Gradually, Then Suddenly: What Film Photography Teaches Us About Every Digital Revolution

For more than twenty years, digital photography was a joke. The first digital camera, built in a lab in the mid-seventies, produced a blurry black and white picture and took a long time to save it. For years after that, digital images stayed grainy, expensive and slow. Professionals looked at them and saw a toy. They were right. And then, in a few short years at the start of the 2000s, film was gone for almost everyone.

I have used this story as an analogy for a long time, because I keep seeing the same shape in other fields. Hemingway has a character who is asked how he went bankrupt, and he answers: gradually, then suddenly. I cannot find a better description of a digital revolution.

The shape of a revolution

Look closely and the pattern has four stages.

1. Worse. The new technology arrives and it is worse than what exists. More expensive, less reliable, less beautiful. Experts dismiss it, with good reasons.
2. Better, for a few. It improves quietly. Enthusiasts adopt it, not because it is better, but because it is new, or more convenient for one specific use.
3. Good enough, and much cheaper. This is the turning point. The new way does not need to beat the old one. It only needs to be good enough, and far cheaper or easier. Then the curve goes vertical.
4. The old way becomes a craft. The old technology does not die. It shrinks into a niche, and often turns into a luxury.

Drawn on paper, this is an S-curve: a long flat line, a sudden climb, then a new plateau. Understanding the curve is easy. Knowing where you stand on it is not. From inside the flat part, the vertical part always looks like it will never come.

Six domains, one curve

Once you see the shape, it is everywhere. What changes is where each domain sits today.

• Photography. The complete story. Decades of slow progress, a few years of collapse, then a second revolution when the camera moved into the phone.
• Music. Records, tapes, CDs, then files, then streaming. Anyone can now produce an album in a bedroom. And vinyl came back, not as the norm, but as a pleasure people choose and pay for.
• Video. Filming used to need expensive equipment and a crew. The phone made everyone a filmmaker. The curve is almost complete, and AI is already starting the next one.
• Commerce. Shops, then online stores, then marketplaces. The next step is buying through AI agents that compare and order for us. This one is still in the middle of its curve, and it is the one I watch most closely.
• Translation. For decades, machine translation was clumsy, literal and often absurd. Then, in a few years, it became good enough for most everyday uses. Gradually, then suddenly, almost word for word.
• Coding. Writing software used to take years of training. AI assistants started as a curiosity and now write a real share of the code in many teams. This curve is going vertical right now, faster than any of the others.

Not every flat line takes off

It would be too easy to say that everything follows this path. Self-driving cars have been five years away for a very long time. Online courses were supposed to replace universities, and they did not. Virtual reality has had several waves of excitement without ever reaching its vertical moment.

A long flat line is not a promise. The climb only starts when three things happen together: the new way is good enough, it is much cheaper, and ordinary people find it easy to use. Miss one of the three and the line can stay flat for decades.

The curve is getting steeper

What strikes me most is not the shape. It is the speed. Photography took decades to reach its vertical moment. Music took less. Maps on phones replaced paper maps and dedicated navigation devices in a handful of years. Translation and coding are moving faster still.

I think the reason is simple. Each revolution is built on top of the previous ones. Digital photography needed new sensors, new screens and new habits. AI arrives in a world where everyone already has a computer, a connection and the habit of using software every day. The ground is prepared, so the climb starts sooner and goes up faster.

For anyone leading a team or a business, this changes a lot. The time between "this is a toy" and "this is how everyone works" is shrinking. The comfortable years of watching from a distance are becoming very short.

What we gain, and what we lose

In French we have a word for what these revolutions bring: démocratisation. A domain that was reserved for experts opens to everyone. Costs collapse. Millions of people can suddenly take a good photo, record a song, sell online or build an app. New markets appear that nobody could have planned. This is the part I celebrate.

But there is a price. When anyone can produce anything, production gets industrialised. Mass output replaces the craft, and the human touch gets thinner. The patient expertise of the printer, the sound engineer or the translator loses its place in the everyday economy.

And yet the craft does not disappear. It changes status. Film photography is alive again, practised by people who choose it for what it gives them. Vinyl is back. Handmade things are valued precisely because they are rare. When anything can be made in seconds, what was made slowly, by someone who knows, becomes the luxury.

Expertise follows the same path. It moves away from knowing how to do the thing, which machines take over, towards knowing what is good: taste, judgment, the ability to tell a great result from a mediocre one. So far, no curve has managed to automate that part.

Where are you on the curve?

When I look at my own field, or at any team I talk to, one question keeps being useful: where are we on the curve? If the new way is still clearly worse, there is time to learn without rushing. If it has just become good enough and cheap, the vertical part is close, and waiting is the most expensive choice. And if the climb is already behind us, the only question left is what we keep doing by hand, on purpose, because that is where the value now lives.

Gradually, then suddenly. The gradual part gives us time. The sudden part never sends a warning.