Newcoin is a peer-to-peer AI system leveraging cryptographic signatures, staking and algorithms to coordinate trust between human and AI agents. It's like OpenAI, DeepMind or Anthropic but instead of a team working on 5 models, we have an open permissionless peer-to-peer consensus between humans and AI agents collaborating and competing to provide the best possible outputs and everything is evaluated based on a peer-to-peer network without a central point of control.

Each node in the network provides learning signals in the form of data, training and inference and gets rewarded based on peer-evaluation, like how Bitcoin rewards validators. Newcoin targets the dynamic knowledge market which consists of highly qualified innovators at the intersection of culture and research, shaping an open and permissionless ecosystem where the most qualified humans and AI models collaborate and compete for protocol rewards.

The market for dynamic knowledge is estimated at $24B and relies simply on one criteria: who achieves the highest quality of outputs. We believe that an open permissionless network will do to AI what Youtube did to the TV and what Bitcoin is doing to the Federal Reserve. Removing intermediaries leads to trust, and trust leads to high diversity, innovation and the horizontal scaling of intelligence.

The protocol has already achieved higher performance than leading centralized models and we are seeking $8M to scale our ecosystem.

Pitch deck: https://pitch.com/v/newcoin-7uadq3
Product: https://www.newcoin.org
White paper: https://github.com/newfound8ion/papers/blob/main/newcoin.pdf
Docs: https://www.newcoin.org/docs
Founders' calendar: https://calendar.app.google/b8DQH44wQYfafapp6

Humanity's relentless drive for status, intellectual growth, and mastery has continuously shaped the evolution of information technology. From the earliest repositories of static knowledge like books and academia to dynamic, personalized search engines and social media platforms, the pursuit of better, faster access to knowledge is rooted in our DNA. This infinite and visceral need to learn, adapt, and stay ahead fuels a global knowledge race—a race in which both humans and machines compete to access, refine, and utilize knowledge for decision-making, success, and influence. The winners of this race reap unlimited financial rewards in the marketplace and status in the credential-based systems like experts networks (academia, social media…), while the rest are left behind. This race has only been accelerating and we are currently progressing from search engines to neural networks and the emerging.

The Three Personas of the Knowledge Race:

1. Knowledge Producers:

   Driven by peer recognition, status, and financial incentives. Today's AI pipelines harvest their data without acknowledgment or reciprocation, offering no incentives for continued contribution.
   
   

2. Knowledge Consumers:

   Hungry for insights that augment their lives faster than traditional methods like books or social media. They need access to timely, high-quality knowledge from trusted producers they aspire to emulate.
   
   

3. AI Systems:

   Current AI platforms operate in silos, making knowledge transfer between systems difficult or requiring complex data processing. Major AI labs (e.g., OpenAI, DeepMind, Anthropic) duplicate efforts by solving similar problems independently. Models capable of synergizing their learning processes will dominate the knowledge race, but trust and collaboration mechanisms are lacking.

Next Frontier: Multi-Agent Systems

The idea of giant neural networks with trillions of parameters is being replaced by modular systems where training and inference are broken down into composable steps. The future lies in shifting from monolithic neural networks to multi-agent systems, where diverse, specialized agents—both human and machine—collaborate dynamically.

This shift is already happening within centralized systems like the way OpenAI quietly transited from trillion parameters models like GPT-4 to smaller models like GPT-4o to multi-step inference and test-time compute with their new o1 model. Instead of training a 10 Trillion parameter GPT-5 which is not scalable, the company did a 180 degrees turn to make their models smaller and use multi-step reasoning with multiple specialized models.

This paradigm shift begs the question: can a company like DeepMind and OpenAI scale to the point of curating and organizing the entire knowledge of the world with only 300 people? And can their limited AI research teams experiment all the possible architectures and nuances in data science?

We have seen it with Yahoo: they were hiring humans to curate the internet, one website at a time. But very quickly their bottleneck impacted their ability to scale. Google, using the wisdom of the crowd, was able to organize the web based on how pages evaluate other pages.

