@stefano this is a synptom of a much larger problem - i had an argument with some people recently who want to use "AI for 6G". I asked "so what's your platform to run the AI on? like do you have server racks in cell towers and run the AI on unikernels, or re-implement the AI on top of Cloud RAN and tolerate the latency, or do you have a non-GenAI tech in mind for your smart cellular service?". you could hear the tumbleweed rolling through their empty heads...stack free