@DM Your metasmoke account has been activated (i.e., been given the Flagger role). If you haven't already read it, there's an Introduction to Participating in Charcoal.
@Jesse A normalized version, {{'00201027039254'}}, of that pattern is already on the number watchlist. You can use !!/bisect-number 00201027039254 to determine which entries on the number lists match that pattern, which would tell you: Matched by 00201027039254 on line 3890 of watched_numbers.txt: 00201027039254 found verbatim
@Jesse That pattern looks like it's already caught by Potentially bad keyword in body and Potentially bad keyword in answer Append -force to the command word(s) if you really want to add the pattern you provided.
@Jesse A normalized version, {{'8884046710', '18884046710'}}, of that pattern is already on the number watchlist. You can use !!/bisect-number 1-888-404-6710 to determine which entries on the number lists match that pattern, which would tell you: Matched by 1-888-404-6710 on line 3891 of watched_numbers.txt: 1-888-404-6710 found verbatim; 8884046710 found normalized
@Jesse They don't overly do it (one answer), they disclose affiliation, and they provide an answer to the question. Bad answer? Maybe, but still an answer nonetheless
@Mast Please provide an adequate reason for rejection (at least 20 characters long) so the user can learn from their mistakes. Use -force to force the reject
@JeffSchaller That pattern looks like it's already caught by Potentially bad keyword in body and Potentially bad keyword in answer Append -force to the command word(s) if you really want to add the pattern you provided.
@Mast I forgot to check their SE profile (I only checked their SO profile). I'm gonna assume from that message that they advertised it on other sites, and was in fact spam
Beyond the 4 posted today, that user has posted that same website in 6 other answers. 2 are questions, the rest are answers. The ones posted on March 1st seem to be similar to the ones posted today (basically copy-pasted), although they do disclose affiliation.
However I am very tempted to spam flag them considering they were all posted in less than 10 minutes and are pretty much copy-pasted. Having a hard time believing that they're linking to their site in good faith because of that.
Also the links in the posts they posted today are broken, which is kind of strange.
I'm going to report the most recent 3 because of what I mentioned above, and the fact that they definitely saw (and reacted) to our comments and they still haven't added disclosure or removed the links from them.
@Fastnlight Or they think the rules don't apply to them because they have so much rep. There are a couple people I can think of off the top of my head that are like that.
@Jesse To avoid SmokeDetector reporting posts too slowly, you can report at most 5 posts at a time. This is to avoid SmokeDetector's chat messages getting rate-limited too much, which would slow down reports.
To each their own I guess, I just reported all of them because this guy seems to like to promote himself a lot. I'm not spam flagging any of the older ones, I opened a mod flag for posts containing a link to the website they spammed today, and might open a second one for everything else. Beyond promoting themself, they also seem to have a lot of answers that are just copy-pasted on multiple questions.
@user16217248 I recorded them all as spam, personally, because IMO the volume here crosses the threshold from 'reasonable disclosure of affiliation to a relevant tool/product' to 'a large portion of their posts advertise affiliated products' which is the fuzzy threshold on the help center page that gives guidance on how not to be a spammer (cc @Jesse)
In short, if most or all of your posts link to or use some product you own/sell/are affiliated with, disclosure doesn't save you
@Jesse for any that are copy-pasted, you can flag for mods and mention that it is a verbatim copy-paste--that's not allowed on SO and mods should delete any duplicates
@TylerH Yeah I might do that later. For the time being I just opened a mod flag for their answers linking their github and mentioned that they have a bunch that are copy-pasted. Might take an hour or so to go through all of their answers ;)
My problem is lots of genAI post flags that violate the 'no genAI' rule that have been around for 6+ months now getting recently declined en masse b/c mods for some reason decided not to enforce the rule as written. So on any given day now I am unable to flag things like spam or NAAs as declined flags pile up and then age away
@TylerH I think most of that is that there's only a handful of approved heuristics for moderating AI-generated content. There's currently only one SO staff member approving them, so it's going very slowly. There are a handful of heuristics that are generally agreed upon to be accurate, but are just stuck in the queue waiting for review and approval, and because they're not approved (yet), they can't be used by a moderator as a valid reason to delete a post.
No, there's plenty of approved heuristics. It's an issue between the policy saying "any use of genAI is not allowed" and the enforcement being limited to explicitly exclude "generated AI code and then modified it myself before posting"
It could easily be fixed by rephrasing the policy to not use the phrase "any use", which would align it with how it is being enforced, or by mods deleting content that freely admits to being genAI-generated
@TylerH There's only 10 approved heuristics (5 strong, 5 weak). In my experience, mods do always delete self-admitted AI-generated answers. There was a time during and slightly before the strike where AI flags were getting declined because of the policy that was in place. Could they have been from around that time?
and it includes questions and answers whose flag declination explanation includes a direct link to the meta post by a mod saying they don't think posts that use code originally generated by AI but human-modified should be deleted, so the cause is pretty clear
(and, for full disclosure, I am in the private heuristics team and in the ai domination chatroom on SO)
@SmokeDetector They do disclose their affiliation, but the fact that the link contains utm_source=stackoverflow kind of makes me want to spam flag it anyway. What do you guys think?
@Vickel The !!/blacklist command has been deprecated. Please use !!/blacklist-website, !!/blacklist-username, !!/blacklist-keyword, or perhaps !!/watch-keyword. Remember to escape dots in URLs using \.
Hmm, I could be wrong actually. I was under the impression that websites could be blacklisted just by being in a post with another blacklisted entry and having 0 FP, but looking at this page that might not be the case.
@Jesse maybe the chicken-farm site was hacked? I didn't click the link, but in google or ddg it shows a resume in Chinese characters which translates to a product catalog which has nothing to do with chicken (CHANEL 1990s Collection Catalog, etc. -. CHANEL 1990s)
Out of curiosity, is there anything in SD that normalizes characters with accents like in that post so it's still able to match things on the watch list?
