kenny 🎩
@kenny
and yeah maybe I'm wrong and it isn't a 10x improvement but what's the drawback to making the change? 1 hour of dev resources to implement and clear statistical evidence to back the decision: https://paragraph.xyz/@yesyes/checking-the-comment-velocity-of-real-users at worst it...makes the following feed "messy"? who cares?
1 reply
0 recast
3 reactions
shazow
@shazow.eth
Maybe the way to validate is to get @woj.eth to try this on @super? Wonder if supercast has good analytics that can be isolated to just other supercast users.
1 reply
0 recast
0 reaction
kenny 🎩
@kenny
my point on this has always been it has to be a network-wide thing to be truly measured because Farcaster is too smol majority of users have to be retrained now to expect to be able to browse following and discover conversations since that hasn't been true since February 2024 (see the analysis above) power users who use Supercast aren't going to be hugely effected imo, they already browse deep
1 reply
0 recast
0 reaction
shazow
@shazow.eth
I agree with your point but I do think supercast is big enough to do a statistically meaningful A/B test on a change like this, assuming the interaction population can be limited to just supercast users.
1 reply
0 recast
0 reaction
kenny 🎩
@kenny
fair enough, I just don't see why we need the test when we ran the statistics from February, across the entire network's dataset, and the impact was disastrous you saw diverse comment threads drop by >50% waiting on a Supercast test to justify (that's still never going to be perfect because most of the network is on Warpcast) doesn't make a ton of sense to me because it's such a pressing problem
1 reply
0 recast
1 reaction
shazow
@shazow.eth
Charitable take: At any given time, MM is experimenting with a number of tweaks, they have no shortage of ideas to try. Some are probably yielding positive results, and bringing back reply bumping is just not high enough priority (perhaps that had other externalities that they considered negative). I don't think any amount of pleading will change that. But another client who proves out that "actually this is good" and maybe even starts securing a larger userbase because of this feature will definitely get attention and prompt to re-evaluate the priorities of things like this. Or supercast will try it and find the same negative externalities that make it not worth it, and Warpcast will have avoided the costly cycle of iteration.
1 reply
0 recast
0 reaction
kenny 🎩
@kenny
that's a fair charitable take 🤝
0 reply
0 recast
1 reaction