Takeaways from TED 2023: why the future is better than you think
Plus thoughts on how the future will be transformed, across tech and culture.
Ideas worth spreading. That has been the long-standing slogan of TED, the media organisation dedicated to sharing ideas around Technology, Entertainment, and Design, since its founding in 1984. After a rocky start (the second TED conference only happened in 1990), TED has gone on to become a global brand, hosting, to date, almost 4,000 TED Talks, extending to 13,000 subsidiary TEDx events, and reaching almost 2 billion views on YouTube.
At the 2023 conference, hosted in Vancouver, Canada, for the tenth consecutive year, AI dominated the conversation. Greg Brockman’s talk, co-founder of OpenAI, was the first to be released on YouTube eleven days ago, and has already reached 1 million views. The attention paid to artificial intelligence is as high outside British Columbia as it was inside.
And rightly so. It dominates my three takeaways, but the first of those is to emphasise just how hard it is to imagine the future.
THE FUTURE IS HARD TO IMAGINE
TED 2023 kicked off with a talk from Angus Hervey, co-founder of Future Crunch, a media group dedicated to inspiring the world with good news, science, and human progress.
He asked Why don’t we hear about good news?
Our headlines are dominated by bad news, despite the objective truth that the world, overall, is better than it has ever been before. His suggestion is that good ideas, like international agreements, take aaaages to create, whereas bad things happen quickly. To an extent this is true, but good things can happen quickly too. In his book, Humankind, Rutger Bregman suggests that good news is scarce on our screens because it is so common. The news, after all, is about what is new. And if good things are happening all the time, then it isn’t news. Bad things dominate the news because they are so rare.
I’m not sure that these ideas are convincing enough. Not all good news is common, after all. Hervey was followed by Stuart Kaufmann, a theoretical biologist, who contributed his own perspective: the news is dominated by bad news that worries about the future, and that’s because it is far easier to imagine a bad future than it is to imagine a good future.
His central thesis is that any invention can be used for more than its intended purpose, and in unpredictable ways. The internet was invented for military communication, and now I’m doing research and writing and emails. Teflon was originally created for use in factories, and now it is in millions of non-stick utensils in kitchens around the world.
Kaufmann argues that it explains the ‘hockey-stick’ level of growth and progress that the world has seen in the past few decades, and you can take the same core idea and apply it to any technology in the world. Ultimately, when you create a new technology, the question of whether that invention is ‘good’ or ‘bad’ does not depend on what the technology is, but on whether people are good or bad1. But critically, it is difficult to imagine how good it can be.
The fact that it is difficult to imagine how good the future can be is exactly the one made by Yat Siu, co-founder of Animoca Brands, one of the world’s leading web3 investors. Siu’s core idea is simple: digital property rights, built upon blockchains, will create enormous levels of value, even if today it is hard to understand how.
And today it is hard to understand how. Digital property rights refer to whether people own their data, or whether it is managed by a third party like Facebook. But as media analyst Benedict Evans points out, whilst Facebook makes billions of dollars from its users’ data, per user, that data is actually just worth a few dollars a year.
Kaufmann’s point is that with new technologies, it is difficult to imagine how the future can be better. And Siu’s argument is that the future can be a lot better with digital property rights. After all, real property rights, for things like real estate, were transformative for the world and underpin almost all of the wealth and prosperity the world enjoys.
That’s a fair reason to suggest that digital property rights could make your data worth a lot more than it is today. Those sorts of rights would empower you to manage your data and sell it, it would empower businesses to be built on it, it would empower new value chains. Data today is largely locked up in the servers of a small number of companies. Siu’s suggestion, given more credibility by Kaufmann’s ideas about where progress comes from, is that if people are given ownership of the data on those servers, entirely new industries, business models, and businesses will emerge.
ARTIFICIAL INTELLIGENCE IS ENABLING ENTIRELY NEW THINGS
As with blockchain, artificial intelligence. TED gave us a taste of the new uses that will emerge from neural networks and machine learning. We heard from Tom Graham (who I met later in the week, lovely man), whose company Metaphysic lets users create fake media. Tap the video for a fake of Tom Cruise.
It was very easy to imagine the bad things that will come from this. It is not so easy to imagine the good, but following Kaufmann, I will believe that entrepreneurs, inventors, and creatives will discover more good than there is bad.
But one of the things it emphasised for me is the importance of ‘blue checks for the internet’. On Twitter, a Blue Check refers to a ‘verified profile’. Someone has verified that you are real, and the Blue Check is a signal to prove that to everyone else. If we’re heading towards a world where anyone can make fake video and audio content, then we need Blue Checks not just for Twitter, but for the entire internet.
Some companies are developing this sort of technology (e.g. Worldcoin is building towards ‘proof of personhood’), though it really should not be a for-profit effort. Similarly, we need ways to track how our likeness is used across the internet. Already, fake songs that sound like Drake and Kanye West have gone viral. And whilst this is scary, it is also an opportunity for these artists, who could have a way to monetise their name, image, and likeness on a greater scale. But what is needed for both internet Blue Checks and tracking our Digital Likeness is a technology that anyone can trust and that isn’t owned by one organisation. Something like a blockchain.
Artificial intelligence is offering much more than this. Karen Bakker, a Canadian researcher, showed us how AI is letting us decode the languages of animals, which communicate in sounds above and below a human hearing range. There is a huge well of untapped knowledge to be gained from this. One insight scientists have discovered recently shatters the traditional stereotype that turtles abandon their unhatched young and disappear into the ocean. Instead, researchers have learnt that adult turtles communicate with unhatched baby turtles and instruct their children how to meet them in the ocean.
Even volcanic vents and plants make sounds that humans have previously never, ever heard. It gives us a completely new dimension to understand the world around us.
And speaking of plants, Jeff Chen gave me a significant piece of insight when it comes to natural medicine. The West traditionally mocks natural medicines, but he points out that natural medicine is significantly underrated. The reason is capitalism. Because natural medicines are, well, natural, it is impossible to patent them. And thus it is impossible to monetise them. And given that, why would any company pay billions of dollars to push natural products through regulatory trials? It is simply uneconomical, and thus there is an entire world of natural medicinal cures that are not approved, but should be, due to the regulatory and economic structures in the industry.
The relevance of all this comes down to artificial intelligence, which with powerful statistical models, is able to reliably demonstrate the viability of natural medicines far more cheaply than before. And not only that: AI can also tailor (and develop) medicines designed specifically for you.
Insight without knowledge.
Jeff Chen’s work, led in partnership with his co-founder Pelin Thorogood (who I also met), is an example of what econ-writer Noah Smith called The Third Magic. I loved his essay about this. Smith outlines three general types of knowledge.
First is history: we write facts down, like ‘this flower is poisonous’ or ‘this is how you make bread’, and thus accumulate information. We write information down and get it later.
Second is science: here, we figure out general principles of how the world works, like biology or chemistry. We derive laws through testing hypotheses and induce new information from those laws.
Artificial intelligence is the third magic. We input data to statistical models, the models perform calculations, and we receive accurate results, like which natural medicines to trust. The difference is that here, we don’t understand how the world works. For history, we understand which flower is bad and why. For science, we understand the rules which make flowers bad and others fine. For AI, we don’t understand the why; we have insight without knowledge. We don’t know what is happening inside the AI model. That is an exciting new model of knowledge, but it is risky: if we don’t understand our knowledge, we do not know when it will fail us; we do not know when the AI might get things wrong. What we do know is that this is a pivotal century in human history.
A Big Question About Culture
As I watched Wangechi Mutu talk on the TED stage about her work, I was struck by the question of ‘what makes this great’. Mutu’s work is intimately tied to her Kenyan heritage. A big part of what has made her work so meaningful is that it is tied to deeply-embedded emotions and cultural concepts2.
In essence, that suggests that in our quest to create great art (whether music, film, written, or otherwise), artists use these kinds of ‘shortcuts’ and apply their talent specifically to something that is meaningful. It isn’t enough to paint a beautiful flower. What matters is the emotion; the message; the meaning.
Ben Zander made this emphatically clear in his own talk. Alongside the talk by Gus Worland, Zander’s talk is definitely worth watching when it comes out on YouTube; I met him a few days later and can confirm he is equally energetic off-stage. But on stage, Zander’s talk was all about the idea that what makes music truly great, rather than just perfect or good, is throwing your emotion into it. The notes are not just notes, they are stories, pieces of nature, ideas… It is the job of the artist not to bring out the image or the sound, but to bring out the meaning.
But does a computer know what it means to cry?
We’re all agreed that AI can create wonderful images. And I have no doubt that it will be able to create art that taps into not just aesthetics, but meaning and emotion. What I’m interested in is how it does that. Because it won’t be by learning cultural history or experiencing heartbreak.
Artificial intelligence will create great pieces of art. But it will do so through statistical analysis, machine learning, and meticulously studying the past. It will take different shortcuts and develop new techniques.
Relatedly, Dutch artist Lonneke Gordijn spoke about her visual practice, and shared her story about programming drones so that they could fly as beautifully as real birds. When her team tried to programme beautiful flying patterns themselves, they came up short. It was only when they replicated the murmurations of birds that their drones achieved real beauty. In this sense, the flying patterns of birds presented a ‘shortcut’ for the human team to create entirely man-made beauty. The reason? Those patterns are deeply ingrained in thousands of years of human history. Replicating them taps into the same human response: that those murmurations are pretty.
So the question is, does AI follow those natural patterns when creating beauty, or does it create new ones? In board games like chess and go, machines learnt human techniques, and then flew straight past them to develop entirely new approaches to playing. It is likely that when it comes to culture, machines will uncover new ways to move their audiences. Of course, AI is just a tool: it will still be humans prompting the machines. But the type of art, music, film, literature that we consider great is about to get tweaked at a very fundamental level. I wonder how, and how much.
Other updates
I wrote a guest post for my friend Peter Imafidon’s blog, The AEA Explorer, about how the builders of paradigm-shifting technology should focus on selling their ideas to the C-suite, rather than anyone else in an organisation. You can read it here.
Last month my start-up, Culture3, announced Culture3 Week and TED Tech. We’re co-creating TED’s first ever dedicated technology event. It’ll be in London in September. Details here.
Sometimes, like in the case of guns, the bad people have a bigger impact than the good people. In these situations it makes sense to regulate (e.g. ban) the technology.