Replying to Avatar hodlbod

Ok nostr:nprofile1qydhwumn8ghj7un9d3shjtnhv4ehgetjde38gcewvdhk6tcprfmhxue69uhhq7tjv9kkjepwve5kzar2v9nzucm0d5hsz9thwden5te0wfjkccte9ejxzmt4wvhxjme0qy88wumn8ghj7mn0wvhxcmmv9uqzqwlsccluhy6xxsr6l9a9uhhxf75g85g8a709tprjcn4e42h053vanaunjh you win, NIP 32 was a mistake. I still think it has valid use cases, but it looks too much like an infinitely flexible sub-protocol and is constantly being misunderstood and misused.

To be honest they felt like a good idea to me when I first read the NIP.

I still like them as tags assigned by a third party. We could rewrite the NIP to be just that.

Reply to this note

Please Login to reply.

Discussion

Basically removing the self-labeling part? That would make it clearer.

And removing ontologies. All labels are just hashtags.

This would simply the spec and makes indexing easier for relays, but we loose the data atomicity

Is it a dataset to train LLMS?

Actually it could be that

dan dropping bombs ;)