The humans behind the artificial general intelligence, or AGI: LegitDAO
LegitDAO is a decentralized autonomous organization, or DAO, that currently regroups more than 300 individuals. They will be the humans behind our blockchain-based AGI.
In order to create a successful AGI, we need to be able to accumulate data, program specific AI models, create chained transformations, train those AI using humans and then use processing power to process the data on the models.
Among other tasks…
Accumulate data
In order to accumulate data, we need to create decentralized marketplaces. Some of those will contain digital value, like the bytescript marketplace. Others will contain physical goods that will be sold by one entity (1), shipped then received by another. At last, other marketplaces will contain services provided by one (1) entity and received by another.
These three (3) examples of marketplaces will contain data that explains how humans interacts with each others in a commercial matter. It will help our artificial intelligence better understand how fraud is created and how to prevent it, among other very useful use-cases.
However, while we train the AI with that data, problems that cannot be automated at the moment will arise in our marketplaces. Those problems will be managed by the LegitDAO entity and voted by its members: they will act as moderators on transactions.
If someone tries to defraud another member of our community by selling a bogus product, or by saying that he didn’t receive his product in good condition, the LegitDAO will act as a middleman in order to judge the situation and act accordingly.
Hire decentralized contractors
In order to make our artificial intelligence grow and evolve over time, we will need to program new AI models, chain model transformations properly, train them using humans and process that data on models.
Program AI models
In order to program new AI models, we will need to hire contractors that will be able to propose new models to our ecosystem. Some of those models might be programmed by themselves, or be open source and analyzed by our contractors to make them fit in our ecosystem.
Create data for AI models
In order to train AI models, we need to have actual data that represent known things. For example, if we want to have an AI model that detects a new molecule that can destroy a given protein and therefore annihalte a sickness, while not affecting other molecules to not kill the individual that receives the medecine, we need our AI to first understand what a molecule is, how they are formed and what are the physics rules to form them and make them react together properly.
In order to to that, we need to create databases of known things and create relationships between them.
Therefore, we will need contractors to create these databases, provide credible information sources and review the validity of that content.
Link data to AI models
After being approved, a new AI model or model chain needs to be linked to data sources and then make that new model learn from that data. Therefore, we’ll need people that understand the context of a given AI model and then assign that data to it.
Process data
At last, we will need to find contractors that have GPU cards and make them process our AI models with our data so that our models can learn from that data. These contractors don’t understand how the AI models work and/or what’s the relationship between the data and the AI models. They’ll just use our tools to execute that work using their graphics cards.
The decentralized components
In order to function properly, all the data created by humans in these interactions needs to be open and decentralized. When someone executes an action, he will use his private key to sign his interaction and if he has the right to execute this action, it will be approved.
In each of these platforms, we need to create a governance system in order to know which person has the right to vote in the decentralized autonomous organization, or DAO.
We then need to create a roles and permissions system in order for the DAO to assign certain roles to users. Like giving the ability to a candidate to propose a new AI model in our ecosystem.
We also need to create a way to pay contractors. Nobody is going to invest their time and resources to our ecosystem if they don’t receive payments. Some payments needs to be direct, like receiving cash, and others needs to be shares in a common venture.
We also need to be able to elect custodians that will host email relays for our platforms and when they receive the order from our DAO to send an email, execute that order. When they receive emails from people, sync that email to our platform so our DAO can read the content of that email. IF a relay becomes rogue, we also need the DAO to vote to change relay easily and punish the rogue one.
Then, we need to create a platform to receive candidate’s résumé, have contractors classify them then contact them. At last, vote on who has the ability to execute certain tasks and assign roles to these candidates.
We’ll also need people that post contracting offers on different platforms on the web, like indeed. Therefore, we’ll need to sync the data from these platforms into our blockchain-based databases.
At last, we’ll need to have a referral system so that people can refer us candidates and receive payments when their referrals execute work in our ecosystem.
Bridges with existing platforms
When we can order resources from other networks, we will create bridges between our ecosystem and other ecosystems. One of these bridge will be with the Golem Network : they already built a marketplace to buy/sell processing power so we will be able to connect to them in order to process our AI models.
Conclusion
Like you can see, the human interactions in our ecosystem can be quite complex. There is a lot of tools to build and/or release, all of them decentralized. They’ll also be open source.
The good news is that most of these tools have already been programmed and just needs to be released.
What’s next
Most of the LegitDAO’s ecosystem connects with the referral system, the currency that the DAO uses and its governance system.
Also, LegitDAO’s already contain around 300 members because these people put resources together, bitcoins to be more specific, in order to launch this system. Some members also referred others and we manually collected that data.
At launch, the initial units of currency of the DAO will be assigned to the DAO members. Each member will also have one (1) governance unit in order to vote in the DAO. At last, if a DAO member that has been referred by another member interact in the ecosystem in a commercial manner, his referrer will receive compensation. Therefore, the referral system will contain initial referrals at its genesis.
My next post will contain all the details regarding how the currency , the governance system and the referrals system works and interacts together.
The post after that will launch that small ecosystem live with the first DAO graphical user interface, or GUI, so its users can interact with the decentralized ecosystem easily.
Stay tuned!
LegitDAO is social
If you want to join the LegitDAO social networks, you can find them on…