The best thing about this idea is how much I hated it.
Like, at least it's proposing something with an actual impact, because why else would I (or anyone) care about it?
That said, I think everything besides "money" is probably a non starter. I can already spoof an AI video to stare at a camera's input feed for me (and would definitely do that, and would happily spin up a service for normies to do it at 50 cents a pop or whatever).
Cooloffs and probabilistic resets only whack humans, any AI or program filling in and spamming stuff is completely undeterred.
Mailing stuff in and money are probably pretty solid. No spammer at scale can afford to spend $1 per throw on whatever bullshit they're doing, and physical mail should work as long as the thing isn't high value enough to merit services or faking (for instance, there are almost certainly services where you can send them a digital picture of the thing and they'll print and mail it for you, any Staples / Fedex / whoever does this already).
As you point out, the overall idea of friction as a service is mostly about separating people who actually care from tire-kickers, and money or deposits probably does that fine already, and it might be tough to expand to the more digital use cases (or the digital "attention" currency, god help us, how would this not be immediately exploited at scale? Like people don't waste enough of their lives staring at the blinky box already? I probably just don't understand what's being proposed there).
I think the lack of friction is a big reason why social media has such a low signal-to-noise ratio now. Since people can get paid for content now it has added even more incentive to use exploits when vying for attention. I think that Substack has tried to steer clear of this dynamic by making email subscriptions the main channel of interaction, but there is a feed and it seems like more of the casual people I know on Substack default to this mode of operation. My hope is that every platform doesn't have to resort to algorithmic recommendation and feed-style presentation to achieve scale and a sustainable business. It does seem like this is the default end state, though.
I think this is right, and made worse with AI. It's not just the lack of friction to post, it's the lack of friction to create content. There's even a potential issue with the lack of friction to consume content; scrolling a feed absorbs less than sitting with a post you sought out. The challenge is always balancing intent and discoverability; I think Substack gets pretty close, but nobody has quite perfected it yet.
This is why I’m kinda waiting out the current wierdness with AI apps. It’s a clear case of an absurd arms race, and I don’t see how it’s sustainable in the long term: it’ll inevitably cause SOME sort of collapse, and that will be when the REAL opportunities emerge. Right now, only marginal opportunities are available, and the effort to work for them is not worth the actual reward.
There's one area where artificially frictitious technology has already been the subject of intense innovation. I recently had the experience of buying a physical product for my infant daughter, which required signing up for an online subscription to activate, this contrary to the product description — "No Subscription Required!"
When I queried their customer service rep regarding this apparent discrepancy, I was informed that I was welcome to sign up for a one-month free trial, and so long as I canceled said trial before the 30 days expired, I could continue using the product without being charged.
On a similar note, ever try to cancel a gym membership? Or an Adobe account (we're sorry to see you go! Now, to finalize the process, simply answer these riddles three...)?
If UFIs ever become a thing, I suspect Planet Fitness product managers will suddenly be a hot commodity.
This is wonderful. It does, however, sound like over-engineering of pre-existing analog processes; i.e going full circle. If we were to just bypass the UFI idea and revert to making analog processes the more rational choice, what do you think could be the challenges?
That's a great question! I think there are two big challenges:
1. Friction, for all its benefits, is painful and expensive. Analog processes mostly died out because there are a lot of small pain points in phone tag, paper resumes, etc. It's hard to fully reintroduce analog friction once it's gone, even if it has useful second-order effects.
2. Platforms like Resy/LinkedIn/etc are both friction removers AND discovery devices. For a lot of people, finding a restaurant and booking it are the same action. If you're mass-market, opting out of Resy, LinkedIn, etc adds the risk that nobody finds you at all.
That said, there's totally a segment of the market for whom that's a feature, not a bug. There will probably be more restaurants that intentionally stay off the platforms, curate a cast of regulars who feel like they're in an exclusive club, and spread through word of mouth.
I see. It’s truly ironic the grave we’ve dug ourselves into, huh? 😅 Let’s hope for the best, I’ve got a good feeling where 2026 is headed especially with good ideas like this spreading on Substack!
It's a better idea than many might think! My decade+ of A/B testing has seen plenty of instances where intentionally adding friction surprisingly improved outcomes, often because it increased the perceived value of the thing people were trying to do/get. Maybe it made the thing seem more exclusive, higher-quality, or even more personalized (e.g. I've run tests where inquiries/purchases increased when gating search results behind a short questionnaire - users felt the results were more relevant, even before we used the info they provided in the search algorithm)
That is super interesting - did the questionnaire actually change results, or was the difference purely the flow?
