London, United Kingdom, May 10th, 2023, Chainwire
Today, PUBLC announced the paradigm-shifting integration of Large Language Models (LLMs) into its search engine, a transformative move that firmly positions the startup as a pioneer in the rapidly evolving search industry. This milestone follows PUBLC’s ongoing mission to create “everyone’s search engine” – a platform built by the web community, powered by cutting-edge AI models, and anchored in a blockchain-based token economy through its token, PUBLX.
Since the release of ChatGPT 6 months ago, the world has witnessed the revolutionary capabilities of LLMs, prompting tech giants like Google and Microsoft to rush to incorporate them into their products. Today, PUBLC stands out as the first search engine to fully integrate LLMs into its core search logic. This integration allows every search query on PUBLC to be processed by the LLM, which retrieves relevant data and sources, delivering answers and insights in natural language. These features further enhance PUBLC’s unique experience by offering users a more engaging, personalized, and visual approach to searching, discovering, and consuming web content.
The integration is also in line with PUBLC’s community-driven approach, which combines advanced AI models with human intelligence. This approach envisions a system where every user can contribute to training the Language Models, following a methodology similar to Reinforcement Learning from Human Feedback (RLHF) used in the development of ChatGPT. Drawing inspiration from collaborative models like Wikipedia, PUBLC aims to engage millions of users worldwide in organizing web content, transforming it into an invaluable resource for answers on any topic or niche.
Committed to creating a search engine that belongs to and rewards the web community, PUBLC challenges the status quo wherein tech giants such as Google, Meta and Amazon dominate global ad revenue, leaving little for the broader web ecosystem. To address this imbalance, PUBLC has implemented a tokenized revenue model that distributes 90% of the platform’s revenue among its community using its Ethereum blockchain-based token, PUBLX. This blockchain-based economic model is tied to the platform’s fiat-based revenue streams, like advertising, giving the token real value and fostering ongoing growth and collaboration. This economic model lays the foundation for a new web economy that rewards users, creators, content platforms, celebrities, developers, and other participants, giving them the value they rightfully deserve.
As a next step, PUBLC will open-source its AI layer, developed on the popular LLM framework, LangChain, while actively embracing the expanding LLM developer community. By doing so, PUBLC will enable developers worldwide to contribute to its AI models and receive rewards through its token model. PUBLC invites LLM developers, open-source projects, and communities to join its mission of creating an open-source LLM project that serves the web ecosystem.
Lior Davidovitch, Founder and CEO of PUBLC, said: “This launch represents PUBLC’s most significant milestone to date. As a modern-day David versus Goliath, our small team has been dedicated to building this ambitious vision layer by layer over the past few years, aiming to revolutionize the search industry and compete with the world’s largest companies. This is just the start, and we plan to launch many features and initiatives in the upcoming months to support our community. We call the web community of users, creators, brands and developers to join us on this mission of creating a search engine that genuinely belongs to everyone.”
About PUBLC
PUBLC is on a mission to revolutionize the search industry by creating “everyone’s search engine” – a platform that is collaboratively built by the web community, powered by state-of-the-art AI models, and anchored in a blockchain-based token economy that benefits the entire web ecosystem.
Check out the PUBLC platform, learn more about the project.
Find $PUBLX on CoinGecko.
Contact
Head of Product
Lora Salomidou
PUBLC
[email protected]
