tl;dr: With limited historical data, a global, 24/7 market and all kinds of actors, a new kind of financial AI is necessary.

I have been a fan of SingularityNet for a while now. I served as an advisor for a time and have a small amount of tokens as well.

What got me excited about it, besides the strength of the team behind it, was the vision…of a world where powerful AI capabilities did not rest solely with the largest, wealthiest players on the planet. It’s a big vision and not easy to implement, particularly on a decentralized platform like Ethereum (or even Cardano to which they recently announced a migration). In fact, it’s really difficult.

But, as I understand it, what decentralized AI can bring to the table is also unique.

In his post, Why Cryptofinance Needs Decentralized AGI, Ben Goertzel, the founder of SingularityNet delves into the reasons why the emerging crypto market is ripe for experimentation with decentralized AI and why they have spun out the SingularityDAO, designed specifically to meet the task.

What is the task?

It’s deploying hundreds (if not thousands) of AI models across the spectrum from Artificial Narrow Intelligence to Artificial General Intelligence in order create more effective prediction models for the smaller cap “alt coins” (say those out of the top 100).

The challenge for most AI approaches (and Goertzel goes far deeper than this in the post) is that they rely on vast amounts of historical data, something which the crypto markets (and even really Bitcoin) can’t readily provide.

What’s more, the crypto markets are subjected frequently to what Goertzel calls “regime change”:

 which in the crypto domain are relentless and extreme. Ongoing improvements in technology, shifts in the regulatory landscape, and expansions in the nature of the crypto investment community (e.g. one year China heavily restricts crypto, the next year a huge number of traditional Wall Street firms come into crypto, etc.) mean that the markets at any given point in time generally have radically different characteristics from what they had a year before.

Plus, it is sadly well known that there is still a lot of market manipulation, as Ben points out when it comes to crypto assets and thus, a model would need to account for “how do market manipulators feel on a given day?”

Lastly, size. Most of the altcoins don’t yet have sufficient volumes of trade in order to provide meaningful data to train an AI.

So, while there are a number of ways to tackle this (e.g. the competitions that Numerai runs), the key is to “let a thousand flowers bloom” (maybe not the best quote) and create a community owned platform that can test multiple models for AI against these smaller alt coins in a nascent market, allowing for experimentation while also rewarding members of the group.

This is the purpose behind the SingularityDAO and I’m excited to see how it evolves.