Or do they just take what marketing people say about the tools at face value? Because they seem genuinely surprised that people don’t like the tools, when even the most ardent AI enthusiasts I’ve seen that use the tools are well aware of their limitations.
it’s a godsend for tagging semantics, but that’s a hyperspecific context and I’m sure there is something out there more suitable for this that I just don’t know about. I’ll keep using my 3b model to tag two million messages until then.