Dijets Method Chain has thus far always taken a back seat when it comes to flexing its use-cases and utility. While both of the value and utility chains took centre stage when it came to updates, releases and generally being in the mix for network activity, the Method Chain stood patiently on the sidelines waiting for its time. Whether it was the intra-chain transfers, minting of new assets or interacting with erc20 tokens, the value and utility chains were the two main points of entry and exit for anyone wanting to do anything on Dijets Network. So naturally it gives me a lot of pleasure in seeing the Method Chain finally come out of the shadows and show off its very first utility.
Just as other chains on Dijets, Method Chain was built for a specific purpose. However, unlike the other chains, Method Chain’s purpose is two-fold. 1. Being the metadata chain for staking, validators, cacids and delegations etc. 2. Being the gateway to “Dijets Collaborative Intelligence Initiative”. An initiative to build a decentralised database of proprietary “methods”. Methods are the digital equivalent to patenting an idea in real world. They can be used to describe both complex logic and a simple trading strategy. Think of Methods as instruments of financial agreements, settlements, strategies, DeFi recipes, ideas and more. Rogue - the method chain’s flagship platform service enables non-programmers and everyday users to build their own Methods simply by describing it in their natural language. The natural language processors and the AI engine for method analysis within Rogue was built using OpenAI’s phenomenal GPT-3 and Codex technology. Non-technical users or individuals with no prior knowledge of programming or smart contract coding can easily declare a DeFi strategy as a Method by sending an instruction in plain english.
By combining the autoregressive 3rd-generation language prediction model developed by OpenAI, and large data sets of method logic, synthesised code libraries and an extensive library of pre-trained provisions specific to Dijets and Methods, we have successfully fine-tuned and trained the AI to a decent starting point where a simple instruction by the user in plain english is enough for it to derive and construct method logic contained within it. By removing one of the biggest entry barriers to decentralised finance for the masses we hope to take big strides towards true global adoption of blockchain technology.
Dijets Collaborative Intelligence Initiative
AI is only as good as the data sets it gets trained on. A state of the art transformer-based deep learning algorithm with unlimited resources would amount to nothing if the data it was trained on only consisted of 2 random paragraphs of disjointed text. We believe that the most extensive data for Method Chain AI should come from the end-users themselves. The Method marketplace, combined with users submitting Methods and Method logic to Rogue platform should be the best way to continue training it in the future with more resources allocated to its refinement (DIP territory this). Using the same principles of deep learning and transformer-based algorithm, the AI should progressively get better at synthesising Methods as more and more users submit their queries to it. The incentive for users will be to earn tokens in return for each Method or the logic contained within it gets qualified and accepted to be added to the dataset as a public domain method that anyone in the world can buy and earn from.
Pavil we’re did you get that information from mate
Way to go 