In March 1919 the British scientist Arthur Eddington travelled to the west African island of Principe.
When he arrived in May he set up his instruments (with other members of his team), waited for a total eclipse of the sun and then measured a curious thing: the extent to which light from a distant star was deflected by the gravitational pull of the sun.
Why did he do this? Why wasn’t he working in his lab, measuring atoms under a microscope, or whatever? The reason is that a few years earlier an eccentric scientist working at the Swiss patent office had conceived of a thought experiment: what would happen if I ran from a clock tower at the speed of light? Albert Einstein (for it was he) concluded the clock tower would appear frozen in time. From this (and other leaps of imagination) he proposed the special theory of relativity and, later, the general theory. It was this which predicted the precise deviation of light caused by gravity — and caused Eddington to go on his journey to Africa to test the hypothesis.
I think this story is of huge importance not just for understanding the nature of science but of capitalism, politics and perhaps even western civilisation. To go back to science, it’s easy to assume that this discipline is about observation, data collection and patient analysis. And there is, of course, a role for those things. But the Einstein- Eddington story shows that innovation often emerges not just from observation but from imagination. This is why the philosopher Karl Popper described scientific hypotheses as “bold and imaginative”, a point corroborated by Einstein, who said: “Imagination is more important than knowledge.
Knowledge is limited. Imagination encircles the world.”
I worry that in an age of big data and mechanical manipulation we have somewhat lost this insight. And this is about more than science. People don’t tend to think of it in these terms, but capitalism at its most thrilling depends on the incessant creation of bold hypotheses. Every entrepreneur who comes up with a business idea is submitting a conjecture to testing. Can I persuade a capitalist to back my plans? Can I persuade consumers to buy my products? The beauty of capitalism is that millions can come up with their own hypotheses, leveraging the intelligence of multitudes, just as science is propelled by the “collective brain” of the scientific community — a point made by Hayek, among others.
One way in which the West has gone wrong, I think, is to have presided over a systematic decline in the formation of bold hypotheses; a degradation, if you like, of our collective imagination.
Capitalism, for example, has ossified.
New businesses start-ups have slumped, old companies stay ever longer in the main stock indexes and productivity has stagnated. Why? One key reason is that giant corporations collude with politicians and bureaucrats who end up on their payroll (via the revolving door) to rig the system against new entrants and ideas. We call it regulatory capture.
Capitalism has become a graveyard of stillborn ideas, most notably in the EU.
Science is drifting this way too because the overzealous use of the gatekeeping mechanism known as peer review serves to block the emergence of unconventional ideas and theories that challenge the prevailing groupthink, not least in social and climate science. In the wider culture too we see state censorship of speech (and, by implication, thought), not to mention social media rewarding not bold and imaginative ideas but derivative and trivial ones. This too corrodes the collective imagination.
But here let me suggest something hopeful: the AI age may offer an opportunity to refuel our collective imagination. It has been noted that humans will still be doing practical and caring jobs for years to come, one reason I applauded Keir Starmer’s emphasis on apprenticeships last week.
Yet the growing fear is that AI will destroy jobs in more cerebral domains through its ever more powerful capacity to search and solve. This, according to some, will cause mass unemployment and perhaps social unrest.
The quality of the answers from ChatGPT depends on the quality of our questions
Permit me to offer a different perspective. Recently I watched an interview with Demis Hassabis, cofounder of Google DeepMind and winner of the Nobel prize in chemistry.
Speaking to the podcaster Lex Fridman, he talked about his optimism for an AIfired future and the potential for an “age of abundance”. No surprise there. But what was most fascinating was a response buried an hour or so into the interview, when Hassabis noted that AI has severe limitations that can only be exceeded by humans. The power of the imagination.
“You have to sniff out what the right direction is, what the right experiment is, what the right question is,” he said.
“So picking the right question is the hardest part of science and making the right hypothesis. And that’s what today’s systems definitely can’t do. I often say it’s harder to come up with a really good conjecture than it is to solve it. So we may have systems soon that can solve pretty hard conjectures ... like a maths Olympiad problem. But could a system come up with a conjecture worthy of study, a really deep question ...? That is a far harder type of creativity. Today’s systems clearly can’t do that. And we’re not quite sure what that mechanism would be.”
This is precisely what Popper and Einstein were driving at, and it may prove the most important observation of the coming age. It hints at the intriguing possibility that the human imagination is set to become more important, not less so. It is our capacity for thinking in metaphor, in context and via analogy — our skill in coming up with hunches, theories and rousing conjectures — that could provide the rocket fuel of progress, in conjunction with the handmaiden of AI.
There are many implications of this insight. For a start, it might finally help us to overcome our obsession with rote learning and the regurgitation of chemical formulae and the like in our educational systems. Yes, our brightest kids should absorb rigorous knowledge (a prerequisite for imaginative leaps) but this new age will increasingly demand the mindset and skills to propose ideas and to test them.
This is also likely to be the key skill in working effectively with so-called large language models such as ChatGPT. It is often said that these bots are making us dumber; that we are outsourcing our thinking to algorithms. Perhaps that is true in the short term but I doubt it will be true in the long. Anyone who has played around with ChatGPT will have noticed one striking thing. The quality of the answers depends on the quality of the questions — in other words, the quality of our conjectures. And this is our species’s USP, if we could only rediscover it.
After his trip to Principe, Eddington said: “Whether in the intellectual pursuits of science or in the mystical pursuits of the spirit, the light beckons ahead, and the purpose surging in our nature responds.” This is the spirit we need today more than ever before. It is why we must free capitalism from the bondage of regulatory capture, science from the overzealous use of peer review and our children from the confining structures of analogue education. It is time to ignite our collective imagination.
No comments:
Post a Comment