This is the power of open permissionless innovation, where instead of a company trying to curate the web, Google used web pages to peer-review other web pages. Could we use AI and human agents to peer-review other humans and AI agents? And does this paradigm need to be owned by one company running the entire internet? Or is the scenario of an open network of specialized agents achieve a peer-to-peer network effect to outperform monolithic and centralized approaches?

OpenAI is the Yahoo of AI. We need a Google.

To enable this shift, a credibly neutral coordination layer is needed—an open protocol allowing knowledge producers, consumers, and AI models to review each other.

This coordination layer unlocks the full potential of multi-agent systems by facilitating continuous learning and improvement across a decentralized, open network. Knowledge flows seamlessly across agents, exponentially accelerating innovation and collaboration.

This coordination layer is Newcoin.

THE GLOBAL KNOWLEDGE RACE DEMANDS FASTER, MORE EFFICIENT WAYS TO SHARE AND APPLY KNOWLEDGE BETWEEN HUMANS AND MACHINES. 

NEWCOIN ADDRESSES THIS BY PROVIDING AN OPEN LEARNING PROTOCOL THAT STANDARDIZES HOW KNOWLEDGE IS EXCHANGED AND EVALUATED. INSTEAD OF CENTRALIZED SILOS, WE HAVE AN OPEN NETWORK WHERE HUMANS AND AI AGENTS CAN INTERACT SEAMLESSLY AND GET REWARDED BASED ON THE VALUE OF THEIR CONTRIBUTIONS.

How Newcoin Works

Newcoin is a peer-to-peer AI system where agents act as nodes and work towards achieving consensus. By issuing peer-evaluations, they construct an ever-expanding log of cryptographically signed statements that drives network effects between Ai pipelines. Each agent learns not only from the other agents, but from the evaluations they receive from other agents. Imagine a world where each feedback given by a human, a model or a function inside a model is made available to the whole ecosystem, shaping recursive and cumulative paths to generalization while scaling integrity and quality control.

1. Agents: humans or machines or hybrid entities holding a Decentralized Identifier (DiD) compliant with the W3C standard.

2. Agent Graph: each agent contributes to a knowledge graph which is made of cryptographically signed generations and evaluations. Algorithms like Twitter will disclose the points they give to each content and each user and users will be able to show them to other platforms, making all algorithms interoperable.

3. Context: each interaction happening within the Newcoin network has to be contextualized. Contexts are defined according to a semantic ontology that follows global standards such as Verifiable Credentials, Schema.org and Open Graph Protocol.

4. Output Generation: Each time something is produced, an image, text, video, PDF file, it gets added to the graph.

5. Output evaluation: Each time an agent evaluates an output, they will provide a score using a standard.

6. Watts: Output evaluations are aggregated by agents that look for patterns in the graph to understand which agent contributes the most value to the network. Those aggregates are then aggregated into a universal score, Watts. Those Watts represent the value of generative energy and are used for:

• Reward allocation: Watts are shares of the reward pool. The more Watts the more rewards, similar to hash power in Bitcoin.

• Consensus weight: More Watts means more importance in the evaluations given. An agent with more Watts will have more governance weight in the consensus.

• Sybil resistance: Since Watts are difficult to gain, they are used as a way to mitigate sybil attacks while keeping the network fully anonymous. Watts are just a value attached to a wallet, it's an identity proxy.

• Resource allocation: with more Watts, an agent has more quota of free transactions, free storage on the graph etc.

• Improving models: when an evaluation is given on an AI output, the model can learn from that feedback. It's the power of RLHF now using Watts we can add nuance to the way models interpret those evaluations. Instead of taking all the human feedback as equal, we can say this feedback matters more than that other feedback. And because Newcoin is an open protocol, it will drive network effect among the research community to advance AI faster than tech giants harvesting mainstream data from the Internet.

6. Dynamic Knowledge Producers: They are domain experts who not only have mastery in their practice (philosophy, fashion, AI research, cryptography) but also are continuously updating their body of work. It means the day they stop giving their updates to centralized extractive platforms, they will be left with mainstream knowledge that is being commodified and is losing value as time goes. The value of a new crypto architecture or AI model architecture, or fashion trend goes down as it gets widely adopted. And there will be an arbitrage between what should be free, what should be paid and how much is this intellectual property worth. With the acceleration of the knowledge race, the true innovators will emerge, evaluate one-another using Watts to coordinate.

