Edition 371 | August 10, 2026 The Dyslexic AI Newsletter by LM Lab AI

Reading Time: 6 minutes | Listening Time: 8 minutes

I've been saying "lifelong learning" for years. In newsletters, on calls, in conversations about what AI does to education.

Today I heard a different term. Continual learning.

At first it sounded like the same idea wearing a different shirt. It isn't.

Lifelong learning describes how long we learn.

Continual learning describes how we live.

That distinction is small on paper and enormous in practice. Lifelong learning is a duration. Continual learning is a default state.

One says you should keep taking classes. The other says learning is no longer a thing you step out of your life to go do.

For a long time, education ran on a simple sequence.

Learn. Graduate. Get a job. Apply what you learned.

We front-loaded the whole thing. You did most of your learning young. You built your skills. You entered the workforce. Then you spent 30 or 40 years spending down what you already had.

That model was not stupid. It fit the world it was built for. It assumed the knowledge you gathered by 22 would still be worth something at 55.

That is not the world anymore.

Technology changes. Industries change. Jobs change. Entire professions change.

And AI is speeding up every one of those clocks at the same time.

So the sequence probably looks more like this now.

Learn, Apply, Discover, Unlearn, Relearn, Adapt, Repeat.

Look at where the hard part sits. It isn't "learn." We are all fine at learn. It's unlearn.

Nobody ever taught us to unlearn. School never graded it. There was no unit on how to put down something you were good at.

I've coached baseball for 30 years, and every season is a new team. What worked with last year's group fails with this one. You either let go of what worked before or you spend the whole season coaching a team that isn't in front of you.

That step deserves its own edition, and it's getting one. Next time.

The front-loaded model wasn't neutral. It scored everybody, once, during the exact window when almost everything ran through decoding speed.

If reading came slowly to you, you got measured in the one phase that counted, on the one channel you were worst at, and then handed the result to carry.

Plenty of us were told what we couldn't be at nine years old and just believed it.

Here's what changes in a continual model. The scoring moves. What a continual world actually rewards is pattern recognition, reframing, asking sharper questions, and connecting things that don't obviously go together.

That's the list. That's most of the dyslexic thinkers I know.

Here's where this gets interesting to me.

Books gave us access to knowledge.

The internet gave us access to information.

Online education gave us access to courses.

Every one of those was a jump in access to content. More of it, faster, cheaper.

AI is different in kind. AI gives us access to a learning relationship.

Something that can explain one idea five different ways until one lands. Something that challenges your thinking instead of just confirming it. Something that spots the gap you didn't know you had, practices with you, gives you feedback at 11pm on a Tuesday, and remembers where you were last time.

For the first time in my life, a continual learning partner is available whenever we need one.

That changes what it means to be educated. Educated stops meaning "finished absorbing" and starts meaning "able to keep going."

Which brings me to the question we ask kids.

"What do you want to be when you grow up?"

It's a fine question for a world where you pick one thing and ride it for 40 years. It's a strange question for a world where the job may not exist by the time they get there.

Maybe the better one is: "How do you want to keep learning as the world changes?"

I think about this with Makena and Kailer constantly. The most useful thing I can hand them is not a list of what to know by graduation. It's the ability to keep learning, adapt, think, question, experiment, and reinvent themselves when the ground moves.

Because it will move. Repeatedly.

We're at a defining moment here, and not because AI can do the learning for us. It can't, and we shouldn't want it to.

It's because AI may finally make continual learning a normal part of being human instead of a luxury for people with time and money.

And if that's true, then education isn't something we finish.

It's something we learn how to keep doing.

Run the unlearn step on purpose this week, since it's the one nobody schedules.

Pick one thing you believe about your work that was true five years ago, and go check whether it's still true.

Not a small thing. Something you'd defend in a meeting. A rule about how your industry works, how your job gets done, what's worth your time.

Say it out loud to your AI and ask it to argue the other side. Ten minutes. That's the whole exercise.

You'll either come out holding it more firmly or you'll put something down that's been quietly costing you.

"Here's something I believe about my work: [state it plainly]. I learned it about [when]. I want to check whether it's still true. Ask me what evidence I'm actually basing it on. Then tell me what's changed in this field since then that I might be ignoring, and make the strongest case that this belief is now outdated. Be direct with me. Don't soften it. At the end, tell me what I should keep and what I should let go of."

Do this once a month and you've built the unlearn step into your life on purpose.

Matt "Coach" Ivey

Founder, LM Lab AI | Creator, The Dyslexic AI Newsletter

Dictated, not typed. Obviously.

Edition 372: "Once I Learn It, I'm Gone." The unlearn step up close, why I've never been able to stay anywhere once I finished learning it, and what that looks like at almost 50.