There is a new Crypto project aimed at building distributed networks for unreliable AI output verification. It is called Mira, released last year and is still in the Testnet phase.
The cryptographic network is centered around Mira’s AI
The project was launched by three AI experts: Aroha Labs, Ninad Naik, Sidhartha Doddipalli and Karan Sirdesai.
In particular, Ninad Naik was previously an AI leader for Uber and Amazon. Mira He holds the position of Chief Product Officer. He has extensive experience in creating AI platforms.
At Uber, he led the development of major market products for the company’s global food and grocery delivery business; Mira He leads product development and research to enable developers and businesses to leverage artificial intelligence in new and impactful ways.
Mira’s mission is Supply Chain AI.
The aim is to establish new standards for how AI is developed, used and maintained, with a particular emphasis on minimizing the complexity of Web2 and Web3 developers.
The project is still in the development stage and is open to developers, creators and researchers who want to explore the potential of infrastructure and participate in the ecosystem.
How does AI cryptographic networks with MIRA work?
The starting point is Pragmatic AI output.
In fact, artificial intelligence often produces outputs containing false information due to the stochastic nature of technologies based on neural networks.
MIRA networks aim to be a decentralized consensus, or unreliable platform, with AI-generated output.
In other words, verifiers are not reliable, instead of having to resort to verifiers who need to place trust.
Therefore, the output of AI is converted into statements that can be verified independently by the user, allowing different AI models to collectively determine the validity of each statement.
Node operators in the network performing these inference-based validation are economically incentivized through hybrid demonstration/empirical mechanisms to induce honest validation.
” `html
Problems and solutions proposed by Mira
” `
Today’s AI systems are great at generating creative and plausible outputs, but they cannot provide error-free output in a verifiable and reliable way.
Therefore, although human supervision is currently required for verification, distributed verification networks are
Mira wants to achieve the same results with a new combination of content consensus made possible by content transformation and cryptoeconomic incentives. This should result in unrealistic operations, both technically and economically.
In addition to validation, the vision of the MIRA project is to achieve a synthetic foundation model that integrates validation directly into the generation process.
Such a simplified approach eliminates the distinction between generation and validation and provides an already error-free output.
Additionally, distributing validation across a distributed network of incentive operators creates an infrastructure that is inherently resistant to centralized control.
In a white paper they write:
“This represents a fundamental advancement. By enabling artificial intelligence systems to operate without human supervision, it lays the foundation for true artificial intelligence, a critical step in unlocking the potential for transformation of artificial intelligence in society.”
” `html
Mira Project Token
” `
Currently, the MIRA network is still on the testnet. Therefore, since the mainnet does not exist yet, its cryptocurrency does not exist yet either.
Nevertheless, in the Crypto market there are already some tokens along with this name or this ticker, but they have nothing to do with the MIRA network projects by Sirdesai, Doddipalli and Naik.
At this point, despite the initial stages of fundraising that ended with a $700,000 collection last year, there is no official roadmap even to know when it will be released.
The main investors are Bitkraft Ventures and Framework Venturesbut Accel, Mechanism Capital, Folius Ventures, Aj Scaramucci Funds I also participated in fundraising.
