Sharing how we think about spam so people can better understand.
1. Most people will only use social networks if they are fun
2. Casting something and getting dozens of spammy replies is not fun
3. If you get enough spammy replies, you'll just stop using the app
4. So, aggressively filtering for spam is necessary to keep your existing users happy
5. However, this means some good new users will hit spam filters. Understatement: this is a bad user experience for those people. It would be great to not have this happen.
6. But given the choice of who to make unhappy — an existing user who has invested a ton of time and energy into the network or a new user that is most likely to churn (top tier social networks lose 50% of users!), then you have to optimize for your existing users.
7. So does that mean you don’t care about new users? No! It’s existential to have new users be able to join Farcaster and find people to connect with — without being labeled as spam.
8. So here’s what we are doing to improve this so new good users aren’t mislabeled as spam:
a. Continually refine the spam model by identifying incorrectly labeled people (that feeds back into the model and makes it better); good existing users on the network reporting incorrect labels is really valuable here
b. Warnings to users before they are labeled spam so they have a chance to adjust their behavior
c. Increase the number of signals our model uses to help filter out bots vs. people
d. Working a new version of channels that give owners / moderators complete control of who appears in their channel (so a Warpcast spam label won’t affect casts in the channel)
Here's another way to look at the spam filtering problem.
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.
Or "Just use proof of humanity to solve spam."
1. Spam is relative: you might not find something spammy but another person might find it spammy.
2. If you are casting in your home feed or to people who chose to follow you, then it's by definition not spammy
3. If you are casting at someone else, then their opinion does matter. It's no different than walking up to someone on the street and trying to strike up a conversation with them.
4. Fundamentally, spam is unwanted inbound that tries to get someone's attention (literally what spam email is).
5. Humans can be spammy and bots can be not spammy.
6. What proof of humanity does is limits how fast new accounts can get created, i.e. if you want to create 1000 accounts, you now have to get 1000 proof of humanity credentials. That's friction.
7. However, if you have proof of humanity, you can still be spammy (and still hook your account up to ChatGPT, etc.).
Over 700 subscribers
Great analysis, it is much appreciated !
Here's what we discussed at Dev Day last week. Our priorities, i.e. where the 14 people on our team are spending time: - Channels - Frames - Retention (feed, spam, DCs, bugs, onboarding) - Hubs I'll do a follow up video with additional context this week. If you have questions on any of these topics, reply to this cast and I'll address in the video. https://warpcast.notion.site/Farcaster-Priorities-1116a6c0c101806fbab8ecfea5c0fa40
Big
Are the 100 users that got a boost changing continuously or are they just locked for now?
It was never 100 users? The group rotates daily.
Wasn't that what V said? Gotcha! 🫡
re: frames, something popped up - I believe it's good to make a standard for auto-completion in frames inputText https://warpcast.com/sinaver.eth/0x41b1f5f3
Thanks, will add to the list. Probably lower priority for now.
Please can you also enable autocomplete for the search bar when searching for users or channels(in particular)? :)
lets goooooo! thanks dan for pushin thru the noise and bein social 🫡
What about growth?
Great list. Do we have a date for migration to channels v2?
https://warpcast.com/dwr.eth/0x375d1ab1
🔥
You mentioned once that most people weren’t casting in channels, can you show us that data? Because I always thought channels were the most active feature of Warpcast
Last time you said most people weren’t casting in channels, can you share the data? Cause I thought most people don’t cast in feed cause of less engagement.
Where / when did I say that?
Your casts are a lot so can’t really scroll, but I remember vividly. You should too, something like most people weren’t casting in channels.
https://warpcast.com/dwr.eth/0x33155ea2 here’s a snippet I could find though @dwr.eth
This is observing that for a given channel, most followers of the channel don't cast. Whereas if you look at most casts, they are in channels.
I read that signup costs will be 1$ early next year. How does this impact storage prices?
We will likely reduce the size of storage slots accordingly. So they will roughly stay the same.
any chance payment will support onchain transaction?
Why is the amount of storage less than before, why this change?
1$ signups? HUGE af
smooth video player when?
I just subscribed to @dwr.eth on /paragraph! Check it out:
The spam filtering dilemma (Put the 3 long casts into a single spot for easier referencing.) https://dwr.email/the-spam-filtering-dilemma
for part 1, point 8a, "good existing users on the network reporting incorrect labels is really valuable here" -- how do you do this? I know how to report users as spam but not how to flag spam filtered users as not spam. If someone is marked spam and I like or reply to them does it count that as a report of incorrect spam label? or if not, could/should it?
Today it's ad hoc to me and Varun. We will like make this process more formal / self-serve.
if useful, I think I would be unlikely to "report as not spam" often, but I already engage with some people who pop up in my mentions but not my priority notifications (assuming they are marked as spam?), and so that to me feels like an implicit "mark as not spam" organic to my existing behaviour, rather than requiring additional explicit action of reporting as not spam, if possible
It will be nice if there are ways of “saying” that a certain person is not as spammer, like replying to their casts, interacting with them, following them…
wtf is up with the 1 ETH fee
I think the spam issue stems from people thinking there is a Farcaster token/airdrop coming. At the beginning of the year, tons of influencers on 𝕏 told followers to come to Farcaster and engage to get the airdrop.
They will continue to think that regardless of what we say.
Yeah, I 100% agree! Just an observation over the last 9 months.
Nice one! thanks for this. ⚡
nice write up. especially good to clarify that 'proof of humanhood' is not the solution to fighting spam. spam is a relative/subjective matter.
Great article. Cool to see if from your side. Glad you guys included a "warning" after the first offence as new accounts fall into the spam bucket more often than not. I narrowcast for "spammy" posts (allowance frames/etc) so only channel followers who are interested will see. I wonder if others are doing this.
I know it's a work in progress, but it's nice to know that you are willing to improve your "spam fitering"! I think a and b are especially important. If there is a way to recover from being misidentified, hope can wait. And with a warning, you can think about which of your actions are being misinterpreted as such. I hope this initiative comes to fruition soon! 🐥
Supercast is not too bad
This was an awesome read. Always looking for solutions instead of whining. Also if we could get new follows and new users more on peoples feed it might be helpful to prevent them being labeled.
How about a system that doesn't ask for proof of humanity right from account creation, but instead asks for it when removing spam status? I think this issue could be solved if we can quickly and properly correct any mistaken labeling.
thank you for addressing this it's great that we'll now receive warnings before being labeled as spam but will this feature also inform us about the specific behavior that triggered the warning? because that would help us understand and correct the issue
Thanks for making this an article! So much easier to point people to this!
Why don't you solve the flw display problem?
alot of people need to read this to further understand this spam filter
hey dan, there was a question i threw to v but i haven’t gotten an answer yet and i am bringing it to you in hopes that you can help out, thanks @dwr.eth https://warpcast.com/freymon.eth/0x2c82d318
well explained!
Smart actually 👌
I mean, casting things and never getting replies will also make you stop using the app 🤷🏻♂️
Thanks Dan!
Is it actually the spam filter which is giving new users a hard time? Independently if you are tagged as a spammer or not - as long as you don’t follow me you won’t get notified that I commented on your post and therefore will never get aware I commented Removing the spam filter will not change anything for “real unspammy” users
I like that it’s “EXISTENTIAL” to have new users join Warpcast. Typos can be so poetic.
Minted
Thanks for the information boss
In the latest blog post, @dwr.eth shares insights on the persistent challenge of spam in social networks. The emphasis lies on prioritizing user experience by refining spam filters while ensuring new users feel welcomed. Reading suggests methods for distinguishing between bots and real users to maintain a fun network.