So, some of you met me and my brother after this wedenesday´s Afternoon Tea (for me was a night tea, but anyway) meeting. Here´s some context: - I insisted he used this methodology - He´s been using AI and introduced me in the last year to some cool AI tools (like Notebook LM, for example) - He´s got over 40 years of computer usage experience (not programming, but usage) Then he sent me over whatsapp (in spanish) the following line: "I'm here building an MVP to audit Google campaigns. So far, so good!! ...I just need to finish watching the videos so it uses fewer tokens." Which I then replied: "It’s NOT just fewer tokens. (ICM) It’s: - fewer or almost no hallucination events - fewer modifications and rework of results - fewer iterations to reach the MVP - more control over changes - easier scalability - faster results - better interaction between the output of one process and the input of another - easier to automate - easier to debug - much, much easier to understand when you pick up the project in 2 weeks, 3 months, or 1 year - much more valuable if you’re going to hand the project off to someone else " After some projects, you realize this yourself. I didn´t go to SKOOL to check the answers, or to make my point....this all came from MY experience after implementating this Methodology for almost 2 months... Every project gets better, faster cheaper (and everything else I answered above). After wednesday´s Afternoon Tea meeting with @Jake Van Clief , I chatted a problem that I encountered after using astro.build , and Jake actually took the time to review it, answer me, and then ask me about his answer ... 🤯 I spent all saturday thinking about the implementation, writing an instruction.md file, and actually reorganized my astro.build project, so that it has an ICM structure (this will be another post). Have you got also this Mind blowing 🤯 moments using ICM⁉️