Earlier this week at WWDC, I finally got a chance to sit down and actually use iPadOS on an iPad Pro. The purpose of the meeting was ostensibly to walk me through the new features, but I greedily grabbed the iPad and started tapping, swiping, and opening web pages — listening and talking about what was new all the while. Turns out that, just like the new OS on the iPad, I can multitask better than I expected.

Jamming through the new features on iPadOS was like a greatest hits album of iPad Pro complaints that have been resolved. External USB drives, direct access to drives for Lightroom, and a desktop-class browser all felt like direct responses to our iPad Pro review last year.

Other annoying things have also been fixed, too. For example, you can finally (finally!) directly change Wi-Fi networks and Bluetooth devices right from the appropriate buttons in Control Center. Apple is also finally bringing quick actions and pop-up web previews to the iPad — thanks in part to its decision to basically merge 3D touch and long-presses.

But as I mentioned when I wrote about how Google Docs worked surprisingly well in Safari, I walked away from my brief time with iPadOS unsure exactly how much Apple has really changed. Did it simply fix all the things people have complained about for years, one-by-one? Or is Apple trying to architect a more fundamental overhaul that would make the iPad feel much less constrained by design?

You can certainly argue it both ways, based on what we’ve seen so far. For example, Apple completely changed how the USB stack works in iOS, ensuring it’s properly sandboxed in user space, while building in support for several different USB protocols. It should be fairly trivial for Apple to expand that in powerful ways going forward.

I’m similarly pleased with the much-simplified cursor control system on the iPad for text editing. One clever thing you might not have seen: if you move your finger really quickly outside the text when moving the cursor, it gets gigantified so you can see where it is and makes it easier to quickly move it to a different paragraph.

I also see a lot of effort going into the iPad’s new windowing system. Yes, the basics of how you arrange windows are the same: split screen and slideover. But the ability to fan out those slideover apps or swipe through them turns a sort-of-nice feature into what amounts to having a spare iPhone sitting on the right hand side of your screen. And since apps can spawn multiple windows, Apple added the ability to get an overview of each app’s open windows quickly, which creates a new kind of multitasking workflow.

Of course, all these windowing tricks are only good if apps support them — and right now, many key iPad apps do not (looking at you, Google). But Apple tells me that sometime next year it will change the rules so that supporting the iPadOS window system will be mandatory for all iPad apps. I’m glad.

But ever since Monday, I’ve kept coming back to the same idea: the new gestures that Apple added are really great but feel a little tacked-on. You can use three fingers to do a bunch of “manipulation” stuff: pinch and spread for copy and paste, swipe for undo or redo. They make sense when you learn them, but you really do have to learn them in a way that wasn’t necessary before.

I don’t think they’re very discoverable. You’d have to do a lot of weird experimentation with multiple fingers to find these new gestures on your own, although Apple tells me it will offer onboarding tutorials for some of them. The company famously made little tutorial videos when it launched the iPhone, but this time around these new power-user gestures are much more complicated and much less easy to figure out on your own.

What I’m getting to is these three-finger gestures are unintuitive, but when I say “unintuitive” I mean something completely different from the common understanding of it. I think that when most people hear “intuitive,” they think “innately understood by humans with no training necessary.” I think that’s wrong, at least when it comes to user interfaces.

I don’t think any user interface — whether it’s a computer or a bicycle — is the sort of thing that humans just innately understand. Nearly everything we do requires training and learning. The difference between an intuitive interface and an unintuitive one is how that learning happens.

With intuitive interfaces, you don’t notice that the learning is happening. One skill flows naturally into the next, more complex skill on a relatively easy learning curve. Take the classic desktop interface: if you step back and look, it’s actually deeply weird! It only feels normal because it’s been around for 35 years. However, it is intuitive: you learn left click, then discover right click, then see keyboard shortcuts listed. Each skill leads somewhat naturally to the next, and there are little hints that these extra tools exist all over the interface, inviting you to try them out whenever you want.

Unintuitive interfaces require training: like classes or little tutorial videos showing you how to do stuff that you’ll hopefully remember later. As I note in the video above, different types of languages are a good example of intuitive / unintuitive divide. Math as a language is unintuitive for most people: you need to take classes to figure it out. But natural languages like the one you speak are intuitive — at least for children: they learn them without even noticing it’s happening just because their parents talk to them.

That doesn’t mean that unintuitive interfaces are inherently bad, either. I want to be super clear about that. If an interface requires you to do a little training but feels internally consistent, you are more likely to remember how things work down the road. And it seems as though the iPadOS user interface does have some consistency, though it’s not perfect yet:

• One finger: open or make a window with the thing you’re tapping or dragging
• Two fingers: select stuff (right now this only works in lists I believe? Apple should expand that)
• Three fingers: act the thing I’ve selected, or undo my last act
• Four fingers: Go home

This seems like it has the potential to be a durable, flexible, and powerful user interface system, one that can work really well for people picking up an iPad for the first time and for power users who are replacing their laptops with iPads. But the real trick is having the interface teach people to go from the first group to the second group without having to focus on teaching themselves how iPadOS works.

Maybe I’m wrong, and when we see more people using it beyond the developers who are brave enough to install the first beta, these advanced gestures will come really naturally. If they don’t, that doesn’t mean the new gestures are bad — but it might mean they’re not very intuitive.