Q&A with Ally Haire: Tackling Compute Power Shortages – Why Decentralized AI is the Future

Q&A with Ally Haire: Tackling Compute Power Shortages – Why Decentralized AI is the Future

The rapid AI development suggests a linear trajectory, similar to the evolution of the internet. However, various forces could disrupt this path, hampering technology development or limiting access. A significant and imminent risk is that demand for computing power may soon exceed supply. Our World in Data indicates an exponential increase in compute requirements for AI algorithms.

In this context, Korea IT Times interviewed Ally Haire, CEO and Co-Founder of Lilypad, a decentralized compute network for AI and other computational processes. Ally, with extensive experience as a Lead Developer Advocate at IBM and a Senior Software Engineer at the University of Sydney, provides insights into the capacity issues faced by AI players and the need for a decentralized approach.

Here are the key insights from our conversation with Ally Haire.
 

 Ally Haire, CEO and Founder of Lilypad

 

Please tell me about Lilypad and the mission and vision behind the project. 

Lilypad is a distributed, permissionless compute network harnessing idle processing power to fuel a decentralized artificial intelligence (AI) stack. The Lilypad team champions compute as the cornerstone of an open web and AI framework striving to unlock access to a universal compute layer as a public good. Developers will be able to seamlessly execute internet-scale computation tasks from smart contracts (on any Ethereum Virtual Machine (EVM)-compatible chain), the command line interface (CLI), or traditional SaaS workflows. 

Lilypad is developed and deployed by a passionate team of cross-functional developers from Web2, Web3, and crypto powerhouses like IBM, AWS, Protocol Labs, Coinbase, Near, Google, and Atlassian, as well as researchers from Stanford University. Our diverse team combines years of business operations and research experience with developer expertise, and our aim is to democratize access to high-performance computing through a permissionless, distributed multi-chain network. With a team that specialises in startup support, distributed system execution, AI research, and Web3 development, Lilypad is designed as a practical, cost-effective solution for compute-intensive activities.   

Why do you think it’s important to make compute accessible and decentralized?

Web2 cloud computing remains one of the largest and fastest-growing corners of the tech industry. However, with the recent explosion in AI research and development, access to compute has become somewhat limited. Our main focus here at Lilypad is to counteract this scenario by creating a distributed network that can run very large scale compute that, in theory, everyone can contribute to. This means going beyond running a little bit of serverless code, and instead enabling users to run really resource-intensive tasks like AI inference, for example.

There is a lot in the news about bottlenecks around AI development due to a lack of compute, but what isn’t discussed as much is the fact that a lot of people run their own servers and run their own hardware, which could be leveraged to fulfill some of this demand. That’s why at Lilypad we are building a peer-to-peer network of sorts that will enable users to contribute their own GPU supplies for others to leverage (while earning rewards), with the ultimate aim of making compute more accessible and cheaper. Lilypad’s infrastructure synchronizes the idle GPU power supplied by users with user job requests, allowing GPU providers to earn Lilybit_ rewards, while contributing compute power to the research and development of hundreds of incredible projects in decentralized science (DeSci), decentralized finance (DeFi), gaming and more.

So, Lilypad is, in essence, a compute coordination platform and operates as the backend for compute where users, protocols, providers and projects can interact and feed into an open infrastructure stack to extend global access to efficient compute power. 

What have been some successful use cases of Lilypad?

There are a few successful use cases of Lilypad, one being Waterlily.ai. Waterlily.ai is a proof-of-concept application aiming to provide an alternative ethical solution to AI-art generation. This is done by creating a new paradigm for AI image generation that pays royalties to the original creator and offers a new revenue stream for artists to explore. Waterlily introduces ethical, efficient, and equitable rail alternatives for the AI-generated art world. Waterlily leverages the transparency and trustlessness of blockchain, with smart contracts, combined with a decentralized compute platform in Lilypad, to deliver an intuitive user experience for generative AI-art while compensating creators for their original work with every image generation call.

Another particularly interesting one is Rare Compute, which is a democratic, non-profit consortium leveraging Lilypad to harness the power of GPU-driven technology to unravel the mysteries of rare genetic diseases. Acting as node providers for Lilypad, Rare Compute prioritizes understanding rare diseases, aiming to improve countless lives and fuel groundbreaking discoveries in human biology while democratizing access to compute power for researchers the world over. It is really quite exciting for us at Lilypad to play a role in such innovative tech development that has the potential to reshape the healthcare industry.

What is in store for Lilypad development? What does the road map look like?

We recently launched the Lilypad Incentivised Testnet (IncentiveNet) after a year of working towards a robust protocol model! Lilypad’s IncentiveNet rewards those who contribute GPU to their network and coordinates workloads to ensure GPUs are utilized to their full potential. As mentioned previously, Lilypad’s infrastructure allows GPU providers to earn Lilybit_ rewards, while contributing compute power to the research and development of hundreds of exciting new projects. 

Participants are encouraged to contribute to network stability and establish sufficient nodes at launch. This entails thorough testing to ensure resilience, detect bugs and vulnerabilities, and validate operational and game theory logic. Participants will be rewarded in the form of Lilybit_ credits, which will be redeemable for the Lilypad ERC20 Token at Mainnet Launch. 

So with that in mind, our focus for the next coming months will be onboarding GPU providers and working tirelessly toward our Mainnet Launch – so watch this space. You can join our community and find any details of our upcoming announcements on our social channels – Discord and Twitter. 

저작권자 © Korea IT Times 무단전재 및 재배포 금지

Originally Appeared Here