The "agentic" buzzword and the promise of having all these quasi-autonomous processes running around handling your email...

The "agentic" buzzword and the promise of having all these quasi-autonomous processes running around handling your email or generating slopcode or whatever, gestures towards the ideal of robot sidekicks. That's partly how the #AI boys are trying to sell their #LLM rubbish, anyway, as if it's like having a whole team of autonomous assistants—except they're NOT autonomous at all because all these "independent agents" are wired into the same centralized "generative AI" service, and hoovering up all the users' data in the process and piping it back to the central service. The behavior of the "agents" is totally under the control of OpenAI and Anthropic and so forth, and indeed that's another selling point: the "generative AI" people keep tweaking their product and offering different "personalities" and so on, which would be impossible if the "artificially intelligent" agents actually possessed an independent sense of intelligence.Instead, it's the central service that's supposedly evolving...