One of my favourite actors is Jonathan Pryce. During a long career in film and stage he has played the IRA leader in Ronin and the Bond villain in Tomorrow Never Dies, and his name on the bill is usually a good indicator of something with a little style. I first noticed him in Brazil by Terry Gilliam, the indefinable (at least in Hollywood money-men terms) Python genius.

Brazil is set in an Orwellian future of control and information, with wonderful twists such as when the police are interrogating you, they charge for the service — “Talk soon sonny or this could seriously affect your credit rating” – an idea seen as funny in the eighties but somehow almost possible now.

However, the beauty of the film is in the information processing devices used by this super-bureaucracy: semi-mechanical computers made of 1930s components and magnifying glasses. Obviously, this is a view of the future with the knowledge of the reality of the 1980s.

Then it occurred to me that you see very little depiction of the IT industry in pre-1950s fiction.

You see airplanes, robots, personal submarines, but not IT. Take Fritz Lang’s 1920s masterpiece Metropolis: predictions come every few frames but there’s nothing resembling IT. In fact, routing is literally done by humans moving dials on a big clock face. Why did the futurologists of the time miss a whole industry?

This is a really fascinating question because without understanding it, our own predictions of technology seem rather random.

Thomas J. Watson, boss of IBM from 1914 to 1952, is reportedly quoted as saying there might be a worldwide market for five computers — and perhaps therein lies the reason why the era of ubiquitous computing was missed.

In no other area in mankind’s history have we seen sustained exponential performance to cost reductions. While it might have been possible to see that computers had some use — five is not zero in practical terms — take-up was ruled out by cost.

Moore’s Law is rather special.

I fell into the same trap when, in 1996, a technically unqualified acquaintance contacted me with a radical idea — to sell music over the internet.

I quickly showed him why this was crazy: a 33kbps phone line would take a few lifetimes to deliver an album connection, and then you had the phone bill... well, you could have had the bands appear live for less and where oh where were you going to store all that data with just 600Mb available per CD.

The safe bet, of course, was that bandwidth, storage and the processing power needed to allow fancy compression would all progress exponentially. How, no-one knew, but somehow it would.

Perhaps the other effect is that ubiquity breeds ubiquity.

When a focus group was asked about the early mobile phone, it came to the stunning conclusion that it might be useful for the odd sales rep on the road. It was hard for them to conceive of a world where phones could take on another use, or that technology would advance so quickly that it would become cheap enough for us to chat on them.

The ubiquity effect has also driven IT.

Not only do we need computers because other people have them, but computers now talk to computers: as a CIO, it’s clear that half the IT industry does nothing but serve itself.

There are now predictions of there being more devices on the internet than there are people on the planet.

Followers of this column may know I am fascinated by augmented reality, the insertion of virtual objects into views of the real world.

At the moment, we can do simple things but I am betting on Moore’s Law that the storage on the devices, their processing power and the connectivity to them will continue to increase as their costs fall. And the worlds these augmentations create will only become more compelling and more real.

The data network folks tell me the network can’t handle all this data, the processor people say devices will get too hot, and I’m told memory prices will stay high as yields can’t be delivered, but I won’t make the same mistake as in 1996.

I will just click my heels, think of Kansas and trust the magic of Moore’s Law.

Understanding future outcomes is tricky but there is one final twist in Brazil that shows just how tricky. The last scene shows our hero, Jonathan Pryce, being tortured by Michael Palin.

He breaks free, escapes, gets the girl and they drive off to happiness in the sunset — then the last shot is his still, dead body in the torturer’s chair. His mind was what could not be restrained by the system and so he escaped in spirit, finally defeating it.

The End? Not quite. Much to Gilliam’s rage, the final scene showing the escape to be virtual and the transcendence of the mind over the physical was cut for the US market. In America, he actually escapes.

When looking forward, even if you could see the ending, will you know what it means?

Mike Lynch is the founder and CEO of UK software company Autonomy