Content
@
0 reply
20 recasts
20 reactions
Dan Romero
@dwr.eth
PSA — please only report casts with a bot replies before "Show more replies" (i.e. "above the fold"). If the replies are already in "Show more", it means we've labeled the account as spammy.
19 replies
15 recasts
74 reactions
downshift - μ/acc
@downshift.eth
good to know! question for you, Dan...for 'useful' bots, is your team planning to allow those to operate freely on Warpcast? (bots like @degentipbot.eth and other automated info sources)
1 reply
0 recast
2 reactions
Dan Romero
@dwr.eth
They already do. And "bots" = spammy accounts operated by humans that use ChatGPT to generate answers. Very few accounts are "programmatic".
2 replies
0 recast
1 reaction
Mark 🎩
@web3withmark
How do you detect the use of ChatGPT? If it’s purely based on text analysis many people (everyone in my Japanese company) will more than definitely be using ChatGPT to translate. I wonder is some sort of “vouch for account” or “organisations” feature could be implemented to help those - but then I don’t know how you’d stop people gaming that 🤔
1 reply
0 recast
0 reaction
FarcasterMarketing
@quillingqualia.eth
They trained a custom model, I don't see a lot of false positives which means the model is good enough for now. I don't think using LLMs to augment your answers is an issue, if it's done right(meaning > 50% of your content isn't AI ), it wouldn't get marked as spam.
1 reply
0 recast
0 reaction
Mikado
@mikadoe.eth
Does that differentiate between users that give a cast to AI asking for a reply and users who give their own reply to AI asking it to fix typos, grammatical errors and refine the text in general? 🥲
0 reply
0 recast
0 reaction
