Hacker News Reader

Comment by yusee | Hacker News Reader

GMail still can't effectively distinguish between real communication and spam. If China overtakes the US in AI, Schmidt and co. are the folks to blame. They are in the drivers seat of intelligence development, yet their efforts are increasingly scattered, and their only consistent focus is on lobbying government.

Dear Alphabet,

Please make your products and services work before worrying about geopolitical balance of power.

View on HN

(While I work on Gmail, this is a curiosity)

What issues do you have with spam? Personally, I get maybe 1 false-positive spam email per year with Gmail, and maybe a few false-negatives (which are less of an issue anyway), for an accuracy rate in the 3-4 9s range. So I'm intrigued as to what you're doing or where you encounter the inability for gmail to distinguish between real communication and spam.

How about: I send someone an email, he answers, his answer is classified as spam. That's as broken as it can get!

(presumably because google doesn't like 'free' servers)

Maybe your friend's email server is misconfigured? Incorrect SPF records, etc?

I think the point was that initiating communication with an address is an extremely strong signal of willingness to receive (vs clicking a link, or replying, etc.). I think it's reasonable to expect to receive responses to proactively-initiated threads even if your correspondent isn't optimized for deliverability.

I disagree, just because I’ve emailed support@visa.com in the past does not mean I want every spam phishing email pretending to come from support@visa.com reaching my inbox.

You make a fair point but you've drawn a slightly broader scenario than I had in mind. Surely if you initiate an email to myfriend@obscureserver.com with the title "Hey buddy" and you get back a reply titled "Re: Hey buddy" from someone alleging to be myfriend@obscureserver.com, you'd want that in your inbox and not spam even with a misconfigured sender on your friend's end... no?

Edit: my ideal UX in this situation would be to get the mail in inbox, with a small notice saying "Unverified" and a mouseover/hover text explaining what that means re: SPF records; from there if you mark it as spam it would treat such unverified mail from that domain as spam on an ongoing basis

yes! I wish I could upvote more.

I implore you to solicit users who will volunteer their spam folders so that you can improve your algorithm. It has always been badly broken for me and many if not most people I know (worst story my lawyer friend who finds important email from judges in this spam folder). Right now my spam folder is 171 emails that are 80% legitimate. Most of them are messages from political groups and businesses that I don't care about, but about once every other month I curse when I find something that does matter. I move some "spam" emails to my Inbox but I do not have time to train your system by correcting all the mistakes. It is so bad that I would prefer not to have the spam removed and see all my email with some kind of "might be spam" flagging to aid my speedreading of the daily torrent.

Lately I've been getting a ton of spam in the form of Amazon phishing attacks. Usually these emails use character substitution or the like to avoid being caught.

I've also noticed recently that spam hits my inbox before being filtered, which is aggravating as that means I get a notification about it.

I get spammed by cybercoders even though I click unsubscribe and report spam all the time.

The other issue I have is that the search function doesn't update... I try to search for an email from the day before and it won't show up.

Maybe Gmail was a has example. How about the Google play store?

I interact with the play store relatively infrequently, so I don't know what you mean. Would you mind elaborating?

This. What is worrying me is core services seem to perform worst than before like GMail, but even search or YouTube.

Yes! Another thread a while back had a good discussion on this. Gmail, google search and YouTube search have all been terrible last few years. Ever since they started touting ML, I think they started to cut costs and it's been terrible using all 3 services whereas they used to be amazing a few years before.

Reply on news.ycombinator.com