Fetch.AI has announced the launch of what it claims is the world’s first adaptive, self-organising smart ledger technology for decentralised transactions.
Fetch is described as a next-generation protocol that enables autonomous economic agents (AEAs) to perform proactive economic activities. The technology allows AIs on both sides of an economic transaction to interface with each other, said the company in an announcement today.
In other words, the system enables AIs to buy and sell digital assets autonomously, from or to another AI, with contracts, payments, and execution all handled automatically.
The company says that corporations can interface their closed, internal networks with the Fetch network without exposing their proprietary IP.
According to the company, Fetch’s distributed ledger system also generates tokens that can be exchanged for utility value on the network, which provides an incentive for operating a node. In this sense, the system includes concepts ported over from cryptocurrency mining.
So how does it all work?
Removing the blocks from the chain
Fetch combines artificial intelligence, machine learning, and distributed ledger technologies to deliver what the company describes as a self-organising framework.
The underlying smart ledger combines elements of blockchain and directed acyclic graph technologies with built-in AI, a consensus mechanism, and a data architecture that “can support millions of agents transacting together”.
Together, the technologies can scale to millions of transactions per second at near zero cost, claimed the company today.
The new consensus mechanism is called Useful Proof of Work. Unlike traditional blockchains, the Proof of Work computations needed to secure the network aren’t abstract, according to Fetch. Instead, they are “redeployed to perform the machine learning tasks that give the ledger its intelligence”.
Using this new system, developers can build autonomous agents that might act on consumers’ or companies’ behalf, and which can be designed to address “almost any task”.
Founded in 2016 and operating in stealth mode until today, the Fetch team includes original pioneers from DeepMind, the AI company subsequently acquired by Google, as well as economics and cybersecurity professors from Cambridge and Sheffield universities.
CTO Toby Simpson – who was part of the initial DeepMind team – said, “Autonomous economic agents are set to revolutionise commerce. They’re digital entities that can transact independently of human intervention and can represent people, machines, or themselves.
“Imagine a world that connects anything to anything, and everything to everything, where data, services, and information get up on their own two feet and deliver themselves with incredible precision.
“The Fetch digital world acts as the ultimate value exchange dating agency: each AI agent sees a space optimised in real time just for them, where the important things are right there.
“Autonomous agents can work together, or alone, to build solutions to complex problems by joining a disparate array of potentially useful data and services into one seamless experience.”
According to co-founder and CEO Humayun Sheikh – who was a founding investor in DeepMind – this radical approach is necessary to address what he called the “inefficiency inherent in today’s economy”.
“Today’s centralised systems and marketplaces result in silos and there isn’t a means to enable complex economic activity,” he said.
“For the first time, Fetch delivers a self-organising digital world that enables the discoverability of data for decision-making, the trading of excess capacity, energy, and computation, or the storage, transfer, and transportation of digital or physical assets.
“When you have a decentralised collective intelligence and AI operating in the right environment, which our open network provides, an autonomous system like Fetch can solve problems before they even arise.”
The company provides a number of working examples. In the travel sector, Fetch says it is working to reduce the burden of planning complex journeys by using expert autonomous agents to organise trips, predict missed connections, and dynamically reroute journeys, rearrange travel plans, and rebook reservations without intervention.
In the energy sector, Fetch said it is working to deliver the most effective energy solutions to households without the friction of switching suppliers, potentially analysing each appliance’s unique energy demands.
And in the supply chain, Fetch says it is working to help the trillion-dollar steel sector to self-manage and optimise its supply chain “autonomously and collaboratively”, from raw material to finished product, including stock management, transport, financing, and insurance.
Jamie Burke, CEO of VC firm Outlier Ventures explained why he has been backing Fetch since 2016: “For the first time we can genuinely say we have not smart contracts, but instead a smart ledger.
“To date, blockchains haven’t been capable of even rudimentary intelligence and anyone wanting to use today’s dumb ledgers for anything complex, dynamic or predictive has been left severely wanting.”
• Fetch plans to provide code on GitHub and announce a range of corporate partnerships over the coming months. Fetch will be providing a comprehensive SDK allowing the development of agents in languages such as C++ and Python, as well as a set of frameworks to allow Web developers to design, configure, and author agents.
Internet of Business says
A fascinating and innovative technology, developed by an impressive team. It’s early days in this new launch, of course, and we plan to bring you deeper analysis and an interview with the Fetch team shortly.
In the meantime, questions must arise around transparency, antitrust concerns, distributed resources, and the risk of such a system automating unorthodox activities. We hope to bring you answers to those questions soon. CM
Coming soon: Our Internet of Supply Chain conference.