stablecoins entering the mainstream. Most people now agree they have achieved product-market fit. This is crypto solving real problems (cheap/fast cross-border remittances, payments, etc) with users not needing to think or care about the underlying blockchain technology. In 2026, we will see a noteworthy office building or apartment building get tokenized on the Ethereum blockchain.

• Autonomous vehicles: Last year, I predicted that autonomous vehicles were going to have a year, and it certainly felt that way. This year will be the first year that I ride in one. I came close on a layover in San Francisco in December. I considered leaving the airport and taking one to Apple Park. But I would have been cutting it too close. In 2026, we will see an insurer refuse to cover a human driver for the first time, marking a clear global shift toward autonomy. Already, none of us should be driving cars anymore looking at current safety data.

• Polycentric world: Some have argued that 2025 marked the end of globalization. I'm not sure that is accurate. I think it marked the end of the US-led post-war world order and the acceleration of a more polycentric world order. It was the start of greater US insularity. In 2026, Canada will start to see the benefits of this shift. What it is doing is shaking us out of complacency and forcing us to look east to Europe and west to Asia, as opposed to just south to the US.

What are your predictions for the year ahead?