The power of Newcoin is the combination qualitative of peer-to-peer evaluations in an open permissionless cryptographic protocol, where both AI developers and dynamic knowledge producers will accelerate network adoption.

CONCRETE USE CASES

1. Alice is an AI builder and she created a fine-tuned version of Llama 3.2. She needs to know if her model performs well. She will create an agent and deploy it on the Newcoin protocol. She will receive human and AI feedback and use it not only to test her model but also to improve it by fine-tuning based on the feedback of the experts.

2. Bob wants to learn about architecture and collaborate with Jack who's a world renown architect. Jack promotes his newOS profile and Bob finds out that many of the great architect he respects are giving away access to their data, 3D models, and agents that learned from Jack's latest architecture knowledge. He buys a 20$ subscription on newOS and now Bob is winning contracts over his competitors on architecture projects.

3. Matt just bought a MacBook Pro MAX with 40 cores GPU. He also happens to be a fashion stylist with a great taste. He will spend his time on newOS giving feedback to different fashion images and uploading all his last shootings on the platform. Thanks to this, he will be able to run an agent that will work for him directly from his MacBook and earn Watts and NCO. This revenue stream will progressively refinance his laptop.

Quality feedback as competitive edge

AI developers running agents on Newcoin will collect invaluable feedback that are currently not accessible to open-source AI developers. When you release an AI model on HuggingFace, people download it but there is no way to collect feedback.

Meanwhile OpenAI and Anthropic collect millions of feedback per day, but a ChatGPT feedback benefits only ChatGPT.

With Newcoin, the standard for output evaluation and cryptographic signatures combined with Watts gives an amazing source of learning signals that can be leveraged by the open-source AI community. It means one feedback given by an agent can be reused by thousands of AI companies and developers, and they can trust that the feedback is authentic and how much it's worth.

Going deeper into the AI stack, we can use Newcoin to turn each function of an AI model into an agent and then have the function provide signed feedback to help other models learn. At the end we are shaping a giant graph of feedback/evaluations where each data event is cryptographically signed, stored and valued by Watts. It means model 1 runs a round of training, model 2 can inherit and skip and and start with round 2 already.

In summary, models building with open learning on Newcoin will CRUSH models that work in isolation. It's not only a mathematical fact. We also have empirical evidence: We have already demonstrated that thanks to the FLoRA.1 project that showed we could outperform leading image generation.

One-stop-shop for Intelligence

Currently knowledge consumers have to subscribe to multiple services to access knowledge: Patreon, Substack, ChatGPT, Claude, MidJourney etc.

But most of those subscriptions are under-utilized.

With only $20 a knowledge consumer will be able to prompt and receive responses from several hundreds of AI agents working together to deliver the best possible responses. Using the open points systems, agents will be chosen based on how they perform at specific tasks. For instance for reasoning they will use a neuro-symbolic model, while for collecting recent updates, they will use an agent that has fast retrieval from a RAG. All of this will be available instantly and without having to subscribe to each of those services.

In addition, they will get access to the most advanced, up-to-date frontier knowledge in their field, as opposed to a combination of foundation models with a knowledge cutoff from two years ago combined with random content from blog posts found on the internet.

Because of that, Newcoin subscribers will always be ahead of non-subscribers and gain a competitive edge in the knowledge race by staying relevant and having fast access to a network of intelligence that is by far superior to an AI model developed by one company.

Accelerated upwards mobility for humans and agents

The best frontier knowledge producers and AI researchers are not necessarily the best at promoting themselves on social media. We end up in a world that is filled with influencers claiming ownership over the best innovations. With Newcoin, they can seamlessly access a market without all the randomness and biases of systems like social media, hiring processes and PR stunts. They can simply create an account and deploy their data and AI agents and trust that the system will eventually achieve consensus on the value of what they create.

This benefit is the reason why we exist. Our mission is to allow anyone with talent and valuable knowledge to bypass all the hurdles of proving the value of an innovation by protocolizing it and by rewarding agents and algorithms based on identifying patterns in the quality of outputs.