The connection between effort and perceived value is probably going to get clearer over time. Everest was a bigger flex before you could pay for an organized trip that handles everything. That dynamic probably isn't top of mind to an end user making a purchase, but over time it's easy to imagine high-value products requiring effort as a status signal (or in the case of something like a survey, a subtle validation that they're valued).
As always, enjoyed reading and thinking about this one!
My sense is that you’d have to approach this in a more universal way, in that perversely any solution for optimal friction needs to have minimal friction itself to get adopted.
Social credit scores have a stigma and are thus likely a non-starter. But proof-of-work was (in some sense) originally an attempted solution for the lack of friction in email leading to too much spam. You could in theory create a currency that can only be earned by various forms of accessible but non-automatable work. Users could then use these units as refundable deposits. It’s nicer than just using money, as it treats everyone’s time at par. And as with CAPTCHA you could conceivably make the underlying tasks additive to some real world good.
Really like this framing - one of the interesting implications with a proof of work approach is that a lot of the existing human centric proof of work approaches (like CAPTCHAs!) seem likely to (eventually) break with AI.
You could do some kind of human verification -> set number of tokens, but if everyone gets equal levels of currency I have to imagine a secondary market pops up with prices per unit. That actually might be more fair than the status quo - everyone gets some level of free actions, and then you pay if you're an excess user - but I can imagine it could lead to some weird outcomes depending how universal it is. It's sort of a subsidy for people who don't do much and a tax on people who do a lot of things?
I really like the idea of making friction do some real world good - having people stare into the camera is good for allocating resources, but probably a real deadweight loss to the world. I wonder if you could feasibly build something that tells people "go volunteer at a pet shelter, get some tokens that let you cut to the front of digital lines." There could be something there!
"A UFI would simplify the process of adding friction back into digital life." A hilarious line on its own.
I love this idea. It is almost like the "Check the box to confirm you are not a robot", but "stare into this webcam to confirm you actually care about this"
The best thing about this idea is how much I hated it.
Like, at least it's proposing something with an actual impact, because why else would I (or anyone) care about it?
That said, I think everything besides "money" is probably a non starter. I can already spoof an AI video to stare at a camera's input feed for me (and would definitely do that, and would happily spin up a service for normies to do it at 50 cents a pop or whatever).
Cooloffs and probabilistic resets only whack humans, any AI or program filling in and spamming stuff is completely undeterred.
Mailing stuff in and money are probably pretty solid. No spammer at scale can afford to spend $1 per throw on whatever bullshit they're doing, and physical mail should work as long as the thing isn't high value enough to merit services or faking (for instance, there are almost certainly services where you can send them a digital picture of the thing and they'll print and mail it for you, any Staples / Fedex / whoever does this already).
As you point out, the overall idea of friction as a service is mostly about separating people who actually care from tire-kickers, and money or deposits probably does that fine already, and it might be tough to expand to the more digital use cases (or the digital "attention" currency, god help us, how would this not be immediately exploited at scale? Like people don't waste enough of their lives staring at the blinky box already? I probably just don't understand what's being proposed there).
I think the lack of friction is a big reason why social media has such a low signal-to-noise ratio now. Since people can get paid for content now it has added even more incentive to use exploits when vying for attention. I think that Substack has tried to steer clear of this dynamic by making email subscriptions the main channel of interaction, but there is a feed and it seems like more of the casual people I know on Substack default to this mode of operation. My hope is that every platform doesn't have to resort to algorithmic recommendation and feed-style presentation to achieve scale and a sustainable business. It does seem like this is the default end state, though.
I think this is right, and made worse with AI. It's not just the lack of friction to post, it's the lack of friction to create content. There's even a potential issue with the lack of friction to consume content; scrolling a feed absorbs less than sitting with a post you sought out. The challenge is always balancing intent and discoverability; I think Substack gets pretty close, but nobody has quite perfected it yet.
This is why I’m kinda waiting out the current wierdness with AI apps. It’s a clear case of an absurd arms race, and I don’t see how it’s sustainable in the long term: it’ll inevitably cause SOME sort of collapse, and that will be when the REAL opportunities emerge. Right now, only marginal opportunities are available, and the effort to work for them is not worth the actual reward.
