Hi, welcome to AI Zone. This is a space where I collect my thoughts on building software with AI agents, the changing role of developers, and the practical realities of shipping AI-assisted products.
Token economics How is it possible that the shady token economics are accepted. We have no clue what is happening when a request is fulfilled, did we get our moneys worth? How can we even verify if the billing accurate / let alone reliable? The bigger threat, once we run...
Read more →Haven’t we seen this before? Significant risks have been identified in cloud computing, the solutions are not even in place and we are all diving full force into AI. The risks of AI and depending on a few providers has so much similarities with the Cloud; We trust our data,...
Read more →Is a bad idea not the core problem of a process, what if you’re already doing something wrong, will AI fix it? What if you have all the money in the world, a top leading development team to built the perfect application, but the core idea itself is fundamentally flawed?...
Read more →Is work faster, or do we have more work to manage? The first week after I drove home from our holiday road-trip, I opened my laptop and thought, let’s give this vendor one more try. Let test it and try once more if I can improve my ‘wip’ game I...
Read more →The guardrail layer If agents write the code, the useful work is no longer typing. It is deciding what they are allowed to do. Before you let an intern-with-infinite-energy touch production, you need a layer that does not depend on a human catching every mistake in the moment: architecture rules:...
Read more →Budgets are the new estimates There is currently no cost estimation tool for using AI at scale / agents. However, departments are already burning tokens, validating if this spend is justified is nearly impossible. The more the path of agents is adopted, and proprietary models are used, the harder it...
Read more →Static analysis as the reviewer You will not read every line an agent produces. Pretending otherwise is how half-baked software ships. Code analysis is the part of review that can run every time, on every diff, without getting tired: linters and formatters so style is not a debate complexity and...
Read more →