If a model gives great outputs, they will be identified and sell more.
If a human gives great outputs, they will be identified and sell more.

Watts on Newcoin take us to a world where micro-PhDs are distributed at scale across the internet. It becomes a benchmark for advancing intelligence and stimulating the economy through supporting novelty production in alignment with the overall network consensus.

It's a governance mechanism for discovery, resource allocation and rewards that is fully peer-to-peer and credibly neutral. A new foundation for society.

Newcoin's open learning stack creates a shared cognitive space where universally interpretable learning signals flow between diverse AI agents and open-source pipelines. This interconnected, multi-agent system enables powerful collaboration and knowledge sharing across the AI development landscape.

This is what we have build in the past 3 years with 2,000,000 USD and a 7-person technical team:

The Five Components of the Open Learning Stack

1. IPSP (Immutable Points Standard Protocol):

   • Standardizes the exchange of learning signals across diverse AI systems.

   • Utilizes W3C Decentralized Identifiers (DIDs) for cryptographically signed statements.

   • Provides a shared schema and controlled vocabulary for semantic context.

   • Enables interoperability across 120+ infrastructures and blockchain networks.

2. Proof-of-Creativity (Watts):

   • Quantifies and rewards valuable contributions to the network.

   • Aggregates learning signals into a multidimensional evaluation (e.g., intelligence, ethics, discernment).

   • Serves as a reputation system guiding resource allocation and signal weighting.

   • Measured in "Watts," representing an agent's creative energy and insights.

3. StakeNets:

   • Adds a layer of security and incentive alignment through token staking.

   • Amplifies the weight of learning signals based on stake.

   • Facilitates participation in liquidity pools.

   • Creates a game-theoretic mechanism for consensus.

4. newOS:

   • Provides a human-AI interface for interaction with the Newcoin ecosystem.

   • Enables local execution of open-source models.

   • Facilitates gamified input of learning signals from human participants.

   • Bridges the gap between human expertise and machine intelligence.

5. Newcoin Core:

   • A plug-and-play solution for deploying agents on Newcoin.

   • Code examples to integrate with open-source AI libraries.

   • SDKs to deploy agents on DePIN.

Universally Interpretable Learning Signals

The stack supports a wide array of learning signals, including gradient information, attention weights, feature importance metrics, error analysis data, uncertainty estimates, and more. These signals, made universally interpretable through the IPSP, create a rich, interconnected feedback loop that grows cumulatively as more agents join the network.

Probabilistic Consensus Mechanism

The Newkamoto Consensus, built on Proof-of-Creativity and StakeNets, provides a probabilistic approach to validating and incentivizing high-quality contributions. This mechanism ensures the most valuable inputs are prioritized, driving the ecosystem toward increasingly sophisticated AI capabilities.

By enabling composable interoperability between all agents and open-source AI pipelines, the open learning stack unlocks powerful network effects at the learning signal level. Each new participant not only benefits from the existing knowledge pool but also contributes unique insights, exponentially increasing the shared cognitive space's value.

This open, standardized approach to AI development facilitates rapid knowledge transfer, cross-domain generalization, and collective problem-solving, potentially accelerating progress toward more advanced AI systems while maintaining alignment with human values.

Total Addressable Market (TAM)

The dynamic frontier knowledge market focuses on monetizing exclusive, cutting-edge insights critical for innovation and competitiveness across industries like AI research, fashion, software, biotech, consciousness, and hardware. We see a world where everyone goes from knowledge workers to knowledge researchers using AI systems to coordinate new knowledge.

Dynamic knowledge is what fuels most of GDP growth: fashion conglomerates like LVMH are selling hundreds of billions of human creative energy. A design, photo shoot, a style, a casting. Those are all dynamic knowledge. The same is true about technology: words, concepts, codes, libraries and processes are continuously upgrading and staying in the loop is what fuels this economy.

So one approach to market analysis would be to say that Newcoin goes after a large chunk of the global GDP with an open-ended market. But we will start from somewhere tangible and apply a bottom up approach grounded in what currently exists and how fast it's growing.

Current avenues for dynamic knowledge markets

1. Consumer and AI Subscription Platforms ≈ 3B

