Not Everything Difficult Needs a Shortcut

This past weekend I was watching The AI Doc: Or How I Became an Apocaloptimist on Netflix, a documentary exploring what AI could mean for humanity and the world we're building around it.

It left me thinking about something much simpler.

Friction.

We spend an enormous amount of time trying to remove it from our lives.

Whether it's waiting. Searching. Getting lost. Writing the terrible first draft. Sitting with a problem we don’t know how to solve.

And now AI can remove some of that friction from thinking itself.

Mostly, that’s exciting.

But I’ve been wondering whether all friction is something we should want to eliminate.

Think about almost anything humans have invented.

Someone, somewhere, encountered a problem.

Something was too slow. Too difficult. Too dangerous. Too far away.

Something didn’t work.

And someone had to wrestle with that problem long enough to imagine something that didn’t exist yet.

The Wright brothers spent four years researching, testing and refining before achieving controlled powered flight.

James Dyson went through 5,127 prototypes developing his bagless vacuum.

At 3M, Spencer Silver developed an adhesive that wasn’t nearly as strong as intended. For years, there wasn’t an obvious use for it. Eventually, that apparent failure helped become the Post-it Note.

Human progress is full of people encountering friction and asking:

“There has to be a better way.”

Of course, the point of invention is often to remove friction.

But I’ve started wondering about what happened before the solution.

Experimenting.

Failing.

Questioning.

Starting again.

Not knowing.

Thinking.

Perhaps the friction they removed from the world was only possible because of the friction they were willing to endure themselves.

Which makes our current moment interesting.

We’re building technology that can remove friction not only from what we do, but from parts of the thinking that happen before we do it.

I use AI regularly. But I’ve become increasingly deliberate about how I use it.

An idea usually starts with me.

I’ll write it down. Sit with it. Come back to it. Pull it apart. Ask myself whether there’s actually something worth saying.

Then I research.

I read. I look for evidence. I question what I think I know. Sometimes what I find changes the idea entirely.

Then AI enters the process.

I’ll use it to challenge the idea, extrapolate, organise what I’m thinking, explore another perspective or help me see something I’ve missed.

Sometimes that sends me back to the research.

Sometimes it sends me back to the original idea.

Then I come back.

I refine.

I remove.

I rewrite.

And ultimately, I ask whether the final piece still contains the key points and value I wanted to communicate in the first place.

The process isn't always linear. New information can send me back to the research, or back to the idea itself.

Used that way, AI doesn’t replace my thinking.

It helps me stay inside it.

And perhaps that’s part of what we need to figure out as these tools become increasingly capable.

Not simply:

What can AI do for us?

But:

What parts of being human should we be careful not to outsource?

Because some friction is inefficient.

Some of it probably should disappear.

But some friction makes us curious.

Some forces us to think.

Some makes us question what we believe.

Some makes us sit with uncertainty instead of immediately escaping it.

And some makes us wrestle with something long enough to actually understand it.

Maybe that’s the distinction that matters.

Not whether technology makes things easier.

But whether, in making everything easier, we accidentally remove some of the things that were making us better.

Perhaps progress is about becoming better at recognising which friction is pointless…

and which friction is quietly making us better.

Not everything difficult needs a shortcut.

Oliver

Mental Performance