It's pretty clear that issues with token leaks have been ongoing for a while now. However, it's pretty clear at this point that the more that these strings are spammed in chat, the harder it'll be to resolve the issue. Since people won't stop bringing it up, I suspect it's much harder to distinguish between
(a) one of the twins leaking a token because of a code issue (true positive) and,
(b) them mentioning a string because the chat feed shoved it into their context (false positive).
While I'm sure there might be some respectful enough to cut it out if asked to, I think it might just be easier to automatically delete messages with known token strings. This might not even have to be a permanent thing if the issue can get resolved to satisfaction, but it's clear that leaving the status quo is going to actively sabotage efforts to fix this bug.
#Add mentions of token leaks to Twitch Chat automod filters
1 messages · Page 1 of 1 (latest)
unrelated but is there a way to fix the twins thinking cerber is a man 😭 they say "he" all the time when talking about her
make your own thread for that
seems pointless to make a whole thread for something like that when this is also something needing to be fixed
they've been doing it for years to multiple vtubers and none of them care atp
so I don't think it really matters
they probably dont care no but its still nice if its fixed
anyway to the point of the thread, it should be pretty easy to add ther leaked tokens to the automod since they're very specific phrases and it wouldn't mess up chatting
at least until the end of the stream
then they can be removed
Exactly, it wouldn't have an impact on chatters who aren't trying to mention it
It'd probably be better to filter Neuro and Evil from seeing such posts in the chat. Either completely or mostly, as they could get a summary from time to time instead, to stay aware of what's going on.
Maybe, but filtering it on the end of twitch chat wouldn't require Vedal's intervention and could be handled by a mod instead
yeah that's way harder
and is never done basically
usually words are just automodded so the twins can't be baited into speaking about them
either because of some current drama or in this case, it baits too much spam
I remember the word beans got automodded because it caused a feedback loop
Ah, good to know that there's some precedent for this. Thanks, BB!
yeah, plenty
has happened like every subathon I think lol
some word gets excessively used, the twins read it, talk about it, causing chat to spam and then it repeats until the mods automod the word
only after enough people complaint, and that after what 2-3h it finally got automodded
I mean yeah it takes awhile but it does happen
I know that people don't wanna be a buzzkill with stuff. But when you get down to it participation is a privilege, not a right.
In NN vedal said the token leak may be fixed in rewrite so 
well he also said he fixed it on tueday so 
code is finicky I guess
this is a good stopgap though
should stop chat spamming it constantly and prevent false positives since vedal did say in chat spamming it makes it worse
Yeah, even if it's just to confirm that it's fixed it still wouldn't hurt to control variables
bro that's so unrelated that it really should be it's own thread

THEY DID IT
seems to be on and off though?
Never mind they're still doing it 
it's weird cuz it stops and then starts again?
are they just changing it slightly, perhaps? 

like there's a bout of spam and then it stops
idk if this is just chat moving on really fast or what
Could be that the mods only put 1 variant of the term into the automod
possibly
if they filter something like <|any text|> it should get it