Love this.
There's one area where artificially frictitious technology has already been the subject of intense innovation. I recently had the experience of buying a physical product for my infant daughter, which required signing up for an online subscription to activate, this contrary to the product description — "No Subscription Required!"
When I queried their customer service rep regarding this apparent discrepancy, I was informed that I was welcome to sign up for a one-month free trial, and so long as I canceled said trial before the 30 days expired, I could continue using the product without being charged.
On a similar note, ever try to cancel a gym membership? Or an Adobe account (we're sorry to see you go! Now, to finalize the process, simply answer these riddles three...)?
If UFIs ever become a thing, I suspect Planet Fitness product managers will suddenly be a hot commodity.
This is wonderful. It does, however, sound like over-engineering of pre-existing analog processes; i.e going full circle. If we were to just bypass the UFI idea and revert to making analog processes the more rational choice, what do you think could be the challenges?
That's a great question! I think there are two big challenges:
1. Friction, for all its benefits, is painful and expensive. Analog processes mostly died out because there are a lot of small pain points in phone tag, paper resumes, etc. It's hard to fully reintroduce analog friction once it's gone, even if it has useful second-order effects.
2. Platforms like Resy/LinkedIn/etc are both friction removers AND discovery devices. For a lot of people, finding a restaurant and booking it are the same action. If you're mass-market, opting out of Resy, LinkedIn, etc adds the risk that nobody finds you at all.
That said, there's totally a segment of the market for whom that's a feature, not a bug. There will probably be more restaurants that intentionally stay off the platforms, curate a cast of regulars who feel like they're in an exclusive club, and spread through word of mouth.
I see. It’s truly ironic the grave we’ve dug ourselves into, huh? 😅 Let’s hope for the best, I’ve got a good feeling where 2026 is headed especially with good ideas like this spreading on Substack!
It's a better idea than many might think! My decade+ of A/B testing has seen plenty of instances where intentionally adding friction surprisingly improved outcomes, often because it increased the perceived value of the thing people were trying to do/get. Maybe it made the thing seem more exclusive, higher-quality, or even more personalized (e.g. I've run tests where inquiries/purchases increased when gating search results behind a short questionnaire - users felt the results were more relevant, even before we used the info they provided in the search algorithm)
That is super interesting - did the questionnaire actually change results, or was the difference purely the flow?
The connection between effort and perceived value is probably going to get clearer over time. Everest was a bigger flex before you could pay for an organized trip that handles everything. That dynamic probably isn't top of mind to an end user making a purchase, but over time it's easy to imagine high-value products requiring effort as a status signal (or in the case of something like a survey, a subtle validation that they're valued).
As always, enjoyed reading and thinking about this one!
My sense is that you’d have to approach this in a more universal way, in that perversely any solution for optimal friction needs to have minimal friction itself to get adopted.
Social credit scores have a stigma and are thus likely a non-starter. But proof-of-work was (in some sense) originally an attempted solution for the lack of friction in email leading to too much spam. You could in theory create a currency that can only be earned by various forms of accessible but non-automatable work. Users could then use these units as refundable deposits. It’s nicer than just using money, as it treats everyone’s time at par. And as with CAPTCHA you could conceivably make the underlying tasks additive to some real world good.
Really like this framing - one of the interesting implications with a proof of work approach is that a lot of the existing human centric proof of work approaches (like CAPTCHAs!) seem likely to (eventually) break with AI.
You could do some kind of human verification -> set number of tokens, but if everyone gets equal levels of currency I have to imagine a secondary market pops up with prices per unit. That actually might be more fair than the status quo - everyone gets some level of free actions, and then you pay if you're an excess user - but I can imagine it could lead to some weird outcomes depending how universal it is. It's sort of a subsidy for people who don't do much and a tax on people who do a lot of things?
I really like the idea of making friction do some real world good - having people stare into the camera is good for allocating resources, but probably a real deadweight loss to the world. I wonder if you could feasibly build something that tells people "go volunteer at a pet shelter, get some tokens that let you cut to the front of digital lines." There could be something there!
Supercharge this idea by using advertising to add friction.
"A UFI would simplify the process of adding friction back into digital life." A hilarious line on its own.
I love this idea. It is almost like the "Check the box to confirm you are not a robot", but "stare into this webcam to confirm you actually care about this"