As part of a hackathon, cafe-society.news added nostr global channel as a feed and set up alby payment verification. I got the same question a couple times about the name of the project, and the answer is "no, dumbass, that would be blockchainathon, blockathon as in this tool allows you to block z@p@thon dorks. https://bolt.fun/project/blockathon

for cafe-society.news, I wish to have a user train a model, then encrypt it to the public keys of multiple subscribers. In order to subscribe, someone will send a lightning payment to the trainer and put their public encyption key in the memo. As most trainers will be hugely successful, it will not make sense to send a private nostr message to each and every subscriber. I would like to encrypt content to to a keybox

PHP: sodium_crypto_box_keypair - Manual

Randomly generate a secret key and a corresponding public key

and post one encrypted message to nostr. My question is does anyone know of a keybox library for nostr keys? It doesn't really break anything to use 25519 keys, but it would be nice to use the same keys.

I am trying to do payment verification in a web app. Assuming the web app user has received a lightning payment with a memo, is there a way for the web app to confirm payment/ look at that memo without an LN vendor's (alby) oauth2 ? I am a cheapskate and don't want that handshake traffic to go through the web server.)

idea for mechanics of subscription business model business/incentives (mpex/trilema inspired subscription model) cafe-society.news will offer ability to assemble, advertise and pay for subscriptions to updated models via nostr messages. draft trainer will generate/configure a public key for each model they use. Each classifier/model in the settings will have at least one of classifier json, keypair, or public key. Until we determine otherwise, different models can use the same key. Flow Control if the classifier has only a public key, the sofware will know to fetch and decrypt the model along with nostrposts. If the classifier has a private key, software will periodically encrypt and post the ML model to noster during/after a training session. if the classifier has no keys affiliated, then local storage is used and copy paste is still a message medium. Purchases: reader/subscriber sends zaps to trainer with the reader/consumer public key in a memo periodically the trainer will encrypt new model to the subscribers and post encrypted model to nostr. reader/subscriber will fetch new models on nostr as they fetch their other posts. Sales,Classifieds/Discovery: models be posted to (probably 30402) event for classifieds along with the results of wink stats() and an address that is monitored for subscription payments. It may also be a good idea to offer a live service to reply with wink metrics for some sample text. This will help the reader find out if the classifier is relevant before spending money. wtf is wink stats()?

wink-naive-bayes-text-classifier - Wink JS - NaiveBayesTextClassifier

tool for building a personal content filter on the nostr global feed. fixed a bug where model resets unexpectedly. Seems stable enough to ask for feedback

daily news

I can discuss at length how decentralized application development is about censorship resistance, and plays an important role in circumventing and deconstructing the militarized pornographic propaganda industry, but really I just want to save money on hosting.

R.I.P. GOAT whistleblower, Daniel Ellsberg. When Snowden and Assange talk about who led the way they inevitably start with this guy. Everyone on nostr should read his work and keep building tools to support exactly Ellsbergs brand of peace making. We are here versus elsewhere for a reason.