Spent $47 on a GPT-4 API experiment to auto-summarize my inbox, now I actually read my emails
It saved me about 2 hours a week digging through newsletters and spam, but it still can't tell a real invoice from a phishing attempt, has anyone found a filter that actually works for that?
Your $47 experiment actually sounds like a decent deal if it saves you two hours a week. But how are you handling the false positives from the invoice filter, the ones that are real but still get flagged as spam? That part seems like it would drive me crazy more than the phishing ones.