The endless scroll
[No, I am not talking about my blog post 😑]
Wow, look at us. We have never had more ways to learn! There are courses and playlists on every platform, videos, books, articles on every topic, and now an AI that will politely explain or role play anything you need. I'm a massive believer in access to learning, so I'm delighted that we've never had more.
But there’s a gap between how much content exists, what and how it's surfaced and how much learning actually happens. I think it’s the most important problem in learning right now. And I worry it's being misread.
A former CEO of mine crystallised this for us in two use cases - just-in-case, and just-in-time learning. Part of the problem is that a lot of content is still built and delivered as just-in-case learning: stockpiling knowledge we might need one day. It's how we were taught to learn, but it doesn't meet every use case. A lot of the time we don't need a whole course or curriculum. We want a quick answer, right now, to the problem in front of us. That's just-in-time learning, and the shift from one to the other changes what good content needs to be.
Two needs
Think of it like this. My husband loves poached eggs. Heck, I love poached eggs. I know from various Saturday morning cooking shows - vaguely - how to poach one. When I finally crack (ha) and make him one for breakfast (he's been waiting 10+ years), I won't be studying the theory of eggs before I turn on the stove. I won't have to take a multiple choice quiz on "Which bird laid this? 🐣" before I pull out the egg carton. I won't need to contemplate and deliver a thesis on whether the chicken came first before I fill a pot with water. What I will need is quick and clear access to a few steps from a source I trust, while the water comes to the boil. That, my friends, is just-in-time learning.
Now compare that with wanting to get qualified in product operations or data, really anything to change or elevate your career. There's no five-step guide here. You need the foundations first: how products get planned, built and launched, how information flows between teams, and what good looks like when it all works.
You learn it before you need it so that when the moment arrives, you recognise it. Compliance training works the same way. Nobody sits through their data protection module hoping to use it that afternoon (let's be honest, it can feel like all hope is gone). But when a dodgy email lands or a customer asks what happens to their data, you know exactly what you're looking at. That's just-in-case learning, the kind that changes how you work, think, and helps you ask the right questions.
Technology meets both
Here's the exciting part. For most of history, formal learning came in one shape: the course. Whether you needed five steps or five months, it was likely delivered in the same medium and you had to fit yourself around it.
Technology can now meet learners where they are. Ask a question in the flow of your work and get an answer in seconds. Or simply ask it to teach you something new, at your pace and your level. That's a real advance, and I'm really all for it. But it raises a question that's easy to skip past: where do those answers and paths actually come from?
More does not equal better
We can open a chat and type "teach me how to…". I get it. AI is patient, it's there at your fingertips at 11:21pm, it pitches things at your level and it never makes you feel daft for asking twice. "That's a great question, Lakhi! You're so smart! Let me put it another way..."
But an answer is only as good as what's behind it. AI often won't show you where an answer came from and how it was sythesised, who the expert was, or whether it's still current. And it answers the question you asked, not the ones you didn't know to ask. (Which is that just-in-case, transformative kind of learning).
The same goes for AI-generated courses. They're cheaper to produce, sure, but the cost hasn't disappeared, it's moved. An expert, one hopes, still comes in and rips them apart.
This isn't a new problem, either. We had it with search engines long before AI. Google a question and you get a few different top answers from ten sources. (I am putting aside my THOUGHTS on Google's zealous use of AI in search). On the face of it, how do you know which link is credible? Who paid for, or optimised, the wrong answer? Most of us pick the top result, or the one that sounds legit. AI simply removes the list and hands you one answer. The question of whether to trust it hasn't gone away - it's just harder to see.
There are millions of courses out there, believe me. So what actually helps you choose the right one? There are signals we can use: reviews and ratings tell you what people enjoyed. Enrolment numbers tell you what was popular. Duration tells you how dense it might be. But none of these things tells you whether it's accurate, current, or right for where you are. The signals that matter are harder to see: who made it, whether someone with real expertise has vetted it, whether it's current and whether it fits the moment or need you're in.
The point isn't to swing the other way, to being handed one single answer that does all your thinking for you. People learn differently, and differently again depending on what they need it for. Some want a short video, some a checklist, some a book they can sit with. Others want to be taught in person with a group of peers. What helps is a selection: a handful of trusted options, not ten thousand, and not one. Enough choice to find what fits you, few enough that every option is worth your time.
It’s a bit like choosing where you get your news. Most of us have outlets we trust and ones we scroll past. Learning deserves the same choice.
Follow the value
Think about the difference between a warehouse full of books and a good bookshop. The warehouse has more books, sure. The bookshop is where you're more likely to find the one you want.
That's where the value is moving. Not to more, but to three things abundance can't fake:
Trust. Who made this, and do they actually know what they're talking about?
Expertise. Is this the real thing, built by someone who has done the work, or a summary of a summary (of a summary)?
Fit. Is this right for me, and for this moment: five quick steps, or the deep foundations?
Generated content is only ever as good as what it was built on, and how well it connects the dots to related subjects and your context. Strip out the experts and the original work, and the machine has nothing worth remixing. I hear worry in the industry that AI can make good, credible content obsolete. No. Technology makes credible content the scarcest and most valuable ingredient in the whole system.
Of course, great content doesn't simply appear. Someone reviews the market, someone commissions it, finds someone spectacular to write it, someone checks it, and someone gets paid for it. If we treat expertise as free raw material, we'll eventually run out of it.
In publishing and any content creation, "rights" is a dry word. It means contracts, territories, formats and licence terms - how the content is used. But behind every one of those contracts is a simple deal. The expert shares what they know; in return they can keep doing the work that makes them worth learning from.
Break that deal and the good stuff will dry up. Experts stop writing. Specialists stop recording. What's left is generated content, made from other generated content, to fill space rather than to teach. I think we can all agree: nah, thanks.
Selection matters just as much as creation. Something has to stand between an overwhelming amount of content and the person who needs to learn, and say: hey, start here. Increasingly, that’s an AI assistant, and that’s a good thing. But it can only recommend well from a catalogue worth recommending, shaped by people who know what good looks like. Whether the learner wants five steps or five months, that pairing is what makes the technology worth trusting.
Built for both
The good news for experts is that the same body of knowledge can serve the poached egg and the career development. It just has to be built that way, and here's what that looks like:
Modules, not monoliths. Self-contained pieces that each answer one question can sit inside a full programme or stand alone as a quick answer, easily updatable.
Properly described. DATA 🤩 Clear descriptions, tags for skill, level, format, topic and when it was last reviewed are what let the right five minutes be found inside a five-hour course.
More than one shape. The same expertise delivered as a video, a structured course transcript, a checklist or a short read, because learners and moments differ.
Visible provenance. Show who made it and why they know. It's the signal learners struggle to find, so make it easy.
Clear terms for use. Knowing how content can be excerpted and used protects creators and opens new routes to learners.
Experts who build like this don't lose out to technology. They become what the technology depends on.
How it reaches us, what it's made of
Technology changed how learning reaches us. The quick answer when you need it, the deep path when you want it, both shaped around the learner rather than the other way round.
What it hasn't changed is what makes learning worth having. Behind every good quick answer and every good qualification needs to sit expert content: made by someone who knows, checked by someone who cares, and chosen for the person in front of it.
So yes, when I finally poach that egg (if you're reading, don't hold your breath, dear), I'll be reaching for the quick steps. But it has to come from someone who's actually good at poaching them.
