There are a few digital coins that claim to be 'artificial intelligence digital currencies', but so far none of the public crypto currencies that make that claim has much to do with artificial intelligence nor with ai networks. Most of them are simply distributed computing networks, i.e., large computers, but some like Numerai are a beginning step to creating ai networks.

Artificial intelligence coins, or ai coin networks, are coins that use human mining or 'micro tasking' as a way to mine the digital currency.

A group of miners constantly update algorithms that provide automated 'answers' to scientific questions, in competition with other miners who create parallel algorithms that ask automated scientific 'questions'. The network is just an accelerating q&a for the specific science being developed.

Many corporations and governments are anticipating this shift and have created mechanisms that are not currently profitable, but which will position them well.

Sciences develop by discriminating between a number of alternate paths.

People are taught, starting in primary school, that 'the scientific method' is how sciences are developed.

Obviously the concept, as it is used in primary school, is oversimplified and doesn't lend itself directly to digital currencies. When these networks develop, those who can make them work will, and those who can't won't.

One likely scenario is that 'forks' of ai networks will develop based on faulty paths, forcing a hard return to the pre fork, as opposed to the current digital economy where forks are strictly gang activity, the fork with the biggest gang wins.


Numerai and similar coins let people construct algorithms that do 'something', and then get paid based on the efficiency of the very narrowly focused algorithm they created.

For example a person creates an algorithm to make stock market predictions. The algorithm might ask for 50 variables, like a) market cap, b) cash on hand, c) length of time CEO has been in that sector, etc. Those variables are entered for ten companies and the result of the algorithm is 'buy company a, don't buy company b'.

Ten people each make their own algorithms, and the one that makes the best algorithm gets the coins.


There are a lot of possible ways to use networks to develop sciences, but the first attempts will be simple variations of the above, but instead of predicting stocks using one 'player', they will develop and test algorithms using multiple sides.

So, for example, while Numerai has a) a simple network with an algorithm, and b) players each with a simple algorithm, ai networks will have

a) in place of the network's 'simple algorithm' will be an algorithm that tests other algorithms by incorporating them into some kind of 'scientific method' variation'

b) and in place of players creating static algorithms to estimate companies, there will be

1) People who create 'software' or fairly developed systems of algorithms, which accurately do something scientific. A 'mathematics' software might add, subtract etc. A chemistry calculator might give the properties that could be expected from creating compounds built with certain elements etc. These might be called 'scripts' rather than software, since they will be small and specific.


2) People who create much simpler software that is like a consumer of that software. This software will need to 'use' the software created by the first group in a practical way.

So far for example you have a group producing 'math calculators' that add and subtract, and a group that makes simpler software that feeds numbers into that first software. It asks the first software 'what is 1+1?, what is 5+5?' etc.


3) People who are 'experts' in that specific science i.e., who can intuitively estimate which developmental directions are likely to be dead ends, and which likely productive.

These people will be somewhat different than comparable scientific 'experts' in a corporate environment.






In Progress