kayabaNerve
commented
1 year ago Apologies for the delayed reply. We generally try to respond within 3 weeks.
Monero primarily doesn't use Discord, yet IRC and Matrix. While there are fundamental questions on the usability of LLMs, the main point is Monero doesn't have DevRel/Customer Support. Accordingly, not only can Monero not well meet your pre-implementation agenda, yet this doesn't benefit/relieve any parties within Monero.
We did collect community sentiment to be sure, and did not find sufficient interest. While we appreciation the submission, we have voted to decline this grant.
Please, let us know of any future endeavors which may better fit Monero.
Candidate Background
Nigel Alford is a technologist and entrepreneur focused on decentralized solutions. In his previous venture, Nigel created a Nervos NFT marketplace for Book authors. Nigel has been involved with various technical projects from blockchains to digital maps. He is dedicated to reducing technical friction to start development and moving faster into projects.
Github: https://github.com/mrnigelalford Linkedin: https://www.linkedin.com/in/nigelalford Twitter: https://twitter.com/AlfordDesign
Project Description
We have built, docIT a private Discord bot (allowing users to “talk to your docs”) that supports multiple document sites, with evolving LLMs. This bot will interconnect Monero's key Github repositories, documentation, related articles, and forum posts, providing users with a platform to ask questions and read responses in a natural language with support links. Our goal is to incorporate docIT with Monero to help lower education hurdles. Importantly, we've built DocIT to prevent hallucinations and respond with sources to give transparent and discoverable answers.
This solution will enhance Developer Relations, Customer Support, and Community Support experiences within a Monero's ecosystem, starting point being Discord.
Our primary objective is to augment the Monero DevRel and Community Support teams, lessen the support hours on Discord, and speed up responses to frequently asked questions. This ensures that both the teams and users get answers more effectively.
Technical Approach
Language: Typescript | Node.js Database: Supabase | self-hosted docker files Hosting location: Digital Ocean Scaling: Increased instance sizes will be utilized. If needed, will setup shared instances across VMs to balance load. CI/CD: Github Actions into Digital Ocean All secrets stored in environment variables Data Sources: Documentation, primary repo, open to pulling chat history or other sources if needed.
UI Development
App will front to Discord Bot Commands: Direct mention docIT with a question. DocIT will respond in a thread with response. All additional communications can exist in a threaded conversation. Happy to create more commands or interactions if needed.
Milestones and Budget
Project Plan (4 week delivery)
Pre-Implementation: Collaborate with the DevRel, Customer Support, and Community Support teams to comprehend their needs, identify pain points, and decide where the bot can be most beneficial. These insights will direct the selection of data sources for implementation.
Milestone 1 - Implementation of Data Sources [2 weeks]: Post the discovery phase, we will incorporate 3-4 high-value data sources into the bot. We will crawl data and create OpenAI emeddings in our db to index and setup search on queries. This milestone will involve refining file loading, index creation, and data source integration.
End state: Have 3-4 data sources that can be queried. Expected bugs: hallucinations, mismatches on query/documents.
Milestone 2 - Testing and Optimization + Deploy to Production [2 weeks]: This critical phase will concentrate on eliminating hallucinations and refining prompt engineering to ensure only pertinent answers are given, with source material and links for users to follow up on. The aim is to make the bot production-ready.
End state: A production-ready bot approved by the Customer / Community Support teams, ready for deployment on the Monero's Discord.