The name of our project will be the crowd-sourced data submission platform. But we must still choose a loose name for the bitcoin education chatbot.

Tasks

• Choose name for bitcoin education chatbot
• Choose name for new data submission platform
• Create banner for new data submission platform

Requirements

Below are general requirements that aren't super designed. Let's have Niku take a first pass at it and we can provide feedback in figma. Comment below if I missed anything and we can update the issue.

Structure

• Designed for desktop only (Chrome preferred. Then firefox if possible)
• Margin checked for 4px spacing.
• Padding checked for 4px spacing.

Chatbot

• Title text that is the name of the bitcoin education chatbot
• Rectangular area for users to type a question to the chatbot
• Area where response is displayed to the user (maybe this is expanded after a question is submitted).

Chatbot States

• KNOWN: If chatbot knows the answer, response is displayed.
• UNKNOWN: If chatbot doesn't know the answer, error message is displayed and CTA to potentially suggest data to LLM
• THINKING: While chatbot is thinking, show a pending logo/icon. (Stretch goal: image of a person thinking or bitcoin logo thinking that's animated)

CTA to submit

• A button below the chatbot users can click. Title of CTA is "Submit New Data". This will bring up a pop-up for our data submission portal.

Footer

• Include link to our github.

Requirements

Below are general requirements that aren't super designed. Let's have Niku take a first pass at it and we can provide feedback in figma. Comment below if I missed anything and we can update the issue.

Structure

• Designed for desktop only (Chrome preferred. Then firefox if possible)
• Margin checked for 4px spacing.
• Padding checked for 4px spacing.

Train the model

• This section should include an area to submit a text or a pdf to the LLM.
• A section to submit lightning address.
• A button to submit data. Data + lightning address should be filled prior to being allowed to submit.

States

• NO DATA: Default state of the website where user sees options to submit data.
• DATA LOADED: State where user submits either a text/pdf and lightning address. User should not be allowed to submit until both are provided.
• PENDING: LLM is reviewing the submission.
• ACCEPTED: Data is accepted. Congratulation message + confirmation payment has been sent.
• REJECTED: Data is not accepted. Message explaining failure. Include link to our github to raise an issue.

Research whether or not we should use lnd's API. and integrate it into our project.

Goals

There are two goals of this training:

1. Teach it to answer technical bitcoin questions
2. Teach it how to filter data

Answer technical bitcoin questions

Initial data we feed should include bitcoin optech glossary, BIP’s, LIP’s, and current version of bitcoin core.

Filter criteria

As an LLM, I want to use various criteria to filter the data so that it doesn’t skew my current work.
The criteria includes:

• Verifying acceptable sources such as: stack exchange (must have at least 50 votes and will accept the highest answer), bitcoin optech, BIP’s, LIP’s, old versions of bitcoin core, and updated versions of bitcoin core
• Must be above 50% relevancy with current database wrt to how related it is. (AI can provide % guesstimate)

Data exclusions

We should exclude information about ARK. This will be the question we submit to the chatbot during our demo and the chatbot will say they don't know anything about it. We will the submit info about ARK to the chatbot and the chatbot will then be able to return a correct response.

This will be a stretch goal once other aspects of the project has been completed.

Here is the logo.

This will be a stretch goal once the chatbot has been completed.

Tryin using alby chrome extension. Might be easiest.