Are there any plans to make the scam filter intelligent?
Your scam filter identifies roughly 50% of legitimate IT (STEM, general IT, others) newsletter emails coming from noreply@1105newsletters.com as scams. If it can't learn to identify legitimate senders and adapt, the filter is virtually worthless and a total annoyance. The only solution is to "disable it (at your own risk)." which is equally useless assistance.
Wšykne wótegrona (1)
There is some discussion on the developer mailing list yesterday. The consensus is at this point more real time information on the level of false positives is needed. Then to either disable it or fix it so email with tracking bugs are not included in the filter. The reality is that while perhaps mislabeled mail identified do contain tracking information used by the sender to get all sorts of click through and profiling telemetry. They might not be a scam, but they are a risk to your provacy.
No decision has been made.