Logo of Abhijeet Dwivedi

Jeet's Law Of Skeuomorphic Thinking

When adapting to new technology, we always start by recreating the familiar past instead of imagining entirely new ideas. You should see 👇 guy's talk on this topic.

steve jobs at aspen

Early TV was basically filmed radio shows. It took time to discover what made TV truly unique: visual storytelling (show don't tell), live broadcasts, shared cultural moments.

When mobile apps first exploded, every big company executive was asking "What's our mobile strategy" - and what happened? Most companies essentially shrunk their websites into phone screens, checking the mobile box without understanding the platform's native potential.

Then products like Uber came that could only exist on mobile. They weren't trying to fit a web business model into mobile - they were creating something entirely new that couldn't exist on the web (not as well).

Jeet's Law Of Skeuomorphic Thinking is the idea that before we make truly native things on a new platform, we go through a phase where we try to make old familiar things first. It's like our collective imagination needs some time to warm up when new technology arrives. First we copy what we know, then slowly discover deep ideas about what makes the new platform unique.

Early cars were designed like horseless carriages. Early electric lights in homes had the same form as gas lamps. Early smartphones had physical keyboards. By the way, even software follows this pattern. The save icon is still a floppy disk. Our documents icon still looks like paper. Our desktop has files and folders - all metaphors from the physical world that we still carry forward.

And now we're in the middle of an AI revolution. And guess what we're doing? The same thing! We're asking AI to write emails like humans, generate images that look like human art, and automate existing workflows. These are all "horseless carriage" versions of AI.

What could only exist now because of AI? Perhaps systems that continuously evolve their understanding as they interact with the world - no human could maintain that kind of constant learning. Or interfaces that adapt to your usage patterns rather than forcing you to adapt to them. Or collaborative problem-solving where AI doesn't just answer questions but gives entirely novel approaches to problems. Even as I write this, you can probably tell that I am limited in my imagination.

The most exciting AI applications won't be versions of things we already do - they'll be things we've never imagined before. I am eagerly waiting for true AI-native applications.