Decentralized AI: GaiaNet’s model building platform

The artificial intelligence boom is undergoing a profound transformation with the emergence of decentralized AI, led by innovative startups like GaiaNet Inc.

“In a centralized AI infrastructure, there is a lot of vendor lock-in, censorship, and unbalanced economics,” said Matt Wright (pictured), co-founder and CEO of GaiaNet. “We’re basically enabling people to take a system, wrap their own data in it, run their own AI model, and choose to monetize the services on top of that. We’re leveraging the best of open source and the power of Web3, as well as open source developer tools, to build your own AI agent.”

Wright spoke with Paul Gillin of theCUBE Research at the “Supercloud 7: Get Ready for the Next Data Platform” event during an exclusive broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed how GaiaNet offers an alternative platform for model inference and decentralized AI development.

Run decentralized AI agents while protecting data

In May, GaiaNet announced it had received $10 million in funding for its decentralized AI technology model, which the startup hopes will appeal to business units that may be hesitant to build AI infrastructure without being able to control the distribution of critical data.

“If a bank or a large financial services company wants to develop its own AI or use something like OpenAI, there is a risk of bringing data into a more centralized ecosystem,” Wright explained. “We enable any company to become an AI company. You can take our open source software, plug it into your vectorized embedding and essentially run your own agent, whereas a lot of other approaches are more about letting you work within their system, within their platform.”

GaiaNet uses node-based blockchain technology as part of its protocol for users. The decentralized nature of blockchain was a key factor in the company’s decision to use it, according to Wright.

“We think it’s critical that we have Web3 to validate that ecosystem and it’s very important to put some kind of stake or collateral on that computation,” he said. “In centralized models, a lot of the computation is given to a small group of people who decide how to share that cost. With Web3, you can actually have a network of participants who offload a lot of that cost or offset all of that cost and share it with the community.”

Attractive business model for entrepreneurs

“You can monetize that data or actually own the intellectual property of that data, and this is a way for you to actually build businesses on top of that AI,” Wright noted. “Whether you’re an app developer, development company, or a commercial company, you can essentially build use cases with our infrastructure and create opportunities to monetize for the end user who needs that kind of insight or a human-readable data format.”

Wright has already seen new use cases emerge for GaiaNet’s decentralized AI development model. One of them involved a 20-year professor at the University of California, Berkeley, who faced an interesting challenge when his graduate students wanted to collaborate with him on previous curricula and work.

“He didn’t have the bandwidth to hire a team of teaching assistants and wanted to build that with an autonomous agent,” Wright said. “With our infrastructure, the development tools we have at our disposal and the large language models, which again are open source, he trained his AI on his previous curriculum. Now he almost has a digital twin of himself and since he owns the intellectual property he created, he wants to monetize it over time.”

GaiaNet also benefited from the wealth of open-source AI models that have flooded into development over the past year, including models that Meta Platforms Inc. has made available to the open-source community.

“From the beginning, we’ve had people like Meta contributing to Llama 3… and without their contribution to open source, we couldn’t do what we do,” Wright said. “We’re really optimistic about making large language models, small language models, and AI learning content available to the community.”

Stay tuned for the full video interview, part of SiliconANGLE and theCUBE Research’s coverage of Supercloud 7: Get Ready for the Next Data Platform Event.

Photo: SiliconANGLE