Why the Future of Technology is so Hard to Predict

He also predicted “The Prolongation of Life” — but there, he might be disappointed in us. We’ve made vast progress in preventing people from dying from infections while still young, but have yet to figure out how to get most people to live much past 100.

More recent predictions by futurists haven’t been quite as accurate, perhaps because they rely too much on extending the latest, trendiest technologies into new realms. One of the most famous living futurists, Ray Kurzweil, predicted back in 1999 that by 2019 robots would educate us, conduct business transactions for us, adjudicate political and legal disputes, do our household chores, and have sex with us.

Even someone as brainy as Kurzweil couldn’t have imagined that in late 2022 the main feature in MIT Technology Review would be headlined: “A Roomba recorded a woman on the toilet. How did screenshots end up on Facebook?”

Worse still, the Roomba is still not as good at vacuuming as a diligent human.

Technology writer Edward Tenner is author of, most recently, The Efficiency Paradox, about the limitations of big data and artificial intelligence. We had a long talk about the trouble with predicting the future of technology, and why, today, the future seems extremely late and not exactly what we ordered. He explained that there are three problems with predicting which technologies will change the world.

The first is what he calls a reverse salient — a sort of stubborn bottleneck, which may explain why we still don’t have a universal cure for cancer, we haven’t extended the human lifespan past a little over 100, and — even with a fantastic breakthrough in fusion energy this month — we have made such slow progress on clean energy.

This year’s debut of ChatGPT looks like it might have broken through a barrier to humanlike artificial intelligence, but Tenner said it’s really just hoovering up vast seas of existing information. “It’s sort of a scaled-up plagiarism in which other people’s ideas and writing are sliced and diced and repackaged.”

To illustrate what it’s missing, he asked it to consider the meanings of the phrase “a rolling stone gathers no moss.” It picked the most common Western interpretation of the proverb — that it’s good to keep rolling along in life.

“On the other hand, in the Japanese sense of aesthetics, moss is really beautiful … so you could say that somebody who is footloose and doesn’t really commit to anything — they will not have this natural treasure,” said Tenner. ChatGPT never considered this view.

There are remaining bottlenecks to useful and trustworthy AI, said Tenner. “A lot of AI now is really a black box process where the AI can’t really explain and defend the reasons for a decision.” ChatGPT can be glib and even creative, but we might not want to put it in charge of anything important.

The second problem with predicting the future of technology is that some inventions just don’t beat rival technologies on the market. A great example was a new kind of refrigerator designed in 1926 by Albert Einstein and another physics genius, Leo Szilard. How could an Einstein refrigerator possibly lose? There was a great need for it because refrigerators at the time used toxic gases that sometimes leaked, killing entire families.

The Einstein-Szilard refrigerator used an electromagnetic field and a liquid metal as a compressor, which got rid of the toxic gas problem but apparently created an annoying noise problem. By the 1930s, scientists discovered chlorofluorocarbons, which were stable and safe for households — but, as the world would learn decades later, were building up in the atmosphere and destroying the earth’s protective ozone layer.

Other examples abound, from Thomas Edison’s direct current, which was usurped by alternating currents, to the Segway motorized scooter, which was supposed to change the world, but never really gained traction — despite the popularity today of e-bikes and motorized scooters.

The final problem with predicting the future: Sometimes, social, cultural and psychological factors keep predictions from coming true. For several years after the first sheep was cloned, there were predictions everywhere that cloned people would soon follow. But society doesn’t really like the idea of cloned people.

Similarly, fears of using gene editing to create the “perfect baby” are probably overblown. Even if Crispr technology makes that possible on some level, the perfect baby probably wouldn’t grow up into a perfect adult, said Tenner. We’re not consistent in what we consider perfect — “you can imagine a wave of [engineered] babies … and by the time they grow up, they’d be obsolete,” he explained. Maybe tomorrow’s parents would try to clone Einstein’s brain, only for their baby Einstein to miss the window for revolutionizing physics and invent a brilliant but forgotten refrigerator.

This year, predictions are reflecting the mood of our pandemic times — gloomy. Earlier this month, the New York Post listed technologies that could bring to life a terrifying dystopian future. The first was quantum computers, which could potentially break all current encryption systems and allow everyone’s money to be stolen. Then there was geoengineering — which could either save us from climate change or kill us all — and killer drones.

And last on the list was the same thing Boyle put at the top if his list in the 1600s: Life extension for the super-rich, illustrated with a photo of a giant rat superimposed on Jeff Bezos. I think Boyle would be more intrigued than afraid, though he might also be surprised that one of the richest men in the 21st century hasn’t invested in a head of “new hair color’d as in youth.”

Bloomberg News provided this article. For more articles like this please visit bloomberg.com.

More Life Events Topics >