I get the ad-like vibe, but I think that's slightly unfair to the paper. "General purpose technology" is being used as an economics term here, not as a claim that ChatGPT is universally useful