AI and the speed of change

I’ve closely tracked the impact of technology on media for more than 30 years now.

The impact of AI is of a completely different character. It is orders of magnitude faster. It moves so quickly that you cannot talk about it as if it were one trend — it is 100 trends wrapped up in one. And it is affecting everything.

Transformations and revolutions

There are two types of technology-driven shifts in the world: transformations and revolutions.

First transformations. As a technology analyst for two decades, I accumulated a great deal of experience with industry transformations.

Imagine for a moment that, like me, you were tracking the impact of social networks on media in 2008, the impact of mobile on software in 2013, or the impact of streaming on television in 2015.

You could see that something big was happening. You could find early adopters, and identify the beginnings of major shifts. You could even predict the transformation of a whole industry. Some elements, like the exact shape of that transformation and the speed with which it would happen, were still fuzzy, but it became clear that something big was happening (for example, mobile becoming the dominant way people accessed web sites, or streaming displacing scheduled television).

These changes took five to ten years to go from early glimmerings to full-blown accelerating trend. That’s very fast for the transformation of a whole industry.

Revolutions are different. Look a little further back.

Imagine that you were attempting to predict the impact of microprocessors in 1971, or the impact of the internet and the Web in 1994. In those cases, with the benefit of hindsight, you can see that the transformation affected every industry and changed nearly everything about how companies and consumers behaved.

Those changes took decades to have their full impact, and continue to reverberate. They were broad, but not rapid. The full shape of a revolution like microprocessors or the internet is impossible to predict, but you can tell it is going to be big, and you have years to track it.

AI is a whole new category of change: fast, broad, and profound

It’s clear to me that AI is more like a revolution than a transformation. But it’s a revolution that is developing faster than even the transformations I tracked.

AI goes through generations of change in months.

It profoundly affects media, marketing, medicine, manufacturing, and management. And that’s just the M’s.

It’s already tearing through the workplace, causing job loss. It’s already upending media creation by making it easy to generate synthetic media, whether that’s text or audio or video. It’s already making entire coding functions vastly different from what they were only months ago. It’s obsoleting job security, IP protections, and data analysis. Transformations, while less profound, still tend to move too fast for industry regulation, which takes years to catch up. Government is hapless at attempting to manage technology revolutions.

There are problems, of course.

Whatever you can say about AI’s consumption of energy, that will change. Already, powerful models are running on personal computers.

Whatever you can say about AI’s wholesale ripoff of intellectual property, that will change. Mass licensing is inevitable.

Whatever you can say about attempting to prevent AI use in college, that will change. Professors are recognizing that AI skill is more important than their narrow ideas about preventing cheating.

Whatever you can say about AI hallucinations, that will change. Models are checking other models. Truth is impossible to define, but AI is becoming more dependable.

And don’t get me started about em-dashes. Do you really think that “tell” will remain three generations from now — in a month or two?

What this means

So AI is a revolution moving at unprecedented speed with unprecedented impact.

Here’s what this means.

Predictions are going to be laughably wrong very soon. That includes predictions about job losses and where the financial value is coming from. No one can reasonably predict where this revolution is going, including the companies like OpenAI that are creating it.

Tactical moves will therefore occupy decision-makers. They’ll be planning for what happens in the next six months or a year. They’ll have to be agile, because their five-year plan is going to be obsolete before the end of the first year.

There will be new centibillionaires. I don’t know who. There will be major acquisitions and bankruptcies. I don’t know which.

But if you not a corporate strategist, there are things you can do.

Keep track of developments in your industry. Try to maintain a longer view, rather than overinterpreting what’s happening moment to moment.

Beware confident predictions.

And develop your skills. Try the tools. Learn to use them, and keep learning, because they will keep changing.

Because the only thing I’m sure of is that those who are ignorant of these tools will soon face obsolescence.