Lecture notes AI — every class, instantly searchable

By the time you finish a four-year degree, you have maybe 50 courses × 14 weeks × 2 lectures = 1,400 lecture notes. Plus readings. Plus slides. None of it is searchable. Most of it gets thrown away after exams.

That's the problem. Lecture notes are the highest-value content in your education, and nobody can find anything in them once the course ends. AI changes the economics.

The setup

1. Upload lecture notes per course as PDFs or DOCX. Photo the handwritten ones with a scan app.
2. Folder per course: CS-101 - Intro to Computer Science / Week 03 - Lecture 6.pdf
3. Build one Assistant per course while it's active.
4. After the course, archive to a "Past Courses" Assistant scoped to all of them.

Real questions

During the course

• "Where did the prof discuss recursion? Cite the lecture date."
• "What did we say about the difference between supervised and unsupervised learning?"
• "Find every mention of the term 'big O' across all lectures so far."
• "What was the example the prof gave for amortised time complexity?"

Cross-course / after graduation

• "In any of my undergraduate courses, did we cover the Bellman-Ford algorithm?"
• "What did my marketing professor say about competitive positioning vs my strategy professor?"
• "Find every course where I learned about Bayesian inference, with the lecture date."

These cross-course questions are the killer use case. They're impossible without a unified searchable archive.

Why this changes how you study

• Recall before exam: instead of re-reading 14 weeks of notes, ask targeted questions.
• Connecting courses: see how the same concept was taught in different courses (sociology vs psychology, micro vs macro).
• Year-end synthesis: "What were the big ideas of this semester?" — useful for writing summaries and reflective essays.
• Career-time recall: five years after graduation, finding "what did my prof say about X?" is genuinely useful.

Tips for higher-quality lecture notes

• Type or handwrite + scan. OCR works on legible handwriting but typed is dramatically better.
• Date and label every note. 2026-09-15 - Week 3 - Lecture 5.pdf beats lecture.pdf.
• Include the slide handouts. Prof's slides + your notes together is more searchable than either alone.
• Add a 30-second "synthesis line" at the end of each lecture note: "The main point today was X." Your future self thanks you; the Assistant uses it for retrieval.

What AI doesn't do

• Replace actual understanding. If you didn't understand the lecture, the notes will be cryptic and the AI will struggle.
• Compensate for missing notes. Skipped lectures = blind spots in your archive.
• Replace flashcards for retention. Use AI for recall and connection-making; use Anki for memorisation.

A workflow over a degree

• Per lecture: capture cleanly, upload weekly.
• Per course: dedicated Assistant during the semester.
• Per semester: archive to a master Assistant.
• Per year: ask the Assistant for synthesis questions.
• Post-graduation: keep the archive. You'll use it in your first job.

The students who do this in undergrad arrive at grad school years ahead — not because they're smarter, because their past is queryable while everyone else's is in a box.