Friction remover

One way to think about AI is as a friction remover.

I was just listening to a podcast from Union Square adventures where Fred Wilson talks about how AI has the potential to create interoperability in places where it previously has not existed. I can think of many examples. For a good while I have been thinking about trying to use AI to line up my photo albums across my different photo storage services: Amazon photos, Google photos, Flicker, Apple photos.

All of these photos services have the same photos, but they have different albums. I would love to go through Google photos and reconstruct the albums that I made in 2007 in 2008 on Flicker.

Actually, I would not like to do that, but I would like to have it done for me! I haven’t tried yet, because a year ago I asked Harper Reed if this could be done and he said it was actually a lot harder than it sounded to identify the same photo across different photo services. But maybe now it’s time to try since the AI tools are a little better.

My personal example aside, the point is that AI is a friction remover. There are lots of products that rely on friction to create switching costs or lock in for the users. And it’s not just products that rely on friction — some industries rely on it! For example, there is a lot of intentional friction in filing US taxes.

AI is likely to change all of that over the next few years. One of my partners runs a lab at UW and some students in the lab built an AI tool that walks through the Comcast phone tree and threatens to leave in order to secure a better price for internet and cable services.

If your product design relies on discouraging users from doing something that they really want to do by creating friction, those barriers are not likely to hold up well.