This makes sense, because A.I. as it has come to be defined has an insatiable appetite for new data. All means fair or foul are being commandeered to feed it. Think of this next time you participate in a survey!
The last stand of A.I., which nowadays is limited to neural networks, needs an immense amount of data to "teach" its language models. And since each new release of an app requires an exponential amount more than baseline, A.I. is reaching a ceiling: incremental improvements will get more and more expense to make, and deliver less and less.
Sounds like a great way to amplify bias or compound problems built on an erroneous data base
This makes sense, because A.I. as it has come to be defined has an insatiable appetite for new data. All means fair or foul are being commandeered to feed it. Think of this next time you participate in a survey!
The last stand of A.I., which nowadays is limited to neural networks, needs an immense amount of data to "teach" its language models. And since each new release of an app requires an exponential amount more than baseline, A.I. is reaching a ceiling: incremental improvements will get more and more expense to make, and deliver less and less.