Literally nobody asked, but here’s my prediction: 2026 will be the year of friction.
The grand promise underlying all technology is to make our lives easier. In other words - to reduce friction. Press a button instead of gathering wood for a fire. Tap a screen instead of riding a horse for two weeks to reach the city. You don’t even need a sponge or run out of hot water because you ordered the smart water heater.
And then 2025 delivered the ultimate promise - Automate everything. Automate the automations. Just write a line and it’s done for us. Just think and it’s done for us. An army of autonomous agents doing everything and we’re not needed anymore. It was also the year we sobered up from that promise.
2026 will be the year we take all the hype and endless tools and infrastructure and push, push, push, until we manage to take that cool POC we built at a hackathon and turn it into a real product for production. It’s not exactly cool technology, though there will definitely be cool technologies involved.
It will be the year where in every possible way, with sophisticated tools and without them, we actually deal with all the bugs. We deal with scale and with tokens running out and do ablations and cut 90% of the mcps and add real security (not just throw another guardrail) and look at monitoring every day and tag a ton of data and get customer feedback and do hundreds of iterations and read every line to make sure it works exactly the way we want it to.
The biggest problem with GenAI is that it gives us the illusion that everything is easy. But anyone who’s ever tried to achieve something knows that making something big work is fucking hard. And despite the noise - AI is no different. The big winners of 2026 are going to be those who are willing to work hard, and smart, to make it work.
