I believe that creating a successful consumer product can be attributed to somewhere between 70-90% luck and 10-30% executing against that luck. A lot of things have to go right: timing, product, distribution, zeitgeist, etc. So much of it is stumbling your way into capturing lightning in a bottle or indestructible product market fit.
This may seem disheartening, but the good news is that it's possible to optimize for getting lucky. When building consumer products, I think one of the best ways to increase the chance of doing this is by building for yourself. Building products and companies is hard, so if you're building a consumer internet application, you might as well build something you want to use yourself. That way, by the end of the day, you are making yourself happy. And that's a win.
Building for yourself means that you are your target customer. You can prioritize the features that you find most useful or fun. Your roadmap is all the things you want to exist. Nobody knows your product or the problem you're trying to solve better than you. And if the product you're building can't make you happy or satisfy your needs, there's little chance it will do so for others.
I've seen a lot of people think about building or trying to build consumer internet applications by drastically overthinking things. They're midwitting their way into a problem that doesn't exist, a product nobody cares about, or an idea they think the world needs based on some hole they perceive to exist in Facebook Marketplace or Events or fill in the blank. This is not the way. Build something you want to exist and use yourself, not something that appears to be some opportunistic arbitrage.
Many durable consumer products take some time to truly inflect, and it takes even longer to figure out how to build and scale the business. On top of that, it's difficult to maintain your energy over that period of time. So you better be enjoying the ride and the product you're building. Build for yourself.
For the past four years my family has spent most of July and August in the Catskills in a town called Accord, NY. We love it there and ended up buying a place here like a lot of Brooklyn transplants did during Covid. Summers are magical up there
One of the things I always look forward to is using the garage gym I have assembled together over the years. Once we head up, I restart a program called TrainFTW. It's a workout regimen organized by an ex-crossfit athlete named Matt Chan and his wife Cherie. I'd describe it as crossfit-lite. It's a mixture of full-body workouts that span weight training, Zone 2, HIIT, and some cardio-focused work, too. More than anything, the premise is to enjoy the benefits of crossfit without getting injured. By a long shot, it's my favorite type of workout.
I also mix it up and play tennis and mountain bike to get a variety of different cardio. I used to road bike more often, but mountain biking is a lot more fun and there are no cars. I love this summer setup for fitness. It's a highlight of my year.
Things in Brooklyn are definitely different. I feel like exercising is more of something I do to check the box than it is fun. I used to be part of a gym and do TrainFTW, but I found it takes up too much time and need something I can just get done in 45-50 minutes in the early morning. Normally I will switch between using the Tonal machine in our basement and riding the Peloton. On the Tonal I have 7 workouts I cycle through that I programmed myself, and on the Peloton I just do
This will be the last generation of global movie superstars.
In the not-too-distant future, everyone will have the tools they need to create a completely AI-generated film that is indistinguishable in quality from a traditional Hollywood movie. This includes the whole shebang: from the main characters and actors to costume and set design. Our only limitations will be the constraints of our imagination. But even that will be augmented with technology. It is easy to draw a line between where we are today with the capabilities of AI tools to where we will be by the end of the decade. And while these statements may seem hyperbolic, the rate of progress across generative AI is accelerating. Things happen gradually, then suddenly. We are approaching the inflection point.
This type of sea change, while unsettling for those who have mastered the status quo, holds immense potential for the broader population. The proliferation of new technologies and tools not only empowers a class of emergent creators but also lowers the barrier to participation and expands the market for industries. This is a positive shift - broadening access to the arts, expression, and creativity to everyone across the globe with a computer and an internet connection - but it disrupts the current norms.
Why would a movie studio build a franchise around a human actor when it can use an AI-generated actor it has created and owns as IP and can uniquely customize across geographies and demographics? Is that dystopian and inhumane, or is it the obvious future? Denis Villeneuve’s recent Dune series is a masterpiece, but the second installment cost $190m to produce. How many orders of magnitude will that production cost come down as people in remote corners of Earth will have access to tools that enable them to create something of equal quality on their phones? Is that an absurd question, an implausible reality, or an inevitable one?
I believe that creating a successful consumer product can be attributed to somewhere between 70-90% luck and 10-30% executing against that luck. A lot of things have to go right: timing, product, distribution, zeitgeist, etc. So much of it is stumbling your way into capturing lightning in a bottle or indestructible product market fit.
This may seem disheartening, but the good news is that it's possible to optimize for getting lucky. When building consumer products, I think one of the best ways to increase the chance of doing this is by building for yourself. Building products and companies is hard, so if you're building a consumer internet application, you might as well build something you want to use yourself. That way, by the end of the day, you are making yourself happy. And that's a win.
Building for yourself means that you are your target customer. You can prioritize the features that you find most useful or fun. Your roadmap is all the things you want to exist. Nobody knows your product or the problem you're trying to solve better than you. And if the product you're building can't make you happy or satisfy your needs, there's little chance it will do so for others.
