FLock.io’s Aptos LLM outperforms ChatGPT-4o in generating Move-specific code, setting new standards in decentralized programming tools and allowing developers to produce secure, high-quality Move code with unprecedented efficiency.
London, UK – October 24, 2024 – FLock.io, a decentralized AI Model Training platform, has partnered with the Aptos Foundation to elevate the capability of large language models (LLMs) in generating accurate code for the Move programming language on the Aptos network. This collaboration enables superior AI-assisted development in Move, enhancing Aptos' unique resource-oriented language for handling secure digital assets.
By leveraging a dataset of community-created Move code and its cutting-edge decentralized AI training capabilities, FLock.io empowers Aptos developers and the broader engineering communities to develop sophisticated models that address Aptos-specific nuances in Move, significantly enhancing code readability and interpretability.
The initial version of FLock.io’s Aptos Move-tuned LLM demonstrated superior performance in testing compared to ChatGPT-4o. This was evident across a range of tasks, from simple "Hello World" scripts to more complex programs such as yield tokenization with AMM trading functionality. Additionally, the FLock model exhibited greater accuracy and adherence to Aptos-specific requirements.
"Our integration with Aptos underscores the tremendous value of decentralized AI,” said Jiahao Sun, CEO of Flock.io. “Notably, the model's superior performance compared to GPT-4o demonstrates that community co-creation, especially in specialized and vertical use cases, can unlock unparalleled potential and groundbreaking innovation. FLock stands at the forefront of this decentralization revolution in AI, where collaborative efforts yield more intelligent and specialized models that surpass the capabilities of centralized systems."
The next steps include the expansion of FLock.io’s dataset with expert annotations, as well as the establishment of benchmarks to further evaluate LLM performance on Move code. In the coming months, FLock.io plans to deploy a production-ready model tailored for Aptos developers. This model will be hosted on FLock.io’s federated learning platform, enabling decentralized training and validation to continuously improve its capabilities with more proprietary Move codebases and adapt to new requirements in real time.
"Move on Aptos is widely known for its security and efficiency, but mainstream LLMs haven't yet offered strong support for it,” said Bashar Lazaar, Head of Grants & Ecosystem at Aptos Foundation. “By developing an LLM specifically tuned for Move on Aptos, we are not only advancing decentralized AI but also simplifying the use of Aptos, making decentralized technology more accessible to everyone."
About FLock.io
FLock.io is a community-driven platform facilitating the creation of on-chain, decentralized AI models. By combining federated learning with blockchain technology, FLock.io offers a secure and collaborative environment for model training, ensuring data privacy and transparency. FLock.io’s ecosystem supports a diverse range of participants, including data providers, task creators, and AI developers, incentivizing engagement through its native FML token. Learn more at FLock.io.
About Aptos
Aptos is the secure, scalable, and feature-rich L1 blockchain of choice for both developers and users—delivering the best performance, the highest throughput, and lowest latency. Aptos is the first blockchain to use the Move programming language, and is designed with simplicity in mind for the best builder experience. Aptos is the top choice for next-gen use cases, real-world applications, and the millions of users that come with them. If it’s happening in Web3, it’s happening on Aptos.
About Aptos Foundation
The Aptos Foundation is an organization dedicated to supporting the growth and development of the Aptos protocol, decentralized network, and developer ecosystem. We lead key blockchain and ecosystem initiatives, such as: fostering the growth of the network, cultivating a dynamic ecosystem, and attracting developers to continually improve the infrastructure of the Aptos blockchain.
