We enter a new era where the ability to use and work with artificial and synthetic intelligence is a human right.
Access to intelligence – the power to innovate, work with and take advantage of higher levels of synthetic intelligence – belong to the people.
By building on still cheap calculated, abundant data and low costs, open source models, we are witnessing a synthetic intelligence cornucopia.
We need to build infrastructure that supports pluralistic development of AI. That’s why we start Thams Network, which is based on Oxford: a decentralized intelligent network to run at the edge, enabling private, censorship, depolitated and decentralized AI through built -in economic incentives and cryptographic evidence.
“I am concerned about the concentration of power and loss of privacy that AI leads to. It is important for us to think of strong technical solutions to this, such as blockchain, ”said main scientist Oxford professor Philip Torr, who called for incessant progress in decentralized AI and handed the AI power to the people.
Decentralization of AI
What does this open source, decentralized market, protocol and incentive layer for artificial and synthetic intelligence mean? The release of Deepseek has several implications, the most important of which is that the open source AI is here to stay and that the future does not turn into a large centralized, company’s (or state) model. AI has changed from the middle to the edge and it is in the future to become more decentralized.
Microsoft has just announced that distilled, NPU-optimized versions of Deepseek R1 will be available on PCs that benefit from one-unit, local treatment, starting with Qualcomm Snapdragon X first, followed by Intel Core Ultra 200V and others. Users will be able to interact with the newer family of groundbreaking models completely locally.
Neural treatment units are specialized computer microprocessors designed to mimic the treatment function of the human brain’s neural network. NPUs that will be displayed on personal devices offer a very effective set of capacities for modelinferencing that unlock the agent paradigm where generative AI can not only perform when directly invoked, but enables semi and fully continuous running Services, ie agents.
The movement towards decentralization is more than a technical upgrade. It represents a fundamental change in how we strengthen individuals. This means promoting AI systems aimed at collaboration, operating innovation, while protecting against the fall gravel in centralized control, says Richard Sutton, widely recognized as “Father of reinforcement learning.” “Reinforcement learning rather than large language models have the key to promoting AI,” he said.
Democratization of AI
Thames Network democratizing access to AI with the first open source decentralized AI marketplace, protocol and incentive layers.
Universal Basic Income – where citizens are offered recurring payments to subsidize their lives – is proclaimed by AI oligels, and especially Elon Musk, as needed. This is not a first approach to people; This is a business-first approach and one that eats away by the community’s drug. The better approach is to democratize access to AI and to enable autonomy and sovereignty to the individual.
With a new intelligence substrate at the edge and with a new economic model, a decentralized intelligent network would enlighten an ecosystem of agents working with humans. Instead of teaching or replacing people, this network will create new opportunities for democratization of human-IA collaboration.
Depolitizing AI
For artificial and synthetic intelligence for the benefit of humanity, it is imperative that it is free of bias and be apolitical without an implicit (or explicit) agenda. Censorship, protection restrictions and access restrictions based on jurisdiction, price and other factors are not the way to create a future where people and AI can collaborate effectively.
At the same time, privacy is key in domains as healthcare. A decentralized intelligent network must be designed with a first approach to privacy and architects on a reliable foundation, ensuring a zero-till security model while balancing governance, risk and compliance.
What can start with hundreds of thousands of models will be built up in a massive wave of hundreds of millions of domain-specific models, curated, distilled and reinforced via retrieval-augmented generation (RAG). The Thems Network will provide the tools and an open marketplace for people to build and make money from their domain expertise, again with a focus on human-IA collaboration.
“We all see the digital world take over our world through the Internet, collection and sharing of data and the current increase of AI,” says Bill Roscoe, director of Oxford Blockchain Research Center. “The world really needs an altruistic development of the rules of digital civilization and an infrastructure to support and control it in a truly collective way.” Thames Network’s mission is to ensure that privacy and collective governance remain at the forefront of technological development.
The convergence of decentralized computing, blockchain tools and governance, crypto-incentive protocols and mechanisms and domain-specific AI models built and curated by human experts points to a future where artificial and synthetic intelligence becomes available, transparent, secure, abundant and cooperative.
Thames Network, which we advertise on the Oxford AI X Blockchain conference today, imagines a win-win world for people and AI. Everything else would be an abdication of responsibility for us as technologists, engineers, researchers and economists.
