Oddbean new post about | logout
▲ ▼
Dustin | 6 months ago (raw) | root | parent | reply | flag +3 
 How would it be different than DVMs?
▲ ▼
DannyMorabito | 6 months ago (raw) | root | parent | reply | flag 
 from what I gather from his message, they run on your machine, think of it as custom private "feed algorithms" that you can use and you control and can sell to other users
▲ ▼
jb55 | 6 months ago (raw) | root | parent | reply | flag +5 
 these run locally, are private, don’t depend on an internet connection. They are turing complete nostr queries that run locally.

DVMs are inherently about asking something from an external resource which may not exist or stop working.
▲ ▼
Rand | 6 months ago (raw) | root | parent | reply | flag 
 bare metal buy
▲ ▼
nielliesmons | 6 months ago (raw) | root | parent | reply | flag 
 Huuuge!
@Don't ₿elieve the Hype 🦊 this is 💯 what I meant with "DVM's" that run locally.
▲ ▼
mleku | 6 months ago (raw) | root | parent | reply | flag 
 you could also do something else, for space conserving reasons to trade off more advanced indexing, while only noting where the data was seen via a compact serial index

events are so small that even simple filter indexes take up quite a bit of space... probably would help to have varints for timestamps event serials, but yeah, easily 128 bytes or so for a basic filter index, if you want more advanced stuff than that it's gonna reach towards being as big as the average 750 byte events themselves

gonna think about that idea though, how to compactly store the timestamps and indexes, these fields occur repeatedly within the index and for access counters as well for the need of managing a garbage collector