Permissionless Innovation — Beyond the Hype & Doom of AI.
urge anyone who wants to build a career even closely related to the knowledge industry to commit to what is happening right now and make AI serve you rather than being at the mercy of it.
The best way to to do that is to first understand what is happening and what is not happening, to separate the signal from the noise and not become overly reactionary. Hype and doom will both lead you down very different, but evolutionary dead ends.
The first thing you need to do is to stop listeting to the news and instead equip yourself with a more philosophical/strategic understanding of what’s going on. The second big thing I would suggest is that you stop thinking about AI or AI agents as language models since much of that is overhyped.
A much better lens is to look at what they can do as language models: that can code and have a runtime environment to test their logic up against. That’s the killer combo. That’s the lens you need to get comfortable with, to understand what’s going on.
LLMs have not fundamentally changed much since GPT 3.5. What has improved though is the scale and quality of the training. The models are not smarter or more rational or any more intelligent, they are simple better trained, and that distinction matters a lot.
Realize three things:
1. Language models can learn anything we have the knowledge to teach them. In principle, there is nothing they cannot learn with one important caveat; Language models can only learn what we have the knowledge to teach them. They cannot create genuinely new knowledge.
2. Language models are not rational but you are. That is where the opportunity lies. You have the problems, the needs, the ideas, the insights, the lived experience, the relationships, the foresight and most importantly the ability to create new knowledge.
3. AI agents that code can do more than just code. Because agents can be taught to code, and because computer science has decades of verification methods built on logic, you can get surprisingly close to building self evolving assistants that help solve your daily problems and act as if it’s rational. Doesn’t matter if it’s sanitizing a csv file, turning it into a database, cleaning up your computer folders or building your own email signup flow. Any chore that you wish you had a solution for is solvable by yourself. Doesn’t have to be an application.
Whats more importantly you can start from anywhere. You can build from scratch, build on top of existing solutions, or refine what you already have. With that, and with the growing ability of language models to write and run more and more advanced code, there are very few problems you cannot tackle alone or with a small team. That is why the knowledge industry is being impacted first.
The opportunity is to find the problems and design patterns and create the knowledge AI cannot produce. Through curiosity and research, through network effects and distribution, through novel ideas and long term conviction, but above all through building and experimenting.
If you have a creative mind, once you get that first aha moment like Jules did, you are going to feel almost intoxicated by the opportunities and overwhelmed by the ability to build without having to ask someone else for permission or to help you.
Execution is now the commodity, big 2nd or 3rd order ideas are what will be scarce. In many ways we are still thinking too small, too scared of committing to the process and let these coding agents explore things at our direction. But the future of information work is more likely going to be a giant real time strategy game letting you play out scenarios, experiments, test, design, validate, simulate, deploy and most importantly, connect.
This may be difficult to accept for many developer, marketeers, scientists, academics and creative, even for me at times, but in many ways we stand in the way of AI’s potential. We are the bottleneck but that also means we decide what part of that potential becomes reality.