Dan Romero pfp
Dan Romero
@dwr.eth
This is not true. If you cast outside of a channel, it just goes to your followers. Casting in a channel gives it the potential to be seen by more people.
18 replies
4 recasts
103 reactions

Tom Jeans pfp
Tom Jeans
@thomasjeans
you’re missing the point I have the same feeling as @zkp casting with no channel feels like kind of a waste unless you’re an auto follow account when users are coming up with hacks like creating their own version of /no-channel there might be an opportunity
1 reply
0 recast
1 reaction

Dan Romero pfp
Dan Romero
@dwr.eth
How do you determine if a cast is from a good user without a power badge vs. a spammy user without a power badge?
3 replies
0 recast
2 reactions

raulonastool.eth 🎩 🏰 pfp
raulonastool.eth 🎩 🏰
@raulonastool
it's subjective, but quality of content is first thing I look for. Do they show some semblance of being human. One thing I'm cautious of is if I feel they appear spammy because they are non-native English speakers. In that case I look at their cast history. If it appears all of their casts are value extractive (fishing for like, follows, tips, airdrops, etc.), then I write them off as spammy. If casts are more sparse, and generally closer to sharing thoughts, observations and opinions, then I know they're human. There's a bit of a Turing test element to it.
1 reply
0 recast
0 reaction

Dan Romero pfp
Dan Romero
@dwr.eth
Sure. But how do you do that algorithmically? What happens when bots use LLMs to appear human but the content is rubbish?
1 reply
0 recast
0 reaction

raulonastool.eth 🎩 🏰 pfp
raulonastool.eth 🎩 🏰
@raulonastool
that's more of a research question. I'm sure there are teams experimenting with DIDs and proof-of-humanity for crypto applications. I think you're right that LLMs make this harder, but I also think there are patterns in the way an LLMs write that can used. How do AI Content Detectors work for school work?
1 reply
0 recast
0 reaction