I've seen a lot of people think about building or trying to build consumer internet applications by drastically overthinking things. They're midwitting their way into a problem that doesn't exist, a product nobody cares about, or an idea they think the world needs based on some hole they perceive to exist in Facebook Marketplace or Events or fill in the blank. This is not the way. Build something you want to exist and use yourself, not something that appears to be some opportunistic arbitrage.
Many durable consumer products take some time to truly inflect, and it takes even longer to figure out how to build and scale the business. On top of that, it's difficult to maintain your energy over that period of time. So you better be enjoying the ride and the product you're building. Build for yourself.
For the past four years my family has spent most of July and August in the Catskills in a town called Accord, NY. We love it there and ended up buying a place here like a lot of Brooklyn transplants did during Covid. Summers are magical up there
One of the things I always look forward to is using the garage gym I have assembled together over the years. Once we head up, I restart a program called TrainFTW. It's a workout regimen organized by an ex-crossfit athlete named Matt Chan and his wife Cherie. I'd describe it as crossfit-lite. It's a mixture of full-body workouts that span weight training, Zone 2, HIIT, and some cardio-focused work, too. More than anything, the premise is to enjoy the benefits of crossfit without getting injured. By a long shot, it's my favorite type of workout.
I also mix it up and play tennis and mountain bike to get a variety of different cardio. I used to road bike more often, but mountain biking is a lot more fun and there are no cars. I love this summer setup for fitness. It's a highlight of my year.
Things in Brooklyn are definitely different. I feel like exercising is more of something I do to check the box than it is fun. I used to be part of a gym and do TrainFTW, but I found it takes up too much time and need something I can just get done in 45-50 minutes in the early morning. Normally I will switch between using the Tonal machine in our basement and riding the Peloton. On the Tonal I have 7 workouts I cycle through that I programmed myself, and on the Peloton I just do
This will be the last generation of global movie superstars.
In the not-too-distant future, everyone will have the tools they need to create a completely AI-generated film that is indistinguishable in quality from a traditional Hollywood movie. This includes the whole shebang: from the main characters and actors to costume and set design. Our only limitations will be the constraints of our imagination. But even that will be augmented with technology. It is easy to draw a line between where we are today with the capabilities of AI tools to where we will be by the end of the decade. And while these statements may seem hyperbolic, the rate of progress across generative AI is accelerating. Things happen gradually, then suddenly. We are approaching the inflection point.
This type of sea change, while unsettling for those who have mastered the status quo, holds immense potential for the broader population. The proliferation of new technologies and tools not only empowers a class of emergent creators but also lowers the barrier to participation and expands the market for industries. This is a positive shift - broadening access to the arts, expression, and creativity to everyone across the globe with a computer and an internet connection - but it disrupts the current norms.
Why would a movie studio build a franchise around a human actor when it can use an AI-generated actor it has created and owns as IP and can uniquely customize across geographies and demographics? Is that dystopian and inhumane, or is it the obvious future? Denis Villeneuve’s recent Dune series is a masterpiece, but the second installment cost $190m to produce. How many orders of magnitude will that production cost come down as people in remote corners of Earth will have access to tools that enable them to create something of equal quality on their phones? Is that an absurd question, an implausible reality, or an inevitable one?
Ride It to the Sky
Ride It to the Sky
4x4 Zone 5 training
at least once a week. I will also try to play tennis once a week after work.
My issue with these workouts back in the city is I get bored of them rather quickly. Lifting weights in a basement isn't necessarily "fun" for me. Sometimes I will mix things up and get into F45 for a couple of weeks or months and that will help. But I've found it's hard to get as excited about fitness when I'm in the city. The best thing for me has simply been to mix it up to stay motivated. I hope to one day be as enthusiastic about my fitness program in the city as I am upstate.
This premise doesn’t just extend to film - it applies to all media. We are already seeing it happen with art, images, text, and music. Video is just an extension of these, and new formats that didn’t exist before will follow. The implications are hard to fathom, but here is a spectrum of scenarios for the future of media and entertainment, all of which will coexist:
We are moving towards a world where content is auto-generated specifically for the sole purpose of entertaining individuals. It will be uniquely customized to suit their needs, tastes, and preferences without the need for explicit definition. Today, companies like TikTok use other people’s content to entertain a mass audience. They, like other scaled (sometimes social) media platforms, understand what we want to watch more than we know what we want to watch, and they capture our attention by surfacing addictive content. In this future, why would a media platform want to be dependent on a group of human creators when it can auto-generate all media itself with AI? We are rapidly approaching this world. Content - from music to video to memes and beyond - can be spun up instantly and uniquely customized to every individual and then immediately discarded. The endless scroll will be truly infinite. Scroller beware of the implications.
