Content
@
1 reply
0 recast
2 reactions
Michael Pfister
@pfista
The only reason you should build a Warpcast-flavored farcaster client right now is if you have a large pre-existing and external community looking for a new home. If 80% of your features overlap with things warpcast already does, you're gonna have a hard time. Strongest differentiators come down to - Net new community - Orthogonal features that have very low likelihood Warpcast will build them in Warpcast isn't stealing your feature. It's your job to understand Merkle/Warpcast's own product strategy and incentives, then build your product in a way that's symbiotic.
19 replies
10 recasts
203 reactions
/naive
@na
or having a better ml model
1 reply
0 recast
1 reaction
Michael Pfister
@pfista
better ml model doesn't matter if there isn't enough good content in the first place
1 reply
0 recast
2 reactions
/naive
@na
maybe but an alt model that's open provides incentives for people who do have a external communities
2 replies
0 recast
1 reaction
Trigs
@trigs
What would be better is to just create the different ML model and get warpcast to give users choice on what algorithm they like best. Don't have to create a whole new app just to modify the algorithm.
1 reply
0 recast
0 reaction
/naive
@na
algorithm enables new app, not the reverse. ml is base quality-of-life tech
0 reply
0 recast
0 reaction
