@itnotes Yeah, strange times we are living in… and this is only the beginning, I fear. Now, how can we educate people to filter out this AI bullshit ? I’m quite sure my client will soon want to feed an Ai with my codebase to audit code and get brilliant ideas💡 And then, will occur the same fight “expertise vs the-obvious-ai-truth” : will client trust me or will it trust the vaporwareAI “very clever ideas”. To get a feeling, I began playing with Claude Code (in a local environment with ollama and a not-stupid-model). Well, while the thing does great to write code, it is not great to play with a large codebase and reverse engineer the design choices on a 20yo architecture with a lot of subsequent changes. I suffered to get the knowledge, but it’s fairly clear in my head. I wouldn’t rely on the llm to properly act on this code. so, my job is safe, for now.