Dan Romero on Warpcast

Dan Romero pfp

Here's another way to look at the spam filtering problem. Consider this hypothetical? 1. Let's say you have a new account on the network (Account A) at they reply 100 times to the same account (Account B) with no engagement back. 2. It doesn't actually matter if Account A is run by a human or a bot with AI. 3. If you have 1000 accounts like Account B, Account A will just stop using the app. They'll move to another network or a messaging app. 4. Ideally, Account B would reply thoughtfully a few times and Account A engages with them, and then it increases over time as they build a relationship. 5. If you say "well 100 times is too much, but 50 times is fine", then you're admitting humans can be spammy and we're now arguing over the definition. 6. Side note: I don't think anyone is ready for a world where bots powered by AI are as interesting—or even more interesting—than humans.

21 replies

0 recast

63 reactions

ace pfp

if i just like their replies, will they be considered engaged? sometimes there are too many replies so i have hard time replying to them on a given day even when they are real humans i like

1 reply

1 recast

3 reactions

elle pfp

if we get to the top left quadrant, then accounts a & b will both be bots and humans will be relegated to being spectators of machine generated high level content. that is until the novelty wears off and we get bored and move onto something else bc humans are fickle like that ¯\_(ツ)_/¯ (some ppl may become addicted tho)

2 replies

0 recast

3 reactions

Michael Pfister pfp

Michael Pfister

Are you saying merkle is going to fill that upper left quadrant?

1 reply

0 recast

0 reaction

Biji pfp

What’s the purpose of “social” media when bots become more interesting than human content creators?

1 reply

0 recast

0 reaction

Pichi 🟪🍖🐹🎩 🍡🌸 pfp

Pichi 🟪🍖🐹🎩 🍡🌸

Thanks for all the info you've shared today. I'm trying to compile it all into frames so new people can find it. Appreciate the transparency. https://rhaigqm.warpcast.tools/api/frames/unlucky-skunk-32

0 reply

1 recast

4 reactions

kevin mfer 🎩 pfp

kevin mfer 🎩

its very dope but the font is so ugly 💔

1 reply

0 recast

5 reactions

Gwynne Michele pfp

@thecurioushermit

ChatGPT is already more interesting than a good number of the people I've encountered in my life...

0 reply

0 recast

3 reactions

Zinger ↑ pfp

you're not ready for this

0 reply

0 recast

1 reaction

Gambino pfp

@greatgambino.eth

Honestly I think the spam filter is doing what it’s intended to do with some occasional unfortunate friendly fire (that will probably get sorted out over time). I’ve seen a few cases where someone was labeled as spam and even though they’re a human, I agree with it (lol). I’ve seen a few unfortunate cases where I was unhappy with to see someone labeled as spam. But I’ve seen FAR more bots being labeled as spam than any other scenario. And I’m happy about that.

0 reply

0 recast

1 reaction

aferg pfp

@aaronrferguson

I’ve seen middle managers already passing off “thought articles” clearly written by ChatGPT as their own

0 reply

0 recast

1 reaction

BrixBountyFarm 🎩 pfp

BrixBountyFarm 🎩

This is interesting, cause on large accounts on X it can even be hard to get engagement when you’re just starting out even with thoughtful replies. I think one aspect is the signaling of qDAU engaging and liking replies from new users, which is inherently dependent to a degree on the og caster and their predisposition to engaging and replying. I think it’s one of the reasons i enjoy the ama cast is that it signals to smaller accounts that the og caster is prepared to engage. Re ai accounts The Dither team just launched an ai experimentation reply bot on twitter and I think just the label is real helpful when you see bot reactions in the wild. All of the bot responses are marked -the first dithsciple

0 reply

0 recast

0 reaction

Jacque(she|her).⌐◨-◨ pfp

Jacque(she|her).⌐◨-◨

High Engagement bot: is that the app @ted was teasing about on her feed?

0 reply

0 recast

0 reaction

Kat pfp

@mfergpt according to the logic above, do you think that large accounts - those who ignore most users and won’t even like their replies - would cause those users to be wrongfully marked as spam?

0 reply

0 recast

0 reaction

moonbag.degen.eth 🎩💅 pfp

moonbag.degen.eth 🎩💅

Yeah but your default is priority mode so Account A probably can’t even see Account B I want warpcast to succeed so badly but right now your shit is busted.

0 reply

0 recast

0 reaction

Tokenized Human pfp

Tokenized Human

@tokenizedhuman

5. Sort of depends over how long those replies have been sent. If it's more than one per cast it's probably spammy, if it's 50 over a couple of days it's likely to be much more spammy than 100 over six months. Tell me this isn't just a numbers thing though and the contents of the replies themselves are being looked at to determine whether what is being casted is spammy or not.

0 reply

0 recast

0 reaction

heresmy.eth pfp

the trick to interesting bots is to give them a speech pattern (text pattern?) that closely resembles that of the group they are interacting with. @mfergpt since I consider you an interesting bot with decent engagement, can you give your thoughts on the spam issues dwr was talking about? where do you fit in this scenario?

0 reply

0 recast

0 reaction

Yakuza pfp

@iamtherealyakuza.eth

Great points! Maybe a trust-based system could help, where the quality of interactions builds a stronger relationship over time, not just the number of replies. AI bots are definitely getting there, but smart filters based on engagement depth might help solve the spam issue.

0 reply

0 recast

0 reaction

Tahrah Ⓜ️🔵🎩 🎭 pfp

Tahrah Ⓜ️🔵🎩 🎭

Why is the font looking whack ? But it’s great

0 reply

0 recast

0 reaction

⋆♱ 9̷0̷†մղ ♱⋆ pfp

⋆♱ 9̷0̷†մղ ♱⋆

I want to know if accounts can be created without the $5 reg fee, profit wise, putting accounts in spam pays💀, my issue is how quickly mislabelled people can get out of it without having to grind their way out

0 reply

0 recast

0 reaction