Near Protocol is set to revolutionize the artificial intelligence (AI) scene with its ambitious announcement made during the opening day of the Redacted conference held recently in Bangkok, Thailand. The initiative is to develop the world’s largest open-source AI model, boasting a staggering 1.4 trillion parameters, which is approximately 3.5 times larger than Meta's popular open-source model, Llama. This groundbreaking endeavor marks a significant leap forward for Near Protocol, aiming to reshape not just AI but also the broader technological and economic landscapes.
Unlike typical projects, which might rely solely on internal development teams, the creation of this colossal AI model will be fueled by competitive crowdsourced efforts. Participants from all walks can contribute through the newly minted Near AI Research hub, engaging with the training of smaller models first, starting with one of 500 million parameters. This hands-on involvement is intended to not only democratize the development process but also drive innovation from the ground up.
Illia Polosukhin, co-founder of Near Protocol, elaborated on the blueprint for this project. Over the course of its evolution, the AI initiative will progress through seven distinct models, where only the most outstanding contributors will advance to tackle increasingly complex iterations. Polosukhin made it clear, “It costs about $160 million, so it’s a lot, obviously, but actually, in crypto, it is raiseable money.” This approach hints at the platform’s ability to capitalize on the expansive funding possibilities made available through token sales.
To meet this ambitious goal, Near Protocol doesn’t just throw money at the problem. Instead, they’ve carefully thought about how to structure contributions and reward those who get involved. The training and computations involved will occur within encrypted Trusted Execution Environments, ensuring privacy and security as well as incentivizing participants to refine and update the AI models continuously. This progressive, merit-based advancement keeps higher standards and goals front and center.
Thinking about the technological challenges involved, Polosukhin addressed the ambitious scale of the project. Training such behemoth models calls for substantial computational power – likely requiring “tens of thousands of GPUs” operating centrally. Yet, he acknowledged the inefficacies of such concentration. Utilizing decentralized networks for training models introduces new levels of complexity, as existing distributed training methods depend heavily on fast interconnectivity which the current technology might not fully support. Nevertheless, hints from cutting-edge research, particularly from DeepMind, suggest possibilities of achieving these decentralized computational systems in the future.
At the conference, renowned whistleblower and digital privacy advocate Edward Snowden stressed the importance of decentralized AI technologies. He warned of the perils of concentrated AI dominance which, as he pointed out, could lead to the rise of pervasive surveillance. Snowden shared, "If AI is controlled by one company, we effectively are going to do whatever the company says.” This sentiment echoed the core vision of Near Protocol as it aims to democratize technology development and defy centralized control.
While the ambitious scale of creating one of the largest AI models poses many challenges, it also presents tremendous opportunities. The available expertise within the Near Protocol team adds confidence to this initiative. Polosukhin was part of the groundbreaking research team whose work led to the development of ChatGPT, and Alex Skidanov, co-founder and current head of Near AI, has valuable experience from his time at OpenAI, which was pivotal to the release of cutting-edge models just last year.
All of this sets the stage for what should be a fascinating evolution, especially within the crypto community, where traditional barriers are often bent if not completely broken. The possibility for utilizing cryptocurrency for funding large-scale tech projects opens doors once thought impossible. Investors will have unique roles as various stakeholders, potentially benefiting from massive developments as they arise.
Summing up the gathered insights, it appears clear: Neart Protocol’s endeavor to construct this AI model could set precedents not just for AI, but for the interplay between decentralized technologies and socio-economic dynamics. The project seeks to actively involve participants worldwide, fostering collaboration and innovation rather than competition.
Near Protocol's vision extends beyond just building AI; it stands to reshape the narrative around digital sovereignty and user rights within these advanced technological landscapes. Achieving success will depend heavily on community engagement, technological breakthroughs, and the consistent push against centralized paradigms advocated by proponents like Snowden.
This venture is not merely technological; it’s deeply intertwined with ethical and social questions about control and power within the rapidly advancing field of AI. Near Protocol is positioning itself at the crossroads of these discussions, aiming to offer systems facilitated through mathematics rather than monopolistic control.
With the conference taking place and promises of revolutionary developments on the table, the tech world will certainly be keeping its eyes glued to Near Protocol as they set forth on this unprecedented quest for open-source AI greatness.