Claude Code, TUCANO, and a New Way of Building Companies

There are those rare moments when you leave a room with the feeling that something fundamental is shifting. Not because a new model has been introduced. Not because someone has presented an impressive demo. But because many individual observations you have made over the past months suddenly begin to form a coherent pattern. That was exactly the feeling I had after attending Anthropic’s first Claude Founder House Event in Berlin.

What surprised me was the fact that the most interesting conversations were not centered around technology. Of course, people talked about models, agents, product development, and the speed at which possibilities are currently evolving. But the longer the conversations lasted, the clearer it became that the real discussion was taking place somewhere else entirely. Not on the level of technology, but on the level of decisions, priorities, and entrepreneurial judgment. One statement kept appearing throughout the day in different forms:

The challenge today is no longer building things. The challenge is building the right things.

Just a few years ago, product development was primarily a question of resources. Those who had capital could build teams. Those who had teams could build products. Those who could build products had the opportunity to enter and shape markets. That logic is now changing fundamentally.

With tools like Claude Code, we are experiencing for the first time that small teams can accomplish things that previously required entire engineering departments. Ideas can be tested within hours. Prototypes can emerge within days. Workflows that would previously have required several specialists can now be developed by small, focused teams. The barriers to product development are decreasing. Speed is increasing. And with that, the actual bottleneck is shifting as well. Another statement that has stayed with me ever since was:

"Pour everything into the part only you can do."

The more I think about it, the more I believe that this sentence describes the current development better than any technical discussion about models or benchmarks ever could. Because as technology becomes capable of taking over more and more tasks, the question of which tasks can only be performed by humans becomes increasingly important. The ability to recognize connections, to frame problems correctly, and to distinguish relevant opportunities from irrelevant ones.

This is exactly the point at which the story of TUCANO began for the three of us.

Looking at Rethinking Places from the outside, one might assume that we implement communication projects, help companies navigate AI, or support executives in building thought leadership. All of that is true. At the same time, however, we have observed another pattern over the past years. The more organizations engage with AI, Customer Value Communication, and digital transformation, the greater their need for guidance becomes. People are not searching for more information. They are looking for interpretation. They are looking for context. They are looking for answers to specific challenges.

With every client project, every workshop, and every strategic conversation, this body of knowledge grows. At the same time, a problem emerges that many growing companies are familiar with. The demand for expertise grows faster than the available time of the people who have built that expertise. At some point, quality does not become the bottleneck. Availability does.

That is precisely why we are building TUCANO. Not as another AI tool or chatbot. Not as a technical experiment. But as an answer to a question that has occupied us for a long time: How can more people benefit from knowledge, experience, and best practices without every interaction necessarily requiring one-to-one support? How can we make what we have learned accessible to marketers, CMOs, growth leaders, and brand decision-makers who do not possess deep AI expertise and who often lack the time required to acquire that expertise themselves, allowing them to benefit from what we know even if we will never work with them personally?

For us, this is not about replacing people. Quite the opposite. The most interesting AI applications we see today do not replace expertise. They make expertise scalable. They help make knowledge accessible, connect ideas more quickly, and support people precisely when they need guidance. The actual value creation remains deeply human. No model understands a client's specific challenges. No model understands the dynamics of a market. And a model certainly does not develop entrepreneurial intuition.

What AI can do, however, is remarkable. It can help structure knowledge. It can help make patterns visible. It can help make answers more accessible and available more quickly. That is exactly why we do not see agents as substitutes for people, but as amplifiers of human capability.

What I found particularly interesting was how openly Anthropic shared how Claude is used internally. The most valuable insights were not about technical capabilities. They were about impact. Again and again, the same question surfaced:

"How do you translate raw model capability into tangible business outcomes?"

In my view, this question captures the essence of the current development. The ability to technically implement something is gradually losing its status as a competitive advantage. Today, almost any idea can be prototyped. Almost any process can be automated. Almost any problem can be addressed technically. The real difference emerges where people understand which problems should be solved in the first place and what concrete value a solution creates for customers. Perhaps that is also why another lesson from the day continues to stay with me:

"Don't be afraid to kill it. Re-rolling beats repairing."

This sentence describes a mindset that may be essential for the current phase of product development. Many companies were built in a world where development was expensive. If something was built, it had to be preserved. If time was invested, that investment could not be lost. But when development costs decrease and iteration becomes possible almost in real time, the way great products are created changes as well.

We experience this every day in the ongoing development of TUCANO. Hypotheses are tested. Agents for studio production workflows are developed. New workflows emerge. Some work immediately. Others disappear again. Features are discarded, reimagined, or rebuilt entirely. Not because the original idea was bad. But because learning has become faster than ever before.

This may be the greatest transformation of our time. Product development is becoming democratized. Execution is becoming less expensive. Speed is increasing. And as a result, the focus shifts from the question: "Can we build this?" to the far more important question:

"Should we build this at all?"

The most important resource for modern founders is therefore no longer primarily capital. Nor is it technology. The scarcest resource is judgment and the ability to recognize which problems are truly relevant. The ability to distinguish opportunities from distractions. And the ability to identify the work that only humans can do.

We often talk about how AI is changing work. In reality, AI is changing something far more fundamental. It is shifting the boundary between execution and decision-making. Execution becomes cheaper. Decisions become more valuable. And that is exactly why the sentence "Pour everything into the part only you can do" takes on an entirely new meaning.

Because perhaps the most important task of founders in the future will no longer be building products. But identifying the few problems that are truly worth solving.

If we succeed in doing that, something emerges that drives us every day at TUCANO. Knowledge becomes more accessible. Expertise becomes scalable. People receive support they otherwise would never have had. Good solutions reach more people. And the knowledge of a few can suddenly help many.

Good things rise.