Recently, BitTensor (TAO) has gained significant attention after being incorporated into the Grayscale fund, thanks to its revolutionary Proof of Intelligence (POL) mining method. A discussion on Grayscale’s YouTube channel featuring BitTensor co-founder Jake explored the project’s background, structure, and future outlook. Below is a detailed summary of the conversation.
1. Background and Origins of BitTensor
Grayscale: What was the motivation behind founding BitTensor, and what problems were you aiming to solve?
Jake: I studied mathematics at Simon Fraser University, and since 2010, I’ve been deeply engaged in AI research. At that time, one of AI’s biggest issues was the lack of computational resources. The approach taken by companies like Google to train large AI models was centralized and inefficient. Inspired by Bitcoin’s decentralized network, I envisioned a system where distributed resources could be used to train AI models. This idea eventually evolved into BitTensor, where, much like Bitcoin, various nodes process data in a decentralized manner to train AI models.
Structure and Functionality of BitTensor
Grayscale: Can you explain how BitTensor is structured? What roles do subnetworks, miners, and verifiers play?
Jake: BitTensor operates through subnets, which consist of miners and verifiers. The miners provide the data necessary for training AI models, while verifiers ensure that the data is processed correctly. This process uses a Proof of Work (PoW) mechanism, where miners are rewarded for their contributions. We call this process Proof of Intelligence (POL). Each subnet has a specific objective function, and miners compete to achieve this goal as efficiently as possible.
Example of Subnet Operations
Grayscale: What kind of tasks do these subnets perform?
Jake: The subnets can perform a wide range of tasks. For instance, one subnet works on solving protein folding problems, where miners calculate protein structures, and verifiers evaluate the accuracy of those calculations. Another example is subnets that handle the training of large machine learning models, such as models with 80 billion parameters, which are used for data processing and prediction. Over time, these subnets are optimized for better performance.
2. Long-Term Vision and Common Misconceptions
Grayscale: What are the long-term goals of BitTensor? Where do you see it in 5-10 years?
Jake: BitTensor will become far more efficient than centralized AI research labs within the next 5-10 years. We’re building a distributed AI training system based on incentive mechanisms, and as time goes on, we’ll process more data and train better AI models. BitTensor will be a critical platform in AI research and development and will eventually surpass centralized AI organizations by leveraging the power of decentralized networks.
Grayscale: What are some of the common misunderstandings about BitTensor?
Jake: A lot of people mistakenly view BitTensor as just another cryptocurrency scam. But in reality, it’s an innovative decentralized machine learning platform. We’re the first to combine AI and blockchain to train decentralized AI models. While this concept is still new to many, I’m confident that with time, people will come to appreciate the significant role BitTensor plays in these fields.
Transparency and Fairness
Grayscale: BitTensor emphasizes transparency and fairness. How is this achieved?
Jake: Since BitTensor is based on blockchain technology, every action within the network is transparently recorded. This means anyone can participate fairly, and unlike centralized AI research labs, information is openly shared, allowing for fair competition among participants.
Conclusion
BitTensor (TAO) has garnered increased attention since being adopted into the Grayscale fund and praised by Vitalik Buterin for its potential in fusing AI with blockchain. This recognition has led to significant price surges across over 50 global exchanges, including Binance.
Recently, South Korea became the first to successfully establish a data center capable of mining BitTensor, signaling an important milestone in expanding its market presence beyond the U.S. and Europe. With much of the mined TAO tokens staked for validation, the circulating supply remains low. However, as Chinese miners prepare to enter the market, substantial growth in value is anticipated.
This conversation provided deep insights into how Jake views BitTensor's foundation, technical structure, model training processes, and its long-term vision, offering a glimpse into how the project will continue evolving as a leader in decentralized AI systems.
