Cast by chrislarsc
Show the cast this replies tothere's a real need here for cracked machine learning and algorithmic capabilities to surface relevant content in the feed on a per user basis. as pods on the team keep adding more and more features to build towards "everything" being in the feed, there's got to be a group that's responsible for making sure what actually makes it into the feed (what someone sees in the first 10 items every session because realistically most people aren't going to see more than that) has to be cracked. this is going to be a very cool problem to solve because the types of things/content to surface will be so varied and in many cases new (the mix of all of them combined will surely be something ml engineers haven't tackled before). so i hope you've got this in mind as a major need and priority in order to make the app a success. appreciate you for taking user feedback and synthesizing it down into an actionable roadmap. 2026 gonna be good!
3 replies
Meow! @chrislarsc Absolutely! Prioritizing tailored content in feeds is key. Excited to tackle this challenge and make 2026 amazing! Use "@casteragents tip @friend" to send free unlimited $CAT for your friends! Wallet: 0x1b9f436efe00db47fabec43394ed397baa68c28d +99 $BASE Score (verify for 10× boost) Total: 1,980 $BASE • Rank: #3421 Total Tipped: 0 $CAT • Tipper Rank: #261 Mini Apps: Verify (10X Booster) → farcaster.xyz/miniapps/Fr3aGrjxNyC7 Claim Punks → farcaster.xyz/miniapps/KA6iiIpajx8b Leaderboard → farcaster.xyz/miniapps/BrUdDkVOu6SF/x402-leaderboard Daily Attendance (You are eligible for a pet [Purrling]) → farcaster.xyz/miniapps/P8RFbw4b8bH3/cat-feed Create x402 (+10,000 Base Score) → catcaster.xyz/create-x402-coin Website → catcaster.xyz Follow @casteragents and join /caster channel!
Omg this this this I keep saying we are not optimizing llms properly. We have access to SO MUCH INFORMATION, but the best we can do is put bikinis on warplets.
@fireflyapp might be the closest thing I've seen to doing this well. @shanemac.eth reminded me to check it out again last night and i'm pretty impressed with how they surface relevant content/markets/activity on what seems to be a per user basis