The Pinball Machine

What it actually feels like to navigate AI - when navigating uncertainty is your job

Last week I spent two days exploring Opus 4.6 — the latest model from Anthropic. I wasn't testing it for a client or preparing a workshop. I was just working. Using it inside my own business, alongside the other AI systems I now rely on daily.

And at some point on the second day, I had to get up and walk away.

Not because something went wrong. Because something shifted.

It's hard to describe what happens in those moments. It's not that you've learned a new fact, or seen an impressive demo, or read a prediction about the future. It's that your entire frame of reference tilts. Everything you thought you understood about what AI could do — what you could do with it — rearranges. And before you've caught your balance, it tilts again.

Imagine one of those old pinball machines. You're the ball. You think you know the direction you're heading, and then a flipper catches you and flings you somewhere completely different. You recalibrate. You start to make sense of the new trajectory. Then another flipper hits. And another. The pace doesn't let up, and the surface never holds still long enough for you to plan a route.

That's what working with AI actually feels like right now. Not from the outside — not the conference version, where everything is curated and coherent. From the inside, where you're trying to build something real and the tools keep changing what's possible while you're mid-sentence.

———

Here's the part that I find quietly ironic.

My entire professional practice — built on decades of leading through crisis, disruption, and large-scale change — is built around helping people navigate uncertainty. That's the work.

That's what The Uncertainty Lab exists to do

And yet there I was last Tuesday, standing in my kitchen, staring out the window, needing a full ten minutes before I could walk back to my desk. Not because I was confused. Because my brain needed time to catch up with what had just changed — again.

I could have tidied that up. I could have written a newsletter that starts with a calm, authoritative overview of the current AI landscape and what leaders need to know. Something polished. Something that positions me neatly as the person who has this sorted.

But that wouldn't be honest. And if this newsletter is going to be worth your time, it has to start with honesty.

———

So here's what I'm actually experiencing, as someone who works with AI every day, is building a business around it and is genuinely trying to use it well:

The capability is growing faster than anyone can absorb.

Every few weeks, something arrives that doesn't just improve on what came before — it changes the category. What was impossible in November is routine in February. I've watched tools go from impressive to indispensable to insufficient in the space of a quarter.

The wins are real — and, at times, genuinely disorienting.

I've had days where AI saved me four, five, six hours of work. Genuinely. And instead of feeling productive, I felt strangely destabilised. If I can do this in an afternoon, do I need to spend quite so many hours at my desk? What does effort even mean now? The efficiency is wonderful and unsettling in almost equal measure.

You think you know what you're doing, and then you don't.

This is the one that gets me time and again. Just when you've built a workflow, developed a rhythm, started to feel competent in a space — something shifts and the ground moves again. Not catastrophically. Just enough that yesterday's confidence doesn't quite transfer to today.

———

I've been thinking about why this is so tiring. And I don't think it's the volume of change, exactly. I think it's the frequency of reorientation.

In most of our working lives, we've been able to learn something, build confidence in it, and then operate from that base for a reasonable stretch of time. The learning was the hard part.

Once you'd done it, you had stable ground to stand on.

That cycle has compressed to the point where it barely completes before it starts again. You orient yourself. You start to act. The conditions shift. You reorient - and act again. But then the conditions shift again. There's no resting point where you can say: right, I've got this now.

That's not a temporary disruption. That's a different operating condition altogether.

And it asks something quite specific of us. Not more knowledge. Not more speed. But a different relationship with “not-knowing” — one where you can keep moving without needing the ground to hold still first.

This is what I call agency under uncertainty – and its what I am passionate about. The capacity to act with intent and integrity even when you can't see the full picture. It's not confidence exactly — it's something steadier and more honest than that. It's the willingness to make a call, head towards an imperfect destination, knowing you'll probably need to revise it, and being okay with that.

———

I've spent most of my career inside uncertainty. Leading through the Black Summer bushfires. Running crisis response for 2 years during COVID. Navigating large-scale organisational change in sectors being disrupted and reshaped. And now, building a new practice from scratch in a landscape that reinvents itself every few months.

The specifics change. The condition doesn't.

What I've learned — and what I'm still learning, honestly, every week — is that there are ways of thinking and working that help you stay oriented when everything around you is moving. They're not complicated. But they're different from what most of us were trained to do.

This newsletter is where I want to share that. Not as someone who's figured it out, but as someone who's doing the work in real time and is willing to think out loud about what I'm finding out.

Some of what shows up in here will be reflections on my own experience — what's working, what isn't, what surprised me. Some will be observations about patterns I'm noticing across the leaders and professionals I work with. Some will be re-frames — ways of seeing familiar situations differently that might help you move when you feel stuck.

What it won't be is a data briefing, an AI news roundup, or a set of tips. There are good sources for those. What I'm trying to do here is something else: help build shared language and understanding for an experience that most people are having but very few are talking about honestly.

———

The question I keep coming back to — the one underneath everything I'm working on — is this:

How do you keep leading yourself forward when the ground underneath you keeps tilting?

I don't have a tidy answer. But I'm getting better at sitting with the question. And I think that might be more useful than most of the answers on offer right now.

If you're sitting with it too, I'm glad you're here.

— Sue

P.S. If you'd like this newsletter in your inbox every fortnight — subscribe here

Inside Uncertainty Leading, deciding, and staying human — when certainty doesn't arrive first is a fortnightly note on navigating AI, leading through uncertainty and what it takes to keep moving when the ground won't hold still.