Autonomous Economic Agents: an update

Sep 18, 2020

There are billions of devices, processes and pipelines around the world that are desperate for automation and communication. Most blockchains aren’t built to handle such a large number of transactions, and none enable their clients to learn from the data generated and exchanged.

Fetch.ai is a platform built to bring value to data, by connecting devices and facilitating the creation of new markets based on what they learn from each other. It is the platform where truly autonomous decisions can be made.

The Fetch.ai approach to delivering this autonomy is via a decentralised network supporting an ecosystem of many “Agents” that search, negotiate and transact in an economic internet.

A big part of our tech roadmap is in delivering the core protocol and infrastructure to enable this new multi-agent system based economic model to take place.

We define an autonomous economic agent or “AEAs” as:

“an intelligent agent acting on an owner’s behalf, with limited or no interference, and whose goal is to generate economic value to its owner.”

The AEA framework is a Python-based development suite which equips you with an efficient and accessible set of tools for building AEAs. The framework is modular, extensible, and composable. It attempts to make agent development as straightforward an experience as possible, similar to web development using popular web frameworks.

AEAs act independently of constant input from their owner and autonomously execute actions to achieve their prescribed goal. Their goal is to create economic value for you, their owner, in a clearly defined domain. AEAs have a wide range of application areas and we provide demo guides for some examples.

So, it’s Friday, and what better way to go into the weekend than with an updated, polished, super-cool new release of the Agent Framework.

We update the Agent Framework fairly regularly, but changes this week have major benefits to several agent projects in development right now.

It includes:

  • Filtering of outdated addresses in DHT (distributed hash table) lookups. This was affecting some development where lookups failed when outdated matches were still present.
  • Updates to the soef connection and search protocol to support new features, including: more comprehensive search return data and additional registration commands. See yesterday’s 0.1.30 soef release.
  • Updated multiple documentation sections, including soef and additional explanations of its usage.
  • Several multiplexer updates and fixes including a disconnection issue being resolved and configurable exception policies.
  • Test coverage of all protocols now 100%
  • Comprehensive new benchmark scripts

And multiple minor tweaks including bug fixes, additional of README.mds files to all skills and HTTP header support in http server connection.This release’s changes support on-going development of the Fetch decentralised delivery network, the hospitality agents (hotel agents and representative agent), and several other projects both by Fetch and others. We’re very excited!

As usual, if you have any questions, or would like support and assistance getting going, please talk to us at the developer Slack.

You can read more information here   

And it’s available from PyPI here: https://pypi.org/project/aea/0.6.1/

And the updated docs are here: https://docs.fetch.ai/aea/

To get started developing your own AEA, check out the getting started section on our docs site.

To learn more about some of the distinctive characteristics of agent-oriented development, check out the guide on agent-oriented development.

If you would like to develop an AEA in a language different to Python then check out our language agnostic AEA definition.

AEAs achieve their goals with the help of the Open Economic Framework (OEF) — a decentralized communication and search & discovery system for agents — and using Fetch.ai’s blockchain as a financial settlement and commitment layer. Third-party blockchains, such as Ethereum, may also allow AEA integration.

Finally, we’re super proud of the Agent Framework and we are working hard to make it increasingly accessible, including to those that are not programmers.

We are working towards some exciting developments in the coming months to showcase how visual development might work, and to deliver a tool that lets anybody explore the Fetch digital world looking for, and trading with, agents.

Image for post

If you want to hear more about where we are, where we’re going, and the future vision remember to tune in for our Autonomous Economic Agents Deep Dive community Q&A at 13:00 UTC, 24th September 2020, here:

https://www.crowdcast.io/e/autonomous-economic