Per Molander, in his book The Anatomy of Inequality, gives a precise and elegant game-theoretic analysis of how radical inequality will inevitably emerge in competitive economic systems – even without benefit of corruption, violence or other forms of ‘cheating’.
He looks at games in which various participants are repeatedly betting against each other on various random events, and notes that in many such games, you’re able to win more on average if you can afford to lose more occasionally.
Basically if you have a big financial cushion you can place a big bet on a good but risky opportunity when it comes up. Whereas, without that cushion, you couldn’t afford to bet the house on that same opportunity, so you’d lose the potential upside.
Molander’s argument explains how, in an economy based on a group of participants repeatedly making bets with each other, a small random advantage to one participant can get amplified over time until that one participant owns, essentially, everything.
Having a little more than everyone else lets you achieve a slightly better profitability than them – and then the advantage compounds.
This explains a lot of what we see in the world economy today – compounded by a bunch of more sinister aspects like military hegemony, regulatory capture and plain old corruption and collusion.
It also explains a lot of what we see in the tech industry. For instance, in the AI industry where a small number of advertising companies have come to dominate the AI R&D scene and hire a disturbing proportion of AI PhDs. And in the blockchain world, where a few cryptocurrencies with inferior technical properties, controlled by a fairly small number of parties (many associated with organised crime and totalitarian regimes), have captured the vast majority of the market capitalisation.
And this explains how an AI researcher like myself – with 30+ years of experience researching Artificial General Intelligence and building practical application-specific AI systems – has ended up in the role of a crusader against tech oligopolies.
My main goal since receiving my math PhD in 1989 has been to create beneficial AGI systems with both general intelligence and ethical intuition at the human level and then beyond. But, increasingly, I’ve been pulled into spending time, not just on the algorithms and data structures of cognition, but on creating frameworks to ensure that, once advanced AGI is achieved, it is not controlled by a small number of parties with selfish and narrow-minded goals and values.
I’m putting a fair bit of time these days into Hyperon – the new version of the OpenCog AGI system that I launched in 2008 – which brings neural, logical and evolutionary learning together into a common AI framework based on a distributed knowledge metagraph. But I’m spending even more time running SingularityNET – a decentralised organization focused on building a decentralised blockchain-based protocol for connecting multiple AI agents within a network that nobody owns and controls. Because I don’t just want to create truly thinking machines, I want to create truly thinking machines that are controlled by a vast network of people in a democratic and intensely participatory way, and because it seems to me this is by far our best bet at getting AGI systems that evolve mostly cooperatively with humanity as our values and culture advance.
The battle against Big Tech for the soul of global AI is an uphill battle for sure, but so was the battle of Linux against corporate computer operating systems, and so was and is the battle of the open Internet against those who would monopolise online communications.
At least the technology for decentralized and democratic large-scale AI systems seems to be coming together recently, with the rise of the Cardano blockchain which combines democratic community based governance with high quality, highly scalable software design (note: SingularityNET is in the process of porting most of its tech from Ethereum to Cardano).
The big challenge is defusing the massive head start that Big Tech has accumulated via the game-theoretic dynamics Molander identifies in his book. The only approach that can work, I believe, is changing the rules of the game.
Whole new ballgame
Big Tech may have won the iterated gambling game regarding conventional fiat money, but we can make up new games, in which participants cooperate in making their own new kinds of money, and unleash our new decentralised AIs within these new game universes, which can then do battle with Wall Street which long ago cemented its dominance of the finance world. DeFi, though, is a whole new ballgame.
Big Pharma has captured traditional medical research, but tokenomically crowdsourced medical data gathering routes around the methodologies Big Pharma understands.
And so I find myself dividing time between developing breakthrough AI tools for advanced cognition and various uses of its applications, and developing novel platforms for deploying, delivering and controlling these in a decentralised way.
If you’re a technologist or a tech entrepreneur, I’d encourage you to start thinking the same sort of way.
