As a seasoned crypto investor with a keen eye for potential, I find Near Protocol’s ambitious AI project to be nothing short of intriguing. Having witnessed the meteoric rise of ChatGPT and its transformative impact, I can only imagine the groundbreaking advancements that could stem from this 1.4 trillion parameter model.

---

At the inauguration of the Redacted conference held in bustling Bangkok, Thailand, I am thrilled to share that Near Protocol has announced an ambitious project aimed at constructing the globe’s most expansive open-source artificial intelligence model. This colossal 1.4 trillion parameter model will dwarf Meta’s current open-source Llama model by a factor of 3.5, pushing the boundaries of what we know about AI development and capabilities.

Through a collaborative approach involving numerous contributors on the recently launched Near AI Research platform, we’ll be developing and refining the project by means of competitive crowdsourcing. Starting as early as November 10th, participants have the opportunity to participate in training a compact 500 million parameter model.

Near Protocol’s ambitious AI model

Over a series of seven stages, our project will expand in scale and complexity. Only top-tier collaborators will transition to working on increasingly intricate and larger models. These models will generate revenue, while privacy will be safeguarded through encrypted Trusted Execution Environments. This setup aims to incentivize continuous updates as the technology evolves.

At the Redacted conference in Bangkok, Near Protocol’s co-founder, Illia Polosukhin, shared that they plan to finance their costly training and computations through token sales, as reported by CryptoMoon.

As a researcher, I find myself in a position where I’m dealing with funds amounting to approximately $160 million. Clearly, this is a substantial sum, but within the realm of cryptocurrencies, such an amount can potentially be raised.

“Then the tokenholders get repaid from all the inferences that happen when this model is used. So we have a business model, we have a way to monetize it, we have a way to raise money, and we have a way to put this in a loop. And so people can actually reinvest back into the next model as well.”

Related: Near patches critical bug that could crash every node on the network

One of only a few cryptocurrency projects capable of executing such an ambitious endeavor is this one: Polosukhin, who co-authored the pioneering research paper on transformers that paved the way for ChatGPT, and co-founder Alex Skidanov, previously at OpenAI during the development and release of the epoch-making model in late 2022.

Skidanov, the current leader at Near AI, admitted that it’s a significant challenge, requiring a substantial obstacle to be surmounted.

Decentralized AI tackles privacy issues

For this project to effectively train such a massive model, we’d ideally need “a significant number of tens of thousands of GPUs, all located in one spot.” This setup isn’t optimal. Alternatively, utilizing a dispersed network of computational resources would be required, but currently, the technology necessary for this doesn’t exist yet. The reason being that existing distributed training methods demand exceptionally fast interconnect. However, he mentioned that recent research from DeepMind indicates it might be achievable.

Polosukhin mentioned that he hasn’t had conversations with existing initiatives such as the Artificial Superintelligence Alliance, but is open to exploring potential alignments. Regardless of the outcome, he stressed that it is crucial for decentralized AI technology to prevail for our collective benefit. As a conference guest speaker, Edward Snowden underscored this idea by depicting a chilling image of centralized AI transforming the world into a global surveillance state.

He pointed out that the current and upcoming technology we’re dealing with might be of paramount significance, and a concerning fact is that if Artificial Intelligence is monopolized by a single entity, in essence, they will have the power to dictate our actions.” Snowden further stated:

“If all AI and effectively, all of the economy is being done by one company, there’s no decentralization at that point. So it is effectively like the only way that Web3 is still relevant, philosophically, if we have AI that also follows the same principles.”

Read More

• GBP EUR PREDICTION
• SEI PREDICTION. SEI cryptocurrency
• CNY RUB PREDICTION
• HBAR PREDICTION. HBAR cryptocurrency
• TRB PREDICTION. TRB cryptocurrency
• HOOK PREDICTION. HOOK cryptocurrency
• MNT PREDICTION. MNT cryptocurrency
• CTXC PREDICTION. CTXC cryptocurrency
• BCH PREDICTION. BCH cryptocurrency
• INJ PREDICTION. INJ cryptocurrency

2024-11-10 15:33