I’ve been travelling around the US this week with a bit of a hectic schedule. So I asked Exponential View member, Gianni Giacomelli to step in and look after the wondermissive this week.
Gianni has been a really valuable member of our Exponential Do community as well, where he has coached dozens of us on tools and methodologies for using collective intelligence. So I asked him to frame some thoughts on the emerging domain.
Enjoy Gianni’s Exponential View!
Honoured to curate EV this week.
TL;DR - I work on designing digitally-augmented, collective-intelligence systems that help organisations innovate inside and outside of their (increasingly blurry) borders.
I’m currently leading research projects at MIT’s Center for Collective Intelligence, work as an advisor to climate-focused startups, and am the former chief innovation officer of a large IT services company. I have worked with software-driven innovation since the early dotcom times, so I am not scared of the current downturn, yet.
I believe that most innovation processes in our organisations and institutions are reminiscent of those built in low-knowledge-transmission times, when existing wetware-based algorithms favoured hierarchical and deterministic decisions. I also believe that innovation-process incrementalism is preventing us from seeing (and harnessing) the powerful systems that surround us – call them “superminds”. In an Exponential Age, we can do way better. Human vs. reptilian-brain better, that is. This is what today’s essay is about.
Thanks for reading,
Sociobiologist Edward O. Wilson once said that the real problem of humanity is [that] we have palaeolithic emotions, medieval institutions, and God-like technology. A positive exponential future will depend on how effectively that trio handles society.
Wilson’s is an insightful yet frightening framing. But, as I will argue, there is reason for optimism as long as we move beyond our fixation with our own individual brains as the centre of action – because they are clearly no match for the challenge at hand. Instead, we can and should deliberately enlist the power of digitally augmented collective-intelligence entities: superminds.
Think about it. Various forms of technology have triggered momentous changes in networks since time immemorial. Think of the emergence of the most powerful societies in history, from the Roman and British empire to the United States and modern China: they are closely intertwined with the evolution of networks and their technology such as roads, language, and media.
One could even intuit a more general power of very simple primitives of collective intelligence. For instance, nodes that are connected and fed simply but not randomly, are able to store information curated through some form of active inference. They may then predict future events, and can help organise lots of things. From tree-roots and fungi, to our brains, to industrial ecosystems and perhaps even some form of planetary Gaia system, network-based intelligence drives successful evolution through new organising structures.
In this essay though, we will focus on something closer to home: organisational, economic and civil-society networks, and how we can deliberately and effectively design and engineer them to fight climate change, build a future of work that works, improve our democratic governance mechanisms – and more.
Superminds, originally conceptualised by Prof. Thomas Malone (founder of MIT’s Center for Collective Intelligence), are not Kurzweil’s Singularity things, or a philosophical or religious concept – at least, they don’t need to be. They simply represent a valuable framing for organisational design in an exponential world.
I describe future superminds as the result of the interplay of networks of people and intelligent machines, thus generating emergent, superior (and possibly exponential) cognitive properties.
Superminds are all around us already. From social media networks to blockchain and modern financial trading markets, from Wikipedia to Google Search to PatientsLikeMe, from Linux to Gitcoin and Bellingcat, from Taiwan’s use of Pol.is to Citizen Assemblies and OpenStreetMap, from drone swarms to Ukraine’s decentralised defence networks, from manufacturing giant Haier to hedge-fund Bridgewater to enterprises using Microsoft Viva Topics, bits of superminds are being harnessed already. (A few hundred examples of them and their components, big and small, here.)
The network is the mind – and it can be hijacked
Network effects underlying collective intelligence have changed the world’s competitive dynamics. Seven of the ten most valuable companies as of 2021 end rely on them extensively: Apple, Microsoft, Alphabet / Google, Amazon, Meta / Facebook, Tesla, and Tencent, use some form of supermind, as they harness communities and markets as part of their operating model. As EV regulars among you will know, Azeem has previously spoken with founders and investors who use network effects – a precursor to collective intelligence. I highly recommend listening to his conversations with James Currier of NfX, and a16z’s Andrew Chen.
Much of these companies’ strengths comes from their use of the cognitive power of billions. Web 2.0 has largely been about concentrating that power into the hands of the few, who could increasingly use AI to monetise its power – often for some sort of commercial recommendation system. In so doing, they’re steering the output of the supermind they’ve summoned in a very specific direction. The upshot is obvious, and might make the difference between sharing facts and propagating disinformation (which human networks tend to like), or between following democratic processes and sleepwalking into civil strife or autocracy. And just maybe, a hijacked supermind could be the endgame of AI gone rogue, which makes Nick Bostrom’s superintelligence nightmare come true. (Oh, if one thinks that good ol’ institutions and their circuit-breaking censorship can do better, remember how China’s social-media signal suppression slowed down a global response to Covid-19.)
Like all powerful technology-based practices, digitally-augmented collective intelligence can be used for good and for bad – and/or incompetently.
And some are using them, with often insufficient scrutiny and regulatory boundaries, as the raging debate on social media content moderation attests. The cat is out of the bag.
