How I stopped taking notes in lectures (and my grades went up)

There's a well-established finding in cognitive science that writing notes by hand in a lecture produces better retention than typing. The reason is counterintuitive: because you can't write fast enough to capture everything, you're forced to process and summarize in real time. The act of selection — deciding what's worth writing — is a form of active learning.

The finding is true, and it misses something important.

The selection pressure that produces better retention also produces an incomplete record. You are choosing, word by word and clause by clause, what to keep and what to let go. The professor says six things in 30 seconds. You write two of them. The four you didn't write don't disappear from the air — they disappear from your notes, permanently, because you were writing the first two.

Across a 90-minute lecture, you make thousands of these micro-selections. You leave with a record of roughly 40% of what was said, filtered through the assumptions you brought in with you about what was important. When you come to revise eight weeks later, you're revising from those assumptions — not from what the professor actually said.

The listening/writing split

Cognitive science is helpful here too. Working memory has limited capacity. Processing incoming speech, encoding it into meaning, and simultaneously producing written output are three competing demands on the same limited resource. You can do all three — just not at full quality simultaneously.

The result is a constant trade-off: write well and miss content, or capture more content and write less clearly. Every student finds their own equilibrium. Most find it somewhere that leaves them with notes they can barely read and a recording they're never going to re-listen to.

The recording problem is worth dwelling on. Most students know they could record their lectures and re-listen. Almost none do it systematically, because re-listening to a two-hour lecture takes two hours. The recording is insurance — something to consult when you specifically can't find a particular thing in your notes. It's not a usable revision resource.

What changes when AI processes the lecture

The insight that transforms this is not that AI is better at taking notes than you. It's that AI separates transcription from comprehension.

When you're in a lecture with Kuulo recording:

• You don't need to write to capture. The transcript is being built automatically.
• Your working memory is free for comprehension — the processing layer that writing was competing with.
• You can make eye contact with the lecturer. You can respond to questions. You can think about what they're saying rather than racing to capture it.

After the lecture, you don't have a transcript to wade through. You have a structured summary — generated in under two minutes from the full recording — with key concepts, the lecturer's definitions, points they flagged as important, and the logical structure of the session. The 90-minute lecture is readable in 10 minutes.

This is not a productivity trick. It is a change in the fundamental relationship between attending a lecture and learning from it.

The practical workflow

Here's what this looks like day to day:

Before the lecture: Open Kuulo, select the template that matches your course — Lecture Notes (generic), Lecture Notes (Medicine Year 1), Law Lecture Notes, and so on. The template shapes how the AI structures the output, so the summary for a clinical pharmacology lecture looks different from one for a constitutional law seminar.

During the lecture: Tap record. Put your phone on the desk, face down, or in your bag. The transcription runs in the background, including through your phone being locked. You sit in the lecture and listen to it.

After the lecture: Tap to generate the summary. Two minutes. A structured note appears: the main concepts with definitions, the examples used to illustrate them, any quotes or phrases the lecturer flagged as important, and a brief synthesis of the session's argument or content. The full transcript is available if you want to find a specific phrase.

For revision: Across a semester, each lecture's summary is searchable. Three weeks before finals, you generate a Revision Pack — all your recordings for a subject, consolidated into a single structured document, sorted by lecture order, with key concepts and the professor's own phrasing preserved throughout.

The note card moment

One consequence of having AI notes available immediately after a lecture is that you can share them.

Kuulo generates a formatted note card — a clean, visual summary of the session — in one tap. The card shows the date, the duration, the main topics, and the first few lines of the summary. It's sized for WhatsApp, iMessage, or Instagram Stories.

When you drop this card in the course group chat fifteen minutes after a two-hour lecture, the implicit message is clear: while everyone else is packing their bags, you have a structured record of the session. The question this creates in the reader — "how did they do that?" — is the beginning of a conversation about the tool. It doesn't need selling. The output is the demonstration.

This is how the tool spreads: not through advertising, but through people seeing a concrete output from a lecture they were also in and immediately understanding what it represents.

Offline: the feature nobody talks about until they need it

Kuulo runs entirely on-device. This matters in lecture theatres more than almost anywhere else.

University buildings built before wireless networking was common — which includes a significant proportion of the teaching spaces at most UK universities — have notoriously unreliable Wi-Fi in auditoriums and lecture theatres. Older buildings, underground rooms, and overcrowded spaces all suffer from the same fundamental problem: the wireless infrastructure wasn't designed for 300 students simultaneously streaming audio to cloud services.

Otter.ai, Granola, Fireflies — all require internet for any AI functionality. In a lecture theatre where the Wi-Fi is unreliable, they fail. Kuulo doesn't use the Wi-Fi.

The exam season payoff

The real value of a semester of Kuulo recordings only becomes apparent in the weeks before exams. Here's what you have that your classmates don't:

Complete records. Not notes filtered through what you thought was important in October. The actual content of every lecture, including the clarifications the professor gave in response to student questions, the examples they returned to, the phrasing they used when they knew something was exam-relevant.

Searchable content. You can search across every lecture's transcript for a specific case, concept, or term. When a past paper question references a doctrine you're not sure about, you can find every time the professor mentioned it, in context.

The professor's own words. Essay answers that use the professor's framing — the specific way they articulated an idea — demonstrate understanding in a way that generic textbook language doesn't. You have the source.

A Revision Pack. All the semester's recordings for a subject, consolidated into a structured document. The conceptual arc of the course, visible as a coherent whole.

The asymmetry between a student who has been doing this all semester and one who hasn't is most visible in the week before finals. One has a structured, searchable revision resource built from the actual teaching. The other has notes from October that they wrote while listening and can't quite read anymore.

The listening/writing split, resolved

There is a version of this that's worth being honest about: AI notes don't make you a better learner automatically. The transcript exists, but you still have to engage with it. The summary is a starting point for revision, not a substitute for understanding.

What AI notes do is remove the constraint that forced you to choose between listening and writing. You no longer have to miss the second half of a sentence because you're still writing the first half. You no longer have to decide, under time pressure, whether to capture this point or the next one.

The cognitive load you used to spend on capture is now available for comprehension. Whether you use it depends on whether you actually pay attention in the lecture. The tool can't make you interested — but it can make sure that when you are interested, none of it is lost.

That's the deal: stop taking notes, and start listening. The notes happen anyway. And they're better than anything you would have written.