Project ideas from Hacker News discussions.

10-202: Introduction to Modern AI (CMU)

📝 Discussion Summary (Click to expand)

Three prevailing themes

Theme Key idea Supporting quotes
1. Curriculum breadth – classical AI vs modern LLMs Many participants feel current AI courses focus too narrowly on machine‑learning and ignore symbolic reasoning, knowledge representation, planning, etc. “I think the problem is the under representation of other branches of AI research: knowledge representation, automated reasoning, planning, etc.” – gignico
“The AI course I took was mostly symbolic methods and some classic ML at the end. Most students were not interested at all and would've probably been more engaged studying ML directly.” – Kaethar
“Nothing on symbolic reasoning?” – sim04ful
“That would be part of whats now 'classical ai'” – cultofmetatron
“That would be the exact opposite of modern” – xdavidliu
2. AI‑tool policy in coursework Courses should allow AI assistants for learning but discourage over‑reliance on them for graded work. “Students are permitted to use AI assistants for all homework and programming assignments … but we strongly encourage you to complete your final submitted version of your assignment without AI.” – aanet
3. Audience awareness & relevance Many people are unfamiliar with terms like “LLM” and the variety of generative‑AI models, so course material must be tailored to the target audience. “It really depends on the target audience, because a lot of people have no idea what they are using is called an LLM or that there are various types of generative AI.” – axseem

🚀 Project Ideas

Generating project ideas…

Modular AI Learning Hub

Summary

  • A web‑based, self‑paced learning platform that lets students build their own AI curriculum by selecting modules on symbolic reasoning, planning, classical ML, RL, vision, etc.
  • Provides interactive notebooks, quizzes, and project templates, giving learners control over depth and breadth beyond the typical LLM‑centric courses.

Details

Key Value
Target Audience Undergraduate/graduate students, self‑learners, educators seeking flexible content
Core Feature Drag‑and‑drop curriculum builder with instant code‑execution sandbox
Tech Stack React + Next.js, Python backend (FastAPI), JupyterHub, Docker, PostgreSQL
Difficulty Medium
Monetization Revenue‑ready: tiered subscription (free tier + $9/mo for advanced modules)

Notes

  • HN commenters lament “courses are set in stone” and lack of symbolic AI coverage; this platform lets them “choose what you study as much as possible.”
  • Encourages discussion on best practices for balancing LLMs with classical AI, and provides a community forum for sharing custom modules.

Course Design Toolkit

Summary

  • A web app that assists educators in drafting AI curricula, automatically generating lesson plans, assessment rubrics, and AI‑policy templates (e.g., “use AI for research only, no AI in final submissions”).
  • Includes a repository of vetted content from multiple AI subfields and a peer‑review workflow.

Details

Key Value
Target Audience University instructors, curriculum designers
Core Feature AI‑driven curriculum generator with policy‑policy integration
Tech Stack Vue.js, Node.js (Express), MongoDB, OpenAI API for content suggestions
Difficulty Medium
Monetization Revenue‑ready: institutional licensing ($5k/year) + per‑course add‑ons

Notes

  • Addresses the frustration that “AI is much broader than LLMs alone” and the need for “structured course to learn how LLMs work” while maintaining academic integrity.
  • Sparks debate on how to balance AI assistance with learning outcomes, echoing the policy discussion in the thread.

AI‑Aware Assignment Tracker

Summary

  • A platform that lets students submit coding assignments with optional AI assistance, automatically logs AI usage, and flags potential over‑reliance before grading.
  • Provides educators with analytics on AI usage patterns and supports “complete final submission without AI” policies.

Details

Key Value
Target Audience Students, instructors, academic integrity officers
Core Feature Real‑time AI‑usage monitoring, plagiarism‑style AI‑generation detection
Tech Stack Django, PostgreSQL, OpenAI API, Docker, CI/CD pipelines
Difficulty High
Monetization Revenue‑ready: per‑institution license ($3k/year) + per‑assignment fee

Notes

  • Directly tackles concerns about “poor final exam results and/or cheating” and the policy that “students should complete final submissions without AI.”
  • Provides a data‑driven conversation about the role of AI in education and how to enforce fair use.

Read Later