Content
@
https://onchainsummer.xyz
0 reply
26 recasts
26 reactions
Amir Mabhout
@amirmabhout
So your feed not only in interesting to you, but also for many other people. But more interesting is insights and hidden gems inside those feed. If you got a clue on what kind of gem your feed has, you can write it in the field below: (continue 👇)
2 replies
0 recast
0 reaction
Amir Mabhout
@amirmabhout
So in short, you creating a curated content out of your feed, where community can subscribe and enjoy the tasty lattes you brew. (continue 👇)
1 reply
0 recast
0 reaction
Amir Mabhout
@amirmabhout
Demo: https://www.loom.com/share/409cf9b7a7db4ff6a99573ca431453e5 Note: input is limited to 10 casts due to problems we faced with neynar and couldnt get any support. Castlattes are best when atleast 1000 casts are inputed.
1 reply
0 recast
0 reaction
rish
@rish
Hi! Sorry where did you reach out and not get support? Happy to help!
1 reply
0 recast
1 reaction
Amir Mabhout
@amirmabhout
Tried writing in /neynar channel. Saw you were out on the weekend so didn‘t bother on TG. Maybe my posts on your channel not visible, since don‘t have a badge. https://warpcast.com/dudeamir.eth/0x8b9b092d
2 replies
0 recast
0 reaction
Manan
@manan
So sorry about that, I just saw you tagged me as well but unfortunately I had the priority notifications turned on and didn't see the cast in mentions. Switched off priority notifications for now. Anything we could still help you with regard to the project?
1 reply
0 recast
1 reaction
Amir Mabhout
@amirmabhout
Thanks, @caruso33 gets in touch since he is in communication! Generally we want a fast way to get the feed, 100 casts/call too low for having a ux friendly solution where we want to have 24h feed (>1000 casts) to analyze and we might consider setting up a postgres.
1 reply
0 recast
0 reaction
Amir Mabhout
@amirmabhout
Maybe pipe is a solution! We dont show raw feed to users, we send in bulk through a data pipeline to return analytics, possibly whole 24h feed in one go. Also LLMs need all data at beginning to filter, chunk, etc and we already have 1 min for compute, waiting also same amount for data to feed to compute kills ux.
0 reply
0 recast
0 reaction
