The Money of the Future will be built from Blockchain and AI Lego

7月 29, 2020

Many criticisms have been levelled at blockchain technology but no-one would accuse the field of lacking ambition, or perhaps as a cynic would say, hype.

A leap forward from humanity’s 4,000-year- old use of gold as a store of value? Sure, but that’s for starters.

A means for cross-border finance that circumvents trillions of dollars of investments by governments and financial institutions? Yes, definitely.

A tamper-proof record of transactions that replaces the double-entry method of accounting used since the renaissance? Already possible.

A mechanism for coordinating human activity that could replace even companies as the prime driver of economic growth? Perhaps.

It’s fair to say that, to a casual observer, very few of the grand claims of blockchain technology have so far been realised. Bitcoin may function as an increasingly stable store-of-value but the complexity around owning and managing private keys is almost as off-putting to many people as asking them to transport and store precious metals themselves. Ethereum may be an open and decentralized super-computer but the experience of spending tens of dollars and waiting several minutes to submit even a simple token transfer would lead many to question what the fuss is all about.

It would be easy then to dismiss the field as being over-hyped or even dangerous, and one that should be avoided at all costs. Easy, and wrong.

Boom and Bust (and Boom)

The internet is an obvious and very recent example of technology advancing remarkably quickly from being over-hyped, slow and difficult-to-use to revolutionary. The dot-com boom and bust in the early 2000s seemed to show that the internet offered an impractical alternative to the way commerce had been carried out up to that point. It was also true that much of the internet’s potential in areas such as video streaming, social media and even shopping were hindered by the technical limitations that existed at the time. Over the years, as each limitation was overcome, new applications very quickly became possible. Fast-forward twenty years and the internet has transformed the economy and society in ways that could hardly have been imagined even at the peak of the bubble.

A similar trend occurred in the blockchain space following its boom and subsequent bust in 2017. Like the internet bubble, many frivolous if not outright fraudulent projects disappeared while innovators continued ironing out the weaknesses of earlier designs. At the protocol level, these include efforts to increase the throughput of transactions that ledgers can handle. Several different approaches including sharding, directed acyclic graphs (DAGs) and off-chain roll-ups have been shown to improve throughput. Many different implementations of these features are in action on live networks or due to be released imminently, and will transform the number of transactions that can be settled on chain.

Why we need an Internet of Blockchains

It should be borne in mind that none of the scaling solutions is a “magic bullet” as each one comes with different trade-offs between performance, user-experience, decentralization, cost and security. Similar trade-offs also apply to other aspects of blockchains. This means that ledgers that act solely as a simple payment system, for example, will out-perform more general compute engines for this narrow task.

Trade-offs also exist in the token economics, where the rate of token inflation reflects a compromise between the currency serving as a store-of-value compared with being a medium-of-exchange. Indeed, it would be surprising if it was possible for a single chain to fulfil all of the roles described at the start of this article.

If we accept the need for more than one blockchain, then the next obvious question is how these different chains should interact with each other. It’s clear that integrating fast payments with chains that focus on privacy can benefit users who can then access both ecosystems. The explosion in DeFi has also shown that exchange of different token types can be used to drive tremendous innovation in financial services. The need for communication between chains is where interoperability projects such as Cosmos and Polkadot come to the fore. These projects both work by allowing transactions to be communicated across chains and, in theory, allow most existing blockchains to be integrated with each other.

AI, Agents and Next-Generation Applications

The integration of different chains is also vital for decentralized machine learning algorithms and multi-agent systems. A mesh of interconnected blockchains could enable one chain to specialize in servicing agents that operate in the energy market while another deals with transportation and yet another supply chain. A customisable mesh network would also enable each chain to comply with different regulatory requirements that apply to different market sectors or different geographical regions.

Collections of agents could optimize the outcomes for their owners by interacting with different marketplaces and each other in real time. An early use case might be DeFi where agents can be used to optimize trades by using either peer-to-peer trading off-chain, a centralized exchange or via one of the many connected chains that are available. In the future, less frequent cross-chain transactions could also allow, for example, a self-driving electric vehicle to supply energy to the grid or deliver a parcel depending on the demand in the respective markets.

Another potential application is collective training of machine learning models where a hub chain can manage financial incentives and user reputations while specialized side-chains keep track of ephemeral records of data and models during training. These side-chains could also support populations of agents that cooperate to provide decentralized oracles and prediction market services back to the hub chain for use in DeFi and many other applications.

The feature that multi-agent systems and interoperable blockchains have in common is that they both involve creating simple building blocks that can be combined in an infinite number of different ways to create new functionality. This flexibility means that the many trade-offs that apply to base-layer protocols can be overcome by hybrid systems that combine these elements. In time, it seems certain that these software, cryptographic, economic and social “lego” pieces have the potential to make the initial promise of blockchain technology into a reality.