Will NEAR Blockchain Lead the Breakthrough of AI?
According to Bloomberg, Llion Jones, one of the eight co-authors of Google’s most influential paper “Attention is All You Need” , has departed from the company to embark on a startup venture, making him the final author to leave the team. This wave of departures commenced with Illia Polosukhin’s pioneering journey to co-found NEAR Protocol, which was subsequently followed by the remaining key AI researchers.
The Departures from Google Started the New Future for AI
In December 2017, eight Google researchers co-authored a pioneering AI paper titled “Attention is All You Need”. The paper introduced a new model architecture called the Transformer, which relies solely on attention mechanisms instead of RNNs or CNNs. Transformers are now key building blocks in the field of Natural Language Processing (NLP) and are widely used in tasks such as machine translation, text summarization, and question-answering systems. Transformers is the technology that underpins popular AI products such as OpenAI’s ChatGPT.
Illia Polosukhin was the earliest of the eight to leave Google. Illia was a major code contributor to Google’s TensorFlow AI open-source project and served as the project lead for Google’s deep learning group. In 2017, he co-founded the AI company NEAR.AI with Alex Skidanov, and developed NEAR.AI into the blockchain network NEAR Protocol.
NEAR has a world-class team. The founding team came from Google, Facebook, and Microsoft with extensive experience in building large scale distributed systems. The engineering and research teams consisted of multiple champions of global programming competitions such as ACM-ICPC, Google Code Jam, and TopCoder Open, etc.. The business teams include serial entrepreneurs and experienced professionals who worked with some of the largest enterprises in the world.
NEAR is backed by leading venture capital firms including a16z, Electric Capital, Dragonfly Capital, Pantera Capital, Coinbase Ventures, and Baidu Ventures. NEAR currently ranks 38th globally with a market cap of $1.4 billion.
The other seven researchers who co-authored the paper have since left Google successively. Almost all of them have gone on to launch AI startups; Aidan Gomez founded Cohere, which offers large language models (LLMs) to corporate customers. Noam Shazeer, the longest-serving Googler in the group who was seen as an AI legend, launched Combined.AI, which lets users create chatbots emulating celebrities and historical figures. Niki Parmar and Ashish Vaswani co-founded Adept, a startup training AI to use existing software and API. Jakob Uszkoreit co-founded Inceptive, which specializes in AI life sciences and is dedicated to designing the next generation of RNA molecules using neural networks and high-throughput experiments. In 2021, Lukasz Kaiser left Google to join its rival OpenAI. Kaiser was one of the co-creators of TensorFlow, Google’s open-source ML platform. The recent announcement of Jones’s departure officially marks the end of an era at Google. Collectively, their ventures have secured approximately $1.5 billion in investments from notable entities such as Nvidia, Tiger Global, Oracle, a16z, and others, culminating in a total valuation exceeding $4 billion.
NEAR is Set to Become the Data Infrastructure for AI
The majority of the authors who contributed to the paper pursued their path in the field of artificial intelligence. However, the first to depart, Illia Polosukhin, embarked on a business venture in the blockchain field and attained remarkable accomplishments. Illia’s initial entrepreneurial direction was also towards AI. However, during the startup period, he realized that the collection, pricing and sharing of data remain as a major obstacle to the evolution of AI. He realized that blockchain and smart contracts are the best solution to solve the data challenge in machine learning; he believed that blockchain can become the cornerstone infrastructure for the future development of AI. Consequently, Illia and Alex decided to repivot from AI to blockchain.
NEAR is constructing a revolutionary blockchain operating system that pioneers an industry-first category: a common layer for browsing and discovering open web experiences, compatible with any blockchain. This establishes NEAR as the primary gateway for users and developers, enabling seamless access and operation of both Web3 and Web2. Currently, the number of NEAR accounts has surpassed 25 million, with a daily transaction volume reaching 446,000. The platform also hosts over 1,000 DApp projects and continues to grow steadily.
Although NEAR has gained significant traction in blockchain, it remains closely intertwined with AI.
“Blockchains can help coordinate decentralized groups of contributors around doing some work according to a set of rules, so they have a role to play in governance of AI as well as globally sourcing contributions to training models.”
Illia states as he takes a deep dive into AI on NEAR. Illia also elaborates on a few use cases:
- Crowdsourced data labeling: a marketplace to reward contributors for high quality data labeling that trains a model
- Governance: managing and monitoring AI transparently and proactively, such as creating alignment dataset via crowdsourcing, curating public datasets for evaluation of models, directing research funding, and providing public and traceable research discussions
- Content management system & on-chain reputation: provenance tracking of all media or content assets when they are created: text, articles, a quote, images, videos, in order to resolve problems of fake attributions, misaligned or illicit content. Authenticated content provides traceability as users interact with data over time, which also offers a more powerful user control over content they want to see with community governance and on-chain reputation system
- Language defined experiences: a user can ask an AI to create a frontend for them combining different components and features. Automating this process will make development accessible to anyone with an account.
- Marketplaces: Allow interaction with LLMs and deep learning models overall more accessible while still allowing monetization around models that are expensive to run and train.
Co-Founder Alex Launches an AI Data Platform based on NEAR
NEAR’s co-founder, Alex Skidanov, is also an expert in AI and machine learning. He won the Gold Medal in the world’s most challenging collegiate level programming competition, ACM-ICPC, and worked for Microsoft and Open AI after graduation. Later on, Alex became the first engineer at SingleStore (MemSQL) and pioneered a sharded database model that served global giants such as Goldman Sachs, Uber, and Samsung.
After completing the research and development of multiple breakthrough technologies within NEAR, such as the renowned Nightshade sharding mechanism, Alex redirected his attention towards AI and founded an AI startup called Jasnah. Jasnah focuses on building smarter large language models, or LLMs, to advance the accelerating field of AI. LLMs typically rely on three main components: data, compute, and model architectures. Jasnah is building LLMs that solve two persistent problems in the field: learning complex topics like logic, math, and science, and identifying when its outputs are factually incorrect. Their unique approach involves training these models in a decentralized way through contributions from a blockchain-based community, using NEAR Protocol.
NEAR is actively paving the advancement of the development of “AI + blockchain” technology; it is promoting various use cases related to AI and machine learning. Backed by a strong technical foundation and a thriving ecosystem of builders and partners, NEAR remains committed to the mission of onboarding billions of users to Web3.
About NEAR Protocol
NEAR is on a mission to onboard a billion users to the infinite possibilities of Web3 with the Blockchain Operating System (BOS). Leveraging its high-performance, carbon-neutral protocol, which is fast, secure, and scalable, NEAR provides a common layer for browsing and discovering the Open Web.
About MetaWeb Ventures
MetaWeb Ventures is a global venture capital firm focused on pre-seed and seed investments in Web3 start-ups. MetaWeb’s Fund 1 is backed by Sequoia Capital, Dragonfly Capital, SevenX, NEAR Foundation, GSR, SNZ and more.
