I assume you can get to the finish line faster if you haven't abandoned it. If speed is not a concern, I'm happy to step in.
What:
y-nostr - prospectus
Why:
yjs is garnering 500k per week downloads at npmjs. y-nostr will provide a convenient way for developers to add nostr infrastructure to a familiar interface. The resulting adoption of nostr will be astounding.
Who:
Cole Albon with help from existing work
- https://github.com/YousefED/nostr-crdt
- https://www.npmjs.com/package/yjs
- y-indexeddb
- y-websocket
- y-webrtc
- y-nostr <— this project coming soon, needs funding:
How:
The existing nostr-crdt has the functionality, but is broken since a recent upgrade of nostr-tools.
- modify the code in nostr-crdt to work with current nostr-tools and repackage as y-nostr.
- swap out nostr-tools for NDK. NDK comes with a best in class approach to security, and we think NDK is probably up to the task.
Developer Background:
- 1999 - modified the Melissa virus to become one of the first commercial email marketing engines.
- 2000 - as a sales dude, wrote a vt-100 terminal scraper to harvest the Insight Corporation open/orphan orders. Successfully traded resulting lead lists colleagues for favors and food.
- 2013 - launched ripple-bandit.com the worlds first dice address on the ripple payment platform.
- 2016 - created blocktool, a tool to denormalize utxo transactions to a corporate analytics friendly format.
- 2019 - partnered with some dude from Nike - 2nd place in the Portland Blockstack hackathon - dice address for stacks this time.
- 2023 - launched git-nostr a pure shell script program to replace git-ssb (secure scuttlebutt) - allows hosting cryptography projects in the EU.
- 2024 - flat broke, need work soon.
Please send money or direct me to an appropriate grant application.
- cole@getalby.com
📢 Collaborative document editing over nostr: Proof-of-concept
@naddr1qqhyxmmvd3skymmjv96xjmmw94hkut2wdaehgu3d2pex7mmx94hkvttrdahxxetswskh5df5v9exzqgdwaehxw309ahx7uewd3hkcq3ql2vyh47mk2p0qlsku7hg0vn29faehy9hy34ygaclpn66ukqp3afqxpqqqzzmk8q5num
So today I called nostr:npub1zuuajd7u3sx8xu92yav9jwxpr839cs0kc3q6t56vd5u9q033xmhsk6c2uc and we were talking about collaboratively editing documents on nostr; I had this very left-side of the curve idea on how to do this, so after kids' bedtime I implemented it to see what it would feel like.
How it works:
A user can invite other users to edit a document (i.e. a long-form, or any other kind of document)
* The invitees can be added/removed at will by the owner of the document.
* The users send ephemeral events while live editing and can hit "save" to propose a new version of the document.
* Everybody sees what the others are doing in real time
* The document's owner chooses which version to make "official".
* Documents can be forked so any of the invitees (or anybody else if the document is allowed to be public) can fork the document and continue working on their own version.
I've been trying without success to get nostr-crdt to sync/share documents related to machine learning model updates. I'm calling them "classifiers". I'm not sure how the ui will work for access control, so I'll be paying attention. https://github.com/colealbon/cafe-society.news/blob/main/src/defaultClassifiers.ts
I vaguely recall the W administration talking heads saying "don't change presidents in the middle of a war!" Willing to wager we hear that old saw soon.
Might be my filters, but I don't see as many obnoxious zapathons as I did 10 months ago.
If someone could tell me what code functionality change happened, I would love you forever
Have no idea why but this fixed a bug
https://github.com/MutinyWallet/blastr/commit/f46e90ecd323650d30b5d2a81a1ec3747ff6daef
Eyeballing it - looks like the "map" on the new line 361 would cause the filter contains x code to run on each item instead of all the items.
cafe-society.news is a tool to do exactly that
I'm looking for honest money. I have some nostr experience (25 github stars for git-nostr) and wrote a tool to train a spam filter against the global nostr feed. I'd prefer to discuss in public, but if you wish to pm me let me know how (which app) to view your message.
2 minute video: local model ML assisted job search with cafe-society.news. avoid duplicate posts and sort by trained, self hosted ML model https://youtu.be/AXoMyDWE24M
Don't know how I missed this for 4 years, but looks like Tangela Tricoli's stinky poodle has been remastered. Nominating Jet Lady as the official Nostr LP. https://www.youtube.com/watch?v=jPI6XztUJhU&list=PLkbCIkhKRjsYgn6paDWKQj2PxI0jFFuSx&index=8
how does this compare?
cafe-society.news 30 day daily visitor metrics:
Mean 58.4
Median 55.5
Mode 57
Range 120
Minimum 31
Maximum 151
Count n 30
Sum 1752
Quartiles:
Q1 --> 45
Q2 --> 55.5
Q3 --> 65
Interquartile
Range IQR 20
Outliers 151
cafe-society.news interesting new idea default "section recipe": https://cafe-society.news/rssposts/shadowysupercoders Consists of njump.me rss/nostr feeds combined, then trained to sort by "dev/not dev" content. You might have to clear your cache or grab it from a private session window.
anyone who submits a PR with their npub self evidently deserves to be in the list. https://github.com/colealbon/cafe-society.news/blob/a9c8bc7f850a3e1aeb7d903ba6fe22fba54308b0/src/defaultRSSFeeds.ts#L154
cafe-society.news interesting new idea default "section recipe": https://cafe-society.news/rssposts/shadowysupercoders Consists of njump.me rss/nostr feeds combined, then trained to sort by "dev/not dev" content. You might have to clear your cache or grab it from a private session window.
I'm trying to set up a paywalled multi party sync service on cafe-society.news to share user trained ML models (a few MBs of json). I've investigated yjs and nostr storage. Is HORNETS something I should investigate?
read the newspaper every day: https://cafe-society.news/rssposts/world
cafe-society.news now works with njump.me. With this improvement, nostriches can now consume individual or groupwise nostr feeds in order to train ML models.
I tried, but looks like njump.me feeds don't get parsed correctly by cafe-society.news state of the art rss parser. Unless someone takes it on, I'll put atom feed parsing it on the todo list. This would void alot of work I would otherwise have to do to grow beyond the nostr global feed. https://njump.me/npub1a2cww4kn9wqte4ry70vyfwqyqvpswksna27rtxd8vty6c74era8sdcw83a.rss
the problem is almost certainly here. https://github.com/colealbon/cafe-society-old/blob/450680a0f987ccfd3cf03c77c1be5c6118d2090c/src/App.tsx#L203C38-L203C38
I tried, but looks like njump.me feeds don't get parsed correctly by cafe-society.news state of the art rss parser. Unless someone takes it on, I'll put atom feed parsing it on the todo list. This would void alot of work I would otherwise have to do to grow beyond the nostr global feed. https://njump.me/npub1a2cww4kn9wqte4ry70vyfwqyqvpswksna27rtxd8vty6c74era8sdcw83a.rss