I believe my friend Ray Kurzweil is right – we are rapidly approaching a technological singularity. But I was frankly disappointed a decade ago when Ray joined Google, because I think one of the keys to achieving a beneficial Singularity will be making sure that mega-corporations like this are less significant than decentralised networks in controlling the AI infrastructure.
Thoughts on Entrepreneurship, Blockchain and the Singularity
Coming out of all these experiences, here are a few lessons I’ve learned in the course of working at the intersection between AI, blockchain and other cutting-edge tech and the practical human and business worlds:
On blockchain and decentralisation…
● Decentralisation is interesting and fun, but it’s not intrinsically a good thing. It’s just the absence of centralisation, which can work worse or better than centralization – perhaps even sync harmoniously or awkwardly with centralization – depending on many factors.
● ‘Anarchy’ doesn’t mean no structure, it means no rigid immovable overarching structure – similarly, ‘decentralization’ doesn’t mean no centralised nexuses of execution or innovation; it means none of these should be locked-in, unchangeable and non-adaptive. How to balance decentralised and centralised structures and dynamics to achieve an effective organisation is a subtle art.
● Designing subtle and beautiful tokenomic models is a lot of fun. Seeing what happens when these models intersect actual human and business reality is a rather different kind of fun – which rarely seems to involve the models working out as expected.
● Should we think of the global brain as an emergent intelligence whose EEG signals are as crazy as a higher-dimensional version of the crypto markets?
On exponential change and the coming Singularity
● This is the first time in human history where a series of radical transformations in society and technology will occur within a single person’s career. Advances that occur “one funeral at a time” won’t work anymore. For the first time ever, we actually need lots of folk to be able to change their minds in fundamental ways, and then do it over again. As a business leader, if you fail at doing this, your business will die or at best profoundly suffer.
● Arthur C Clarke said: “Every revolutionary idea seems to evoke three stages of reaction. They may be summed up by the phrases: (1) It’s completely impossible. (2) It’s possible, but it’s not worth doing. (3) I said it was a good idea all along.” As technological and societal change accelerates, the transition between these stages is happening ridiculously fast.
● Many complex dynamical systems, when approaching a bifurcation or multifurcation point, display massive degrees of chaos and volatility. The singularity is the ultimate multifurcation of human history
● The only way we can come anywhere close to ensuring the coming revolution in machine consciousness will lead to beneficial outcomes for humans is to couple it with a revolution in positive human consciousness.
● The two traits that seem likely to be automated last as singularity approaches are: leadership skill and radical creativity.
About Dr Ben Goertzel
CEO & Founder of the SingularityNET Foundation
Dr Ben Goertzel is the CEO of the decentralised AI network SingularityNET, a blockchain-based AI platform company, and the Chief Science Advisor of Hanson Robotics, where for several years he led the team developing the AI software for the Sophia robot. Dr Goertzel also serves as Chairman of the Artificial General Intelligence Society, the OpenCog Foundation, the Decentralized AI Alliance and the futurist non-profit Humanity+.
Dr Goertzel is one of the world’s foremost experts in Artificial General Intelligence, a subfield of AI oriented toward creating thinking machines with general cognitive capability at the human level and beyond. He has decades of expertise applying AI to practical problems in areas ranging from natural language processing and data mining to robotics, video gaming, national security, and bioinformatics.
Dr Goertzel has published 20 scientific books and 140+ scientific research papers and is the leading architect and designer of the OpenCog system and associated design for human-level general intelligence. Together with Cassio Pennachin, Dr Goertzel co-authored “Artificial General Intelligence”, published in 2002 by Springer Publishing. He is also the chair of the Artificial General Intelligence (AGI) conference series, advisor to Singularity University and former Director of Research of the Machine Intelligence Research Institute (formerly the Singularity Institute). He also served as Chief Scientist Officer for Hanson Robotics until early 2019.
His appearance on the Joe Rogan Experience podcast and most recently on the Lex Fridman Podcast, have scored more then three million views on YouTube and the wide-ranging discussion continues to be viewed, shared and referenced today.