Humans will embrace AI tools to augment their creativity and explore new artistic frontiers. We’ve come a long way since the days of cavemen painting pictures with twigs and pigment blocks. We are now armed with tools that expand our capabilities and creativity to new dimensions. In 1997, world chess champion Gary Kasparov was the first knowledge worker/athlete/artist (pick your own descriptor) to lose his job to AI when he was defeated by Deep Blue. Since then, it has been believed that between human chess players and AI chess engines, the most skillful player is actually a combination of human and AI, or a chess cyborg. Media is experiencing its chess moment, and creators now have AI partners to help them push into new frontiers. Over the past two decades, some of the largest and most powerful networks in the world emerged around media. These networks built around “social media” grew on the backs of user-generated content - photographs from point-and-shoot cameras, 140-character quips, memes, videos taken with smartphones, entries on message boards, and more. The combination of humans and AI is going to produce new atomic units of media and expression, and around these novel (and weird) formats will emerge tomorrow’s global networks.
There will always be an important market for authentically human creations, and as AI-generated media floods the digital realm, the value of “deep reals” will grow. Live experiences that showcase human talent will counterbalance the manipulation of bits, and its scarcity will drive demand. The same goes for unassisted, human-created media. While these formats may be more niche and never achieve the reach or ubiquity that their AI-generated or AI-assisted counterparts will - a human-directed, produced, and acted film may never have the same “popularity” (ie views, impressions, etc.) as a perfectly optimized-for-engagement AI-generated movie - they will still be critical foundations of culture.
Media exists on a spectrum, and the introduction of AI tools is stretching and adding new dimensions to it. Innovative tooling is critical, but I am most interested in the networks that emerge around these novel “Human + AI” formats, as well as the experiences and infrastructure that help deep reals thrive. While this may feel like a particularly daunting moment for those who have thrived in the status quo, I'm optimistic about new opportunities for exploration and the ever-expanding canvas of expression and creativity that can fit in the hands of everyone across the globe.
4x4 Zone 5 training
at least once a week. I will also try to play tennis once a week after work.
My issue with these workouts back in the city is I get bored of them rather quickly. Lifting weights in a basement isn't necessarily "fun" for me. Sometimes I will mix things up and get into F45 for a couple of weeks or months and that will help. But I've found it's hard to get as excited about fitness when I'm in the city. The best thing for me has simply been to mix it up to stay motivated. I hope to one day be as enthusiastic about my fitness program in the city as I am upstate.
This premise doesn’t just extend to film - it applies to all media. We are already seeing it happen with art, images, text, and music. Video is just an extension of these, and new formats that didn’t exist before will follow. The implications are hard to fathom, but here is a spectrum of scenarios for the future of media and entertainment, all of which will coexist:
We are moving towards a world where content is auto-generated specifically for the sole purpose of entertaining individuals. It will be uniquely customized to suit their needs, tastes, and preferences without the need for explicit definition. Today, companies like TikTok use other people’s content to entertain a mass audience. They, like other scaled (sometimes social) media platforms, understand what we want to watch more than we know what we want to watch, and they capture our attention by surfacing addictive content. In this future, why would a media platform want to be dependent on a group of human creators when it can auto-generate all media itself with AI? We are rapidly approaching this world. Content - from music to video to memes and beyond - can be spun up instantly and uniquely customized to every individual and then immediately discarded. The endless scroll will be truly infinite. Scroller beware of the implications.
Humans will embrace AI tools to augment their creativity and explore new artistic frontiers. We’ve come a long way since the days of cavemen painting pictures with twigs and pigment blocks. We are now armed with tools that expand our capabilities and creativity to new dimensions. In 1997, world chess champion Gary Kasparov was the first knowledge worker/athlete/artist (pick your own descriptor) to lose his job to AI when he was defeated by Deep Blue. Since then, it has been believed that between human chess players and AI chess engines, the most skillful player is actually a combination of human and AI, or a chess cyborg. Media is experiencing its chess moment, and creators now have AI partners to help them push into new frontiers. Over the past two decades, some of the largest and most powerful networks in the world emerged around media. These networks built around “social media” grew on the backs of user-generated content - photographs from point-and-shoot cameras, 140-character quips, memes, videos taken with smartphones, entries on message boards, and more. The combination of humans and AI is going to produce new atomic units of media and expression, and around these novel (and weird) formats will emerge tomorrow’s global networks.
There will always be an important market for authentically human creations, and as AI-generated media floods the digital realm, the value of “deep reals” will grow. Live experiences that showcase human talent will counterbalance the manipulation of bits, and its scarcity will drive demand. The same goes for unassisted, human-created media. While these formats may be more niche and never achieve the reach or ubiquity that their AI-generated or AI-assisted counterparts will - a human-directed, produced, and acted film may never have the same “popularity” (ie views, impressions, etc.) as a perfectly optimized-for-engagement AI-generated movie - they will still be critical foundations of culture.
Media exists on a spectrum, and the introduction of AI tools is stretching and adding new dimensions to it. Innovative tooling is critical, but I am most interested in the networks that emerge around these novel “Human + AI” formats, as well as the experiences and infrastructure that help deep reals thrive. While this may feel like a particularly daunting moment for those who have thrived in the status quo, I'm optimistic about new opportunities for exploration and the ever-expanding canvas of expression and creativity that can fit in the hands of everyone across the globe.
