A press release on August 29 revealed that FLock.io and IO.net collaborated to bring the first reported Proof of AI consensus algorithm. The algorithm will enable networks to verify the integrity of nodes working in their ecosystems. The consensus mechanism is an AI-native Proof of Work algorithm.
The PoAI consensus mechanism allows decentralized physical infrastructure networks (DePIN) to verify nodes by completing complex AI training tasks. The mechanism is notably AI-native, directing resources toward meaningful AI tasks. Nodes can also earn rewards from IO.net and FLock.io, which are DePIN and an AI training network, respectively.
FLock.io is a decentralized AI training platform that focuses on federated learning. The platform incentivizes its community for democratized model creation, development, and alignment. FLock.io has collaborated with other companies, including Chainbase, Ritual, and more, to promote its AI training.
Proof of AI to verify legitimate nodes in a decentralized network
The PoAI consensus mechanism enables decentralized AI platforms to verify the quality of computing from their nodes. The verification ensures honest and high-quality compute AI training inputs. PoAI quantifies DePIN nodes’ integrity in a decentralized and AI-native way.
FLock.io continuously provides challenges that nodes need to complete. The challenges generate data through inference on an LLM. The data specifies behavior from GPU devices based on factors such as Teraflops and Video RAM. Some necessary stats produced include latency, score deviation, and data correctness, supplied to IO.net nodes to compile judgments.
FLock.io CEO Jiahao Sun mentioned that AI engineers and end users need to trust the quality of the compute resources they receive. Sun believes that PoAI is the key to achieving that.
“Compute underpins the entire AI development process, which is why we’re starting with that, and we are delighted to join forces with io.net, a true leader in its field.”
Jiahao Sun, FLock.io CEO
FLock.io looks at long-term AI development
The PoAI consensus mechanism connects FLock.io’s decentralized training platform to DePIN platforms. FLock.io believes that in the long term, the PoAI verification will extend to AI training and federated learning. The platform’s team is also discussing the possibility of a dedicated AI chain.
IO.net’s CEO and co-founder, Tory Green, predicted that PoAI’s arrival would lead to massive improvements in AI model training and inference over decentralized compute networks. Tory also mentioned that he was confident that GPU node operators and the AI/ML developer community would welcome the Proof of AI mechanism.
Both CEOs confirmed the critical role decentralized AI GPU networks play in the long-term success of decentralized AI. The platforms still recognize how many nodes try to outsmart the systems using methods such as maliciously tricking the system into believing that they have more computing resources than they do.
FLock.io and IO.net agree that robust deterrence measures are necessary to prevent bad actors from taking advantage of the network rewards. The platforms also recognize the issue of solving node integrity. Both teams believe that the Proof of AI consensus mechanism will be a major breakthrough in solving this issue while creating a foundation for long-term decentralized AI advancement.
