iXnote
Get iXnote
It matters more than you might think. Hit record on a lecture, drop the transcript into your AI of choice, and out comes a summary, a set of flashcards, a mind map, and a little podcast of two synthetic voices talking through your syllabus. It feels fantastic. You've got materials. You've got output. You've got a folder that looks like a term's work.
The catch is that you could have generated all of it from the corner of a coffee shop without watching a minute of the lecture. You could have written a couple of scripts and automated the whole pipeline. The AI ingested the lecture. You didn't.
I built my first AI app in 2008. I've led AI at the world's largest tech and entertainment companies for 10 years, and my day job is helping people harness AI to enhance their output and job satisfaction. So take it from me: you have to be deliberate about what you let AI do for you, and what you choose to do yourself. A tool that hands you a perfect set of notes you never think about has created the illusion of productive learning, nothing more.
I built iXnote to help my own continual learning (the AI space moves fast). It doesn't help you look like you're learning. It helps you actually learn.
Here's how that works in practice. Turn FollowAlong on during a one-hour lecture and it won't hand you a 2,000-word transcript of every word said. It gives you about 40 small cards. The key points, stripped right back. They're deliberately thin: a hint, a reminder, a breadcrumb to help you find your way back to a lecture you actually paid attention to.
FollowAlong puts none of the cards in your graph. It hands you a list of suggested cards and then it stops. It won't decide which of the 40 are worth keeping, and it won't draw the connections between them. That's your job, the actual work that proves and supports learning.
Because to connect two cards, you have to understand both and the relationship between them. You can't fake it. Pulling the right eight cards out of the 40 and connecting them the way they actually connect is only possible if you followed the lecture. If you drifted off, you'll sit there with 40 cards and no idea which ones matter. The graph you build is a readout of how well you understood the material.
There's a real worry going around that AI lets students look like they've learned when they haven't. The AI did the writing and the connecting, so the work looks done, but nothing went in and there's no ownership of it. Now think about how you'd actually test whether someone understood a topic. One of the simplest, hardest-to-fake ways is low-tech: write the key facts on cards, put them on the table, and ask the student to pick the ones that matter and explain how they connect. You can't do that on a topic you slept through. Choosing and connecting is the proof.
That is exactly what iXnote asks of you, every lecture. It takes the typing off your hands so you can keep your eyes up and your attention on the room, and it leaves the part that makes you smarter, the choosing and the connecting and the quiet click of understanding, where it belongs. With you.
The other tools will do more of it for you every year. Better summaries, slicker flashcards, more convincing podcasts. The materials will keep improving, and the gap between having the materials and knowing the material will keep widening.
So use AI to study, but choose your tools with care. If you actually want to learn, judge each one by a single test: does it make you do the thinking, or does it do the thinking for you? Pick the ones that keep you in the room.
The exam is where you find out whether you chose well. Nobody wants to open the paper and realise that all the recording and summarising and endless tidying of notes never turned into anything they actually know.
Do the thinking while it counts, and you walk in ready.