I often feel like I live two lives.
My day-to-day is spent either building new products and playing with AI, or following the timeline on Twitter/X, studying how people from around the world are being pulled toward the new frontier. People who are all trying, in their small ways, to contribute to building the future. Then there is the reality around me. Most people I meet have barely scratched the surface of what AI can do for them, or simply can't be bothered. I recently met someone who works at a family-run SME that still prints out emails for their archive. In 2026.
You can draw whatever conclusion you want from that. Here is mine.
Twitter is an echo chamber of the top 0.1%. Most of the world is not there yet, and nobody really knows how long the diffusion of new technology will take to reach all corners of daily life and the economy. Tech progress has always moved faster in the minds of the people building it than in the lives of the people it is intended to serve.
At the same time, working in tech gives you a glimpse of the future before almost everyone else. You use tools that will, sooner or later, become accessible enough to change how ordinary people live and work. You see what is coming more clearly than most. It is like looking at the future through a blurry glass. Every once in a while you see something that gives you real conviction about how things will play out, but most of the time it is not clear. Tech has always been the engine of progress, and the people closest to the engine feel the heat first.
That proximity has led me to believe that everything that can be verified will eventually be solved by LLMs. That is actually a thought from Dario Amodei, who runs Anthropic. The playbook is out there, and the race is unlike anything before it, because the potential reward dwarfs even what the last generation of big tech unlocked. As Naval says, English is the new programming language, so everyone can build. And today's real developers are the researchers building the models.
If that is true, then the easy problems in software will be handled extremely well by AI, and whenever something becomes easy, its value depreciates. The returns will flow to whoever is working on the things that are still hard.
Which means all of us, if we care about contributing something meaningful, need to move up the difficulty curve.
What we consider hard today was often borderline impossible just a few years ago. The frontier has shifted. Problems that once required decades of institutional effort or resources only governments could deploy now have a real shot at being solved by a small team with the right insight, competence, and enough conviction. Think of Peter Steinberger building OpenClaw alone.
That is one of the reasons I built Questd.
I kept feeling lost. The world is big and complex and messy, and I had no good answer to the simplest question: where do I even start? What is actually worthy of my time? Which old problems now have a real shot at being solved within my lifetime?
Questd is designed to answer that question. To help you find the problems that genuinely matter, so you can take your most precious resource, time, and pour it into something harder and more important than what the default path would have you working on.
In the process, maybe, make the world a little better.