Chris Dixon
@cdixon.eth
I used to think venture investing is 80% an intellectual test, 20% emotional test. Now I’d say it’s the reverse. I’d say the same about my experiences as a startup founder. Probably true of many other long-term activities.
18 replies
7 recasts
122 reactions
🌹 zach harris 🥀
@zachharris.eth
@cdixon.eth will we ever see a hunch client on top of Farcaster protocol? Given the slew of new channels, I’d love a recommendation engine that surfaces up high-quality FC signal. Feels like an opportunity… NS
1 reply
0 recast
1 reaction
Chris Dixon
@cdixon.eth
Ha, yeah, miss Hunch :) and it would be 100x better with the new AI tech :) I think it would be cool if someone built a new recommendation engine where the user owned the taste data and only let the AI borrow it on a per task basis. So data is portable and apps compete for users without relying on data lock-in.
3 replies
1 recast
12 reactions
Yassine Landa
@yassinelanda.eth
And have the AI pay for the user taste data proportionally to how much it make better for everyone else.
1 reply
0 recast
0 reaction
Chris Dixon
@cdixon.eth
Yes. Also if I did it again I’d start with a vertical (eg books) and run the “bowling pin strategy” vs trying to make a generalized recommendation system from the get go.
1 reply
0 recast
2 reactions
🌹 zach harris 🥀
@zachharris.eth
Books didn't work for Small Demons, but they were ahead of their time. Going narrow vs wide probably sense as an architectural choice as a specific folksonomy with a few phylums is likely easier than trying to boil the ocean with a taxonomy that has lots of nested phyla. https://www.youtube.com/watch?v=XOoOUBkdUMc
0 reply
0 recast
1 reaction
