AI Writers
Yatendra Kumar
Problem
Writing can often be time-consuming, tiring, and occasionally frustrating. Whether it's composing an email, jotting down a to-do list, or drafting a blog post, the challenge lies in the mental effort required to organize thoughts and ideas coherently. Moreover, prevalent issues like writer's block can compound these difficulties. There is a clear need for intelligent tools that can assist in generating and organizing content efficiently.
Solution
AI-powered writing assistant using Flutter and the GPT-4 model from OpenAI. The assistant, tentatively named "AI Writers", will assist users in generating and organizing content efficiently. This includes automatic to-do list generation, blog post drafting, and outline creation.
Similar to the notion AI, we can keep a button or just press space for "ask AI"
Goals
The successful completion of this project will involve:
A fully functional API wrapper using Rust that uses AI to generate to-do lists, create outlines, and draft blog posts.
A user-friendly interface that encourages engagement and facilitates ease of use.
Comprehensive documentation covering the architecture of the application, functionality, usage instructions, and the integration process of the GPT-4 model.
A well-structured thoroughly commented codebase for easy understanding and future development.
Implementation Design
The application will be developed using Rust & Flutter, which allows for cross-platform compatibility.
OpenAI's GPT-4 model will provide AI-powered writing assistance.
Rust will be used for network requests to the OpenAI API.
Timeline:
Week 1 (19/06/2023):
Tasks:
Milestones:
Gain a comprehensive understanding of GPT-4 capabilities and functionalities
Gain in depth understanding of appflowy best coding practices integrated in CI/CD.
Determine if there are existing Rust libraries available for GPT-4 integration
Week 2 (26/06/2023):
Tasks:
Milestones:
4. Develop a basic Rust wrapper for OpenAI API with minimal functionality.
5. Understand the basic structure and components of the AppFlowy backend.
Week 3 (03/07/2023):
Tasks:
Milestone:
5. Build an end-to-end demo system using gpt-4 with proper prompting
Week 4 (10/07/2023):
Tasks:
Milestones:
7. Upgrade Rust wrapper for OpenAI API with advanced features.
8. Gain a deep understanding of AppFlowy backend, including data flows and dependencies/
9. Raise a PR for issue. (raised pr: #3009).
Week 5 (17/07/2023):
Tasks:
Milestones:
10. Finalize application design and UI.
11. Begin the process of integrating GPT-4 Rust Wrapper with rust backend structure.
12. Successfully integrate GPT-4 Rust Wrapper with rust backend structure.
13. Initiate the development of the to-do list feature.
Week 6 (24/07/2023):
Tasks:
Milestones:
14. Complete the development of the to-do list feature
15. Initiate the development of the outline creation feature
Midterm Valuation
Gained comprehensive understanding of GPT-4 and AppFlowy's coding practices.
Identified potential Rust libraries for GPT-4 integration.
Developed a basic Rust wrapper for OpenAI API.
Explored AppFlowy's backend structure.
Built a demo system using GPT-4 with proper prompting.
Enhanced Rust wrapper for OpenAI API with advanced features.
Delved deeper into AppFlowy's backend complexities.
Resolved a bug in AppFlowy (PR: #3009).
Finalised application design and UI (similar to Notion AI).
Initiated and completed GPT-4 Rust Wrapper integration with backend.
Began development of auto-generating to-do list feature and completed it.
Initiated development of outline creation feature.
Week 7 (31/07/2023):
Tasks:
Milestones:
16. Achieve modular and scalable Rust wrapper API code.
17. Completed tests for the API with 100% coverage.
Week 8 (07/08/2023):
Tasks:
Milestones:
18. Completed and polished frontend Flutter interface.
19. Full test coverage for the entire frontend interface.
20. Deep understanding of the integration flow from backend to frontend.
Week 9 (14/08/2023):
Tasks:
Milestones:
21. Successfully fixed the Evernote importer issue (#2971).
22. Successful integration of the Rust wrapper API with the Flutter frontend.
23. Completed tests for the integrated system.
Week 10 (21/08/2023):
Tasks:
Milestones:
24. Complete the project documentation.
25. Address and fix any bugs or issues that may arise during testing.
26. Initiate the migration of current OpenAI features to using Rust wrapper.
Week 11 (28/08/2023):
Tasks:
Milestones: 27. Successfully migrate all current OpenAI features to use Rust wrapper 28. Continue comprehensive testing and debugging of the entire application 29. Ensure all identified issues are resolved and application is fully functional
Week 12 (04/09/2023):
Tasks:
Milestones:
30. Successful project review and preparation for final live deployment.
31. Successfully prepare the application for deployment, with all features working as expected
32. Completion of a comprehensive blog documenting the entire project.
Impact
The successful implementation of this project will provide a powerful tool for users who require assistance with various writing tasks. By automating these tasks, users can focus more on their ideas and less on the mechanical aspects of writing.
Given the increasing demand for AI-powered tools and the widespread use of mobile devices, this project can potentially benefit a wide range of users.
Risks and Mitigation
Risk: Unfamiliarity with the GPT-4 model.
Mitigation: My recent completion of a course on prompt engineering from OpenAI gives me a solid foundation to understand and utilize the GPT-4 model effectively.
Risk: API limitations, changes, or costs associated with using GPT-4.
Mitigation: For cost mitigation, the application will require the users to provide their API keys to access GPT-4, allowing the usage costs to be borne by the user directly.
Risk: Time management due to other commitments.
Mitigation: Adherence to a strict schedule and prioritization of tasks will be key in managing this project alongside other responsibilities.
Last updated