Amara’s Law says that whenever there is a new technology, we tend to overestimate its impact in the short term and underestimate it in the long term. I saw firsthand when this happened in crypto: this is seen as revolutionary ideas whose primitives don’t exist yet, which is going full speed ahead, trying to solve it on-the-fly; and, in retrospect, they are ideas that were ahead of their time.
This is not negative; It is normal in the sense that, every time new technological primitives emerge, we have to rethink the status quo as a society and decide if we are willing to update it, or if we rather want to preserve it.
This was the case with DAOs, Proofs of Humanity, and certainly also with Smart Contracts. Perhaps the latter are the best example: revolutionary technology that works well, but, on a human level, is difficult to understand, just as the “cloud” was at the time and perhaps “AI” today. Technology lovers like me understand them and use them simply because we are curious about them. But not all of us are early adopters. There are institutional clients, governments, regulators; all of them with valid questions that rarely have obvious answers. This back-and-forth between the order of things and progress is what marks the real impact of technologies.
Technologies themselves sometimes face limits; for example, the GPU shortage that has existed since the times of crypto, and is now repeated in the times of AI. Hype comes and goes, but technology continues to advance. As the technological and social challenges are resolved (regulation, understanding of technology, track record), it is incorporated into the new world.
In this sense, for those of us who are always building, everything makes sense: for there to be agentic b2c commerce, you need infrastructure that does not fail: fast, secure and programmable digital payments; Resilient Language Models with uptime comparable to other mainstream technologies and larger context windows; secure user interfaces, based on full consent and intent; etc These are not trivial technological problems, nor has the social sphere reached a consensus; For this reason, it cannot be immediate.
This is not a discouraging post; I think it is, in your case, quite the opposite. I have seen it being built right now, in an incredible time. Every month there is a hackathon around the world where someone is pushing an LLM to its limits, looking for points of failure and writing solutions around it.
Mitchell Hashimoto, the creator of, among other things, Terraform, talks about how there are more containers that have been deployed in recent months than in the rest of history, because the Claude Codes and v0s of the world need them. I myself have found that it is very difficult to design these types of systems without container orchestration.In short, I believe that the path to the new internet will not be easy, but it will not stop either. It’s a great time to be a builder, but we have to keep an open mind and be very careful of the marketing and propaganda that promise a panacea for tomorrow. Software development has never been so easy and difficult at the same time: easy, because the trivialities are taken care of and you can start building today; and difficult, because the frameworks are not resolved yet and, if you want to push these technologies to the limit, you or your collaborators are going to reach the limit of the infrastructure, which is where it is time to build today.