I’m deeply skeptical of most commercial claims about AI “personalizing” education—at least in the short term. My doubt isn’t about the underlying technology, but about how poorly it’s likey to be executed. (I ought to write another post on this, because it also shows great promise)

But there is room for real-world wins in knowledge and education.

One of the most useful tools I’ve built is a system that turns any article or website into a podcast episode with high-quality voice narration. I use it daily—catching up on news while driving or listening to long-form Tolkien essays before bed.

Today, I showed Gavrel how to use it.

Gav devours audio. He’s only 14 and has already logged over 13 months on Audible. We can’t get him books fast enough.

He’s not a fan of reading text—but he loves to learn.
So today, I taught him how to ask AI to generate a custom article and add it to his podcast feed.

Prompt (I do not know what these things are):

Please write a story on the v1v2 mouse-tank (mauz?) and the rat-tank landship concept.

Give me a long story article that is very detailed, uses trustworthy and historical information that you can find on the web and is optmized for a podcast, then add it to my podcast feed.

Result several minues later: a 12-minute podcast on… whatever those WWII German weapons were.

This is a small win at home, a way for my 14-year old to generate articles tailored to him in the way that he learns best. Not without risk, the AI might get some things wrong, but because its pulling from general web knowledge it will be roughly as accurate as what he could Google himself, but in a fraction of the amount of time.

There are so many ways that use-cases like this can be beneficial.

The real catch is this: for it to be tailored, it also can’t be a mass-market product, that isn’t how customization works…

P.S. Listen to more examples:

On Gondolin via Tolkien Gateway

On Larry the Chief Mouser:

My morning report, dynamically generated for me each day: