Pitch your startup story at [email protected] Please don't forget to join our ML Subreddit
The underlying concepts of “deep learning,” which evolved decades ago, are required for self-driving automobiles and other AI-based applications. This necessitates the development of training models based on patterns found in the human brain. This, in turn, necessitates a vast amount of compute capacity, which TPUs (tensor processing units) or GPUs (graphics processing units) running for long periods of time may provide. Most AI developers, however, cannot afford this compute power, and are forced to rent from cloud computing platforms such as AWS or Azure.
Gensyn, founded in 2020, claims to be a cost-effective alternative to this paradigm that also provides a bigger scale compute capacity. The Gensyn network, which is based in London, connects GPUs and pays their owners for running deep learning training jobs given by users. Gensyn has established a means to verify that deep learning tasks have been completed correctly using a blockchain network, triggering rewards via a token.
According to the researchers from the team at Gensyn, nearly half of the world’s potential computing power – from personal computers to data centers – is squandered due to idle hardware. Because the calculations are exclusively employed to maintain consensus and the lower level labor is effectively underutilized, mining proof-of-work cryptocurrencies now provide for a considerable return for part of this idle compute, albeit at a high environmental cost. The Gensyn network claims to use this computing power for beneficial work. Gensyn gives developers on-demand access to high-scale deep learning to compute a trustworthy source despite being decentralized, allowing them to monetize their unused compute resources in an ecologically conscientious manner.
In the recent seed funding round, the company raised around 6.5 million with the primary aim of enabling machine engineers to train AI models not only at a larger scale but also at a cheaper cost.
For Advertisement or Content Creation Service, Please Contact Us at [email protected] or check out our ad page here