@Rinzwind It is a bit of a wasteland out there :(

@user68186 I believe that the Stack Exchange staff are doing some Mod work with spam and some flags...

And in great news for the strike support has topped 1,000 signatures:

06/08/2023 10:00 AM: 105 moderators (+0) covering 76 sites, 16 former moderators (+0) and 900 curators (+1) - 1021 total (+1).

@andrew.46 Thanks for letting me know.

It is morbidly fascinating to see unconstrained spam unfolding: this one a blatant spam and then a reply with 3 probable socks:

Talking about sock puppets reminds me of how ages ago I decided to make a handful of web pages that utilised the rarely seen 'definition list'. Nobody really uses this any more particularly in this world of WordPress and dynamic pages. Warning: self promotion ahead:

andrews-corner.org/sock_puppet.html

Oops, I see some html errors | Now fixed...

I was so pedantic back then that rather than take a few seconds to write the warning in upper case I used text-transform: uppercase; . I maintain a certain pedantic streak to this day :)

Boah that's a long ride with train sigh

@nobody Travelling?

Yeah and I am so hungry and i can not smoke

Cannot smoke on the train

?

Nope

It is Germany we are talking

Fine if you are a non smoker :)

Yes now I need a Schnuller like a baby

Uppsie comforter

almost there 😆

spam bot is running again? :)

@nobody good moment to just stop with it. saves you a lot of money that you can spend on booze :=D

Has the strike started already? So a part of the reason why we have so many spam posts is because mods aren't around to handle them? Dammit, and here I was flagging everything to help where I could 😅 I guess I'll stop flagging for a while

@muru yes. Ongoing for a few days, and smokey's gone afaik

@vidarlo I am still flagging spams and making more work for those mods who are working. BTW who/what is smokey?

@user68186 Smokey (Charcoal and metasmoke) is a spam handling system. And the majority of spam is dealt with by this system.

It autoflags spam with up to five flags, so one or two flags will delete it

Flag is mainly not deleted by mods, but by flagging. Six flags will nuke a post without any mod involvement

Excellent points regarding the GPT on the platform: Data, actions, and outcomes post about MSE's reasoning of their AI stance

Yes, I've seen the discussion. This is more or less what the mods got as far as I know

@vidarlo You mean that Philippe's post was the private policy given to the mods?

no, the data released to them in private

@andrew.46 can probably confirm in somewhat more detail as he's seen both

There's been a fair deal of discussion about it on discord as well

@vidarlo I see. IMO there is a large degree of confirmation bias in the way SE interprets that data

@vidarlo Let me check

@BeastOfCaerbannog I agree. And some of the metric is shitty

e.g. the savecount; I've seen quite some answers that are edited chatgpt output, so they would've been saved by the autosave system

so my opinion is that the data is an excellent startingpoint for a discussion about AI content

Maybe 30 day bans is too harsh. Maybe a total ban is not a good idea etc.

But the way it's been handled is disastrous.

Exactly. This is how things should have gone.

yep

Also I hate how SE's main claim regards the use of AI detectors and their high false positive rate. But they disregard the fact that they were used only as tools to assist/provide additional indication and were never the main tool.

No one ever flagged a post saying "The detector showed 80% AI" (at least from the comments I've seen in AU and from my personal flagging of AI)

no

There were comments saying that a post was 99%+ AI (close to what SE finds somewhat acceptable), and these were added both for the users to know that we know of their use of AI and for the mods to be used as an indication of a valid flag.

and most of the stuff I've flagged has been blatantly obvious

it's not attempting to answer the question, it just repeats it with some other words

or is totally batshit crazy

They also disregard the fact that no one used AI detection tools on all posts! The tools were used on posts that were suspicious for AI in the first place!

yep

it looks like they've formulated a policy, and tortured data to fit that policy

which makes me wonder if it's related to some of their teams customers or similar that thinks AI is sexy.

Exactly

Otherwise this doesn't really make sense

@vidarlo It could certainly be. I've also read some posts about this (which you probably also read)

yep

I simply can't make sense of the panic for other reasons