Tomorrow’s problems, solved with tomorrow’s intelligence
There’s a lot that we can do with intentionally engineered superminds.
Think about using them for fighting climate change, and specifically to hasten the adoption of relevant technologies. The cycle historically takes decades but for climate change, we just don’t have that time. I believe we can do better today than we ever did in previous technology-propagation waves. Hyperspecialised collective-intelligence “utilities” could accelerate the spread of high-momentum/low-signal content (both granular practical enablement and broader learning materials), and support the identification and engagement of relevant people (experts and practitioners). The upshot would be that the new granular, practically implementable knowledge could be quickly matched with the early majority of mainstream users to cross, in the words of Geoffrey Moore, the innovation chasm.
Decentralised-science superminds could benefit from a potential revolution in collaboration (think knowledge graphs and AI transformers) and incentives (think tokenisation).
The Future of Work agenda can also benefit from, for instance, a better understanding of the drivers of the Great Resignation by using a network lens; or by rekindling the serendipitous connections between “weak-tie colleagues” that are often severed when working from home, and yet are so important for innovation and culture.
I am sure there are examples in the Exponential View community too. I am familiar with this one, for instance - as they plan to use collective intelligence as the basis for trusted and transparent verification and tracking of impact of nature-based carbon solutions. If you know of others, I would love to learn about them.
Where do we go from here?
One of the hardest skills to find in innovators is the ability to think in systems – superminds are systems-design for organisations, and I believe they can be learned by extending current innovation practices. My MIT colleagues and I call this “supermind design”. My money is on the prediction that our children, and some of us, will have jobs such as supermind designer, supermind engineer, and supermind operations lead. I am open to (crypto denominated) bets on this – or perhaps let’s just discuss.
Intelligent networks, made by large numbers of people and machines, are a new organisational design ready for widespread adoption. A hint at how to build them is below.
In essence, a supermind is a system with four high-leverage design points:
1. Illuminate and connect the nodes: i.e., right people, and machines, e.g., knowledge graphs pinpoint the right nodes in the network and make them discoverable to each other, irrespective of how distant they are.
2. Incentivise the system: rewards e.g., extrinsic ones through Web3, norms and culture, trust and goals.
3. Curate and feed information to prevent the network from becoming insular e.g., NLP; automated data discovery to identify patterns people miss; deliberate diversity of narratives; computer-aided generative design.
4. Collaborate: Remove friction in the process of generating and applying ideas. E.g., NLP translation and Transformer based summarisation, DAOs.
Unsurprisingly, Robert Metcalfe himself sees value in these developments.
Engineering a supermind requires answering many design questions. Two of the most common ones are:
1. How large should it be? There are advantages and disadvantages of scale (e.g., diversity of ideas, representativeness vs burden of coordination). In general, keep things no larger than your coordination mechanisms can manage.
2. Should it engage the nodes in collaboration, or in competition? NumerAI relies on competition between users, and that makes it scalable – the signals from thousands of nodes are assembled seamlessly in a central metamodel. Others, like Wikipedia, Linux, PatientsLikeMe or Bellingcat work through more collaboration. As a rule of thumb, competition is faster to start, incentivise, and somewhat easier to scale, but in many cases individual nodes can’t do the job independently (for instance, if they need each other’s skills or data points). Again, interdependency burns energy in a supermind, so keep it as light as appropriate.
(There’s a lot more to it. For more detailed guidance, take a look at my open source book.)
Our society has created huge challenges through the interplay between our palaeolithic brains, our archaic institutions, and our technology. But we can take the logical next step to build superminds that overcome them, and withstand the thrust of an exponential world.
What you’re up to – notes from EV readers
Laure Claire Reillier is organising a debate about the future of digital platforms at Platform Leaders on Jun 7, 2022.
Ben Welby and his team the OECD are opening up their Good Practice Principles for public consultation before the planned launch in September. They also published a policy note on Designing and Delivering Public Services in the Digital Age.
Anil Seth’s Dreamachine experience is live in multiple locations in the UK (tickets are free and going fast).
Eliot Peper's new novel Reap3r follows a quantum computer scientist, virologist, podcaster, venture capitalist, and assassin who come together to untangle a twisted enigma that will change the course of future history
Andreas Widmer’s new book The Art of Principled Entrepreneurship argues that we have lost sight of the human person as the centre of the economy, and that investing in employees is the only opportunity with the possibility of infinite returns.
Eduardo Plastino and Euan Davis published “The Future of Us”, a field guide to help companies navigate the “net-zero era”.
Jonathan Lu and Shauheen Etminan, co-founders of VCENNA, will be presenting the ethnopharmacology of psychedelic substances in Chinese and Iranian medicinal history at the Ethnopharmacological Search for Psychoactive Drugs conference May 23-26 in London (livestream available).
To share your projects and updates, fill out your details here. Because of space constraints, we prioritise updates from paying members and startups I have invested in. (You can become the former by subscribing, if you have not already, and the latter by getting an intro to me via a trusted contact.)
How happy are you with today's wondermissive?
Sign in or become a Exponential View member to join the conversation.