Fly the Aircraft First

Somewhere over the paddocks north of Melbourne, my flight instructor reached across and pulled the throttle to idle.

"You've just lost your engine. Where are we going?"

I was maybe fifteen hours into learning to fly, and I had a plan. It was a good plan. It had a heading and a height and a checklist, and it was now completely useless, because the plan assumed an engine. What I had instead was a glider with a very bad glide ratio, a couple of minutes of altitude, and an instructor who was not going to help.

That moment did as much for my confidence as it did for my nerves (to say it was uncomfortable is an understatement. Not convinced I did a smashing job, but I must say, I did land the plane). What I learned was that plans are one thing, but the ability to constantly assess the situation in an ongoing loop is another. This vicious loop keeps the bars in corporate cities across the globe in business, and has been well studied in the aviation and business worlds.

The loop

The loop has a name, taken from fighter pilot John Boyd, who flew F-86s in Korea and spent the rest of his career trying to understand why American pilots were winning dogfights against a MiG that could out-climb and out-turn them. His answer was not a better aircraft but a faster one at changing its mind: the Sabre's hydraulic controls let it flick between manoeuvres quicker than the MiG's cables could.

Boyd's translation of that quicker ability to go from input to output was never that you should think harder than the other guy. It was that you should move through the loop faster, so that by the time they react to what you've done, you've already taken the next step. Out-cycle, don't out-think.

Business borrowed OODA decades ago and mostly treated it like a poster. But as we head into 2026, I think it finally means what Boyd intended. Strategy is no longer a plan you defend; it is a loop you run faster than the market. Every surprise becomes data for the next cycle, not evidence that the previous one failed.

The magenta line

In 1997, Warren Vanderburgh, an American Airlines training captain, stood before a room of pilots and gave a talk that is still passed around in flying circles today. He called the new generation "children of the magenta line", named after the coloured track the flight computer draws on the navigation display. His concern was straightforward: the automation was so effective that pilots were beginning to manage the computer rather than fly the aircraft. He worried that when the automation quietly handed control back, some had lost their instinctual feel for it. As bonkers as it sounds, you can feel your aircraft when flying it. You know when it is operating normally, you know when something tiny is off, you can just 'feel' it and then you go to diagnose what the feeling scientifically is.

Now, Captain Vanderburgh was not anti-automation. In fact, the magenta line and countless other automations have made flying safer than it has ever been in history, with you statistically more likely to die from a ladder falling on you than on a commercial flight (fun fact: IATA reckons you would need to fly every day for about 49,000 years to be in a fatal accident). His point was instead about levels. When things get confusing, step down a level of automation, not up. Fly the aircraft first, then work out what the computer thought it was doing.

For most of my career, confidence in a decision was something you could 'feel' because you earned it on the way to it. You sat with the numbers, you argued it round the table of smart humans, you slept on it, you rewrote the deck a billion times, and by the time you said "this is what we're doing" the conviction had been built up in the working it out. Now the answer arrives first, either from an expo, a competitor, or from AI. The working out, the part that used to make something yours, is the part that more often than not now gets skipped.

That is not a crisis of meaning. It is a magenta line problem. The machine drew the track, the track looks right, and somewhere in the last eighteen months you stopped being sure you could feel the aircraft underneath you.

MIT studied this (as they seem to do for everything, or is it just my research bias leaning to them because they have a famous name? A debate for another time). Anyway, MIT's Media Lab wired up 54 people with EEG caps and had them write essays, some with ChatGPT, some with a search engine, some with nothing but their heads. The ChatGPT group showed the weakest brain connectivity, the lowest sense of ownership over what they'd written, and 83 per cent of them could not quote a line from an essay they had finished minutes earlier. Small study, but the finding was that the aid does not just help you do the task, it does the task, and the muscle that used to do it dwindles away. Then we have the ambient noise, things like executives being, on average, in 23 hours of meetings a week and rating more than half of them as low value, and digital interruptions and notifications constantly stealing our consciousness and therefore interrupting intelligent thought. Combine those, and you get a cockpit where the autopilot is superb, and the pilot is tired, distracted, and slightly unsure who or what is actually flying.

Herbert Simon saw this coming in 1971, or at least guessed, and was maybe correct. "A wealth of information creates a poverty of attention." Either way, he was right. I wonder if he ever thought of a day where the information would one day write back to us with an attitude and personality.

Unusual attitudes

There is an exercise in flight training I genuinely hated experiencing but loved forcing cadets to do, and it is famously referred to as the most important thing you have to learn when learning how to fly. The instructor takes control and tells you to close your eyes (or worse, puts a stupid hood on you that blinds your vision of everything other than instruments). They fly around for a bit, roll you into something ugly, a steep turn with the nose dropping, or a climb bleeding off speed and about to stall, and then say "your aircraft". You open your eyes and you have a few seconds to read the instruments, understand what the aircraft is doing, and recover before it does something you cannot recover from and everybody dies (no pressure).

The exercise is called unusual attitude recovery, and what it teaches you is not the recovery. It is the reading of the situation. The whole point is that your body is lying to you (your inner ear is convinced you are flying straight and level when you are in a 60 degree bank and falling), the instruments are telling the truth, and you have to be able to interpret the truth quickly, unaided, with your heart trying to remind you of the fragility of its existence.

The hard thing to teach is that this isn't a "trust your gut" kind of situation. Pilots who trust their gut over their instruments in these situations fail the test. Judgement in the cockpit is not instinct. It is the trained ability to look at the data, know what it means, and know what to do about it, on your own, in the time available.

That, I think, is the skill the confidence crisis is really about. Not whether you use the aids. Of course you use the aids, but whether you can still read the panel when the aids fail or are shut down by government or a risk committee.

Where the loop breaks in an AI world

AI is extraordinary at Observe. It sees more than you ever could. It is getting very good at Decide, in the sense that it will hand you an answer with a confidence that is mostly unrelated to whether the answer is right. And Act is yours and, to some extent, probably always will be.

The step it cannot do for you is Orient. Orient is where the observation meets your experience, your context, your understanding of who this is for and what has to be true. Orient is the bit Boyd said mattered most, and it is the bit that we skip first when the machine draws the line. The problem is, when every competitor upgrades their strategic tooling at once thanks to new tools, relative position doesn't move; you're running just to stay in place. The real race shifts to whatever isn't equally available to everyone: your data, your speed, your judgement. In other words, how you Orient.

The leaders pulling ahead are not the ones using AI fastest. They are the ones who can still tell you why, unaided. They fly the aircraft first.

An experiment, not a framework

I trained to be a pilot, hold a commercial licence, and still fly, but career-wise I happily fell into retail instead. But since I'm talking planes, let me squeeze the analogy of all its plausible juice. There is a well-known adage in flying circles: "There are three secrets to a great landing. Unfortunately no one knows what they are." Instead, you just have to treat each landing as its own event and do your best.

So, in an ugly attempt to land this article, here is what I'd suggest you try next. Before your next significant decision this week, pull the throttle. Close the laptop. No summary, no draft, no chat window. Take a blank page and write, in your own words, what you observe, what you think it means, and what you would do. Draw it if the words are hard; sometimes triggering our creative brain can help us orient around a problem or data better. Or, if it is a bigger decision, walk on it. Walking triggers so many parts of the brain, and the air helps us think.

Then open the tools and compare notes.

If the machine's answer and yours agree, good, you have a decision, and you have the added bonus of knowing why. If they disagree, even better, because now you have something to orient around. And if you find you cannot fill the page at all, that is not a failure. That is the instrument reading you needed. It tells you which muscle wasn't firing, and muscles come back with use.

The point is to stay the kind of pilot who could fly the leg without any of the instruments, but use the instruments without being blinded by them.

Your aircraft.