Sumil
@sumil
Folks, what's your take on Dorsey's AT Protocol? @dwr.eth @cassie https://blueskyweb.xyz/
7 replies
0 recast
0 reaction
Cassie Heart
@cassie
Extremely limited documentation, not nearly enough to formulate a concrete opinion. All I can really say is from the perspective of what I see now, and having to make conjectures. Here's the high level:
1 reply
0 recast
0 reaction
Cassie Heart
@cassie
- This took three years? How many people were working on this? - The spec offers no opinion on how indexers _should_ or even _can_ work. While it's understandable to not decide the best approach, there is no guidance at all, which suggests in part they may not have even tried to build one outside of a naive one.
1 reply
0 recast
0 reaction
Cassie Heart
@cassie
- DID (in terms of the bloated but also unclear standard) is a headache, and even they know it, so they have a placeholder definition. - They already outline a call for media embeds, which is incredibly naive for a federated protocol without a solid p2p (at scale) video streaming protocol in existence yet.
2 replies
0 recast
0 reaction
Cassie Heart
@cassie
- Why didn't they just incorporate distributed search and forego indexers? Given they don't have E2EE baked into their design it's certainly feasible. Again, they've been at this for three years? - Zero incentivization model. Not even a consideration.
2 replies
0 recast
0 reaction
Eric Kuhn
@ekuhn.eth
What I thought was strange was they are using a DID but decided to make up their own which they are clearly saying will be depreciated. The DID longevity is essential
0 reply
0 recast
0 reaction
