Hearing that Microsoft Recall uses an AI model trained to observe the GUI, it all feels so… bizatine to me. Might be that my experience is coloured by previous attempts at UI automation testing but surely there’s a better way to do this at the API layer.
Well, I guess it includes a collective action problem to solve, too.
And it doesn’t guarantee all that sweet, sweet shareholder investment cash that seems to drive most decisions to turn to AI features recently (this “stock-price driven development” is one of the reasons why I try to avoid working at public companies).