Working on a new DVM. AMA.

Reply to this note

Please Login to reply.

Discussion

sickk improvement

wat did you eat for dinner today

This is dalle btw. It's not about creating the image, it's about what's happening on the image.

did you have a bot problem w the dvm?

or are you going to use dv to tackle a bot problem on nostr?

Mute these nostrich house bots with 2 clicks ๐Ÿ˜…

Will it manage to make pretty image-to-image portraits of women?

People are asking the wrong questions.

The right question would be:

What does it do?

The answer is:

Get rid of botfarm bots in your feed.

Oh, wow, really? Gonna get the 250k bounty?

I don't know ๐Ÿ˜…

It's live on noogle.lol/filter

nostr:note17r404yqau6csmceltx2w3x00l5azv2jk87pdzu7wc4yn5r6uhfusestp5u

Yeah that's why I'm building it ๐Ÿ˜… but it make so much sense.

So, you're training it based off your own npub data or a particular relay or...?

This one is looking for nip05s/names/profile descriptions associated with known bot farms and returns all it's npubs. For now it's targeting specifically nostrich house.

Thoughts on a WOT based DVM?

I personally haven't yet looked into best practices to determine a WOT, but yes it would make sense for so many use cases.

The censorship1984 DVM ias designed so that it expects either 0 inputs, then it uses your follow list, or an arbitrary amount of npubs as inputs, then it uses reports only from these. The inputs could be manual or the outputs from a component determining a wot. As an example.

Dumb question, what is a DVM?

A Data Vending Machine. Imagine Dvms as microservices connected to nostr, doing any kind of data processing from image generation to note discovery, media transcription, or as here, feed curation

How we gonna hide from the dmca hit squads

Sorry, I missed this thread until now (maybe I am blocked ๐Ÿ˜„) - anyway, not all nostrich.house are bots!