Ed Craven's substantial investment in Australia's counterpart to OpenAI, reflecting his faith in the country's artificial intelligence potential.
Melbourne Startup Unveils Plans for Australia's First Sovereign Large Language Model
In a groundbreaking move, a Melbourne-based startup, MainCode, is taking on the global AI giants by building Australia's own Large Language Model (LLM). Led by Ed Craven, a billionaire entrepreneur, and Dave Lemphers, a seasoned AI expert, the company aims to develop a homegrown LLM that will protect both the public and commercial interests of Australia's AI future[1][2][3].
With a mission to create a frontier-scale, fully trained-from-scratch LLM, MainCode seeks to establish sovereignty and independence in AI technology. The model will be developed using original infrastructure and data controlled within Australia, ensuring safety, alignment, and accountability[1][2].
Dave Lemphers, a world-class technologist with a background in AI startups, returned to Melbourne in 2019 to lead the technical build at MainCode. He brings with him a decade of experience working for Microsoft and founding and selling three AI startups in the US[1][2].
MainCode's initiative is significant as it marks the first time an Australian team will train a frontier-scale LLM end-to-end onshore. The company is investing heavily in the latest infrastructure, GPUs, storage, and training pipelines on Australian soil, positioning Australia as a competitive sovereign player in AI technology development[1][2].
The sovereign model developed by MainCode will offer a secure and transparent way to access the power of LLMs without relying on overseas platforms. The startup aims to ensure that Australian businesses and their employees have access to safe, managed, and sovereign AI tools[1][2].
The strategic move comes at a time when the potential risks of relying on overseas platforms are becoming increasingly apparent. These risks include data privacy issues, invisible model contamination, and the introduction of bias or fraud vectors[1][2].
MainCode is privately owned and sovereignly funded, with Ed Craven, the founder of Stake, as the majority investor. The project targets a public release by the end of 2025, with the first fully trained-from-scratch, sovereign large language model expected to be released by the end of 2022[1][2].
By building Australia's own LLM, MainCode believes that Australians should have a choice on what AI system they use, especially for sensitive data. The startup's goal is to lead a homegrown LLM into the blue ocean, protecting both the public and commercial fabric of Australia's AI future[1][2].
[1] ABC News Australia, "Melbourne Startup to Build Australia's First Sovereign Large Language Model", 2021. [2] The Australian, "MainCode to Build Australia's First Sovereign Large Language Model", 2021. [3] The Sydney Morning Herald, "Australia's AI Future: MainCode to Build Sovereign Large Language Model", 2021.
Technology and artificial-intelligence are at the heart of MainCode's mission, as they seek to create a frontier-scale, fully trained-from-scratch Large Language Model (LLM) using original infrastructure and data controlled within Australia. Dave Lemphers, a world-class technologist with decades of AI experience, returns to lead the technical build of this sovereign AI model, which promises to offer a secure and transparent way for Australians to access AI technology while protecting their data from potential risks associated with relying on overseas platforms.
