What is Gensyn?
Gensyn is a hyperscale, cost-efficient compute protocol for deep learning models. The machine learning compute protocol seeks to unite the world’s computing power into a global supercluster that is accessible to anyone. The vision of Gensyn is to disintermediate cloud oligopolists like AWS and Google by lowering the cost of compute.
Gensyn can be described as a blockchain-based marketplace for computing power. It connects developers seeking to train their machine learning (ML) model with compute providers. The protocol tapps into the idle machine learning resources around the world and will potentially be able to increase the available ML compute power by 10-100x.
The following provides an introduction to Gensyn.
The recent rise in applications of artificial intelligence (AI) has sharply increased the demand for computing power needed to run the necessary algorithms. With the computational complexity of AI systems expected to double every 3 months, the resource requirements for building and using such models are becoming excessively high, thus introducing significant access barriers.
With the resource-related access barriers limiting AI advancement and usage, novel ways need to be found to provide access to compute in an inexpensive manner. Present day solutions such as AWS are oftentimes very expensive because of the market’s oligopolistic structure.
These challenges are further complicated by existing providers not being able to cost-efficiently leverage all the available compute of the network as well as shortages of critical components such as semiconductors.
Gensyn is a layer-1 protocol that seeks to provide deep learning computational resources. The protocol rewards supply-side participants that provide their computing power to the network. This compute can then be used by Gensyns customers seeking to perform machine learning (ML) tasks.
Smart contracts are being used by the Gensny protocol to facilitate the programmatic distribution of compute tasks and payments in the network. The fundamental challenge of verifying completed ML work is approached by an intersection of cryptography, game theory, optimisation, and complexity theory.
