One last question about the Amazon discussion earlier...should I be looking at EC2 and how to set that up? Or AWS? Lambda doesn't look like it does what I want. I did find "beanstalk" mentioned, but haven't looked at that yet
It makes sense to have one brand for services in general, and subdivide into types. I only know EC2 (computing) and S3 (storage), the rest is some mumbo-jumbo to me.
I'm pretty sure when people say "AWS" without further detail, they probably refer to EC2.
[ SmokeDetector ] Bad keyword in body, bad keyword in title, blacklisted website in body, link at end of body, pattern-matching website in body: Lumella Eyes Serum by brittneywaddell145 on drupal.stackexchange.com…
yeh it's always around 3.30pm here which is around 11 am India, for some reason I suspect some of the spam comes from there, though I'm not sure @zaq do we have stats on where the spam comes from?
Spammers from India mostly do Manual spam, which is not very well detected by the SE Spam filter. In other countries, they mostly use bots which is pretty well detected. Don't have evidence on that, but that is what I suspect
@Magisch yes but we're having a discussion about it and spam o clock in particular intrigues me and I suspect that's cos it's around 11 am Indian time and the spammers are at their pcs ;)
@ƎpᴉʇʎXD that's also for auto spam, but that the spam post is not posted on the network as it is stopped by the SE filter, so account nuke is delayed. But, the auto-spammers account also gets blocked
[ SmokeDetector ] Bad keyword in body, link at end of body, link following arrow in body, pattern-matching website in body: Booty Pop When considering by PhyllisDumond on gaming.stackexchange.com (@angussidney) …
Failed attempt to spam? http://stackoverflow.com/a/38929264/1364007 (click edit to see what they were trying to do - there's a link to a site in the source).
@Magisch Maybe? Remember, it's a lot of work to train new moderators - even non-green ones are coming from sites that are a couple orders of magnitude smaller than SO, at least.
At this point, throwing more moderators at the problem is starting to not scale well.
I swear the flag queue is sentient. I can go on vacation for two weeks and not handle anything, then show up and handle two hundred in a day... and somehow there are two hundred more flags to handle that day than any other day.
but most of these surely are fairly mundane and there are established ways of dealing with them
in the same way that a couple hundred reviews in you get a feel for how to handle most cases in the review queues without thinking more then a couple seconds
Sure, but it's mentally exhausting. Five 'other' flags could have five different problems. It's better when you can store one procedure and do it over and over and over.
don't get me wrong, its a heavy and important job, and you guys are doing amazing work, but it doesn't sound like wizardry to me, just tedious on an otherwise unprecdented scale
Context switching is terrible. I can sit down and do ten sockpuppet cases in a timeframe, or ten plagiarism cases, but I can't do five of each in the same timeframe.
But yeah, I have thoughts about how we could harness other folks (cough @ArtOfCode and co.) for the cases where we need trustworthyness, but not really election-level PII access and such.
I'm sure SE staff will bring forth something sensible
one of the reasons I spend so much time on this site is that I'm convinced that the staff at least mostly know what they're doing - which is something thats refreshingly new about a site
