Umami AF on Nostr: Re: last boost, I sometimes discuss AI with an MIT professor, and although he's at ...
Re: last boost, I sometimes discuss AI with an MIT professor, and although he's at the forefront of the technology, he's also deeply skeptical of it.
The training data is the issue I've been harping on, in a quasi-professional context. It's hard to train something when that was never one's intent. Why would you train something on imperfect, real-world data that includes things like bad days, unsatisfactory billing rates, rush jobs, low-quality translations, manipulative behavior, etc.?
The training data is the issue I've been harping on, in a quasi-professional context. It's hard to train something when that was never one's intent. Why would you train something on imperfect, real-world data that includes things like bad days, unsatisfactory billing rates, rush jobs, low-quality translations, manipulative behavior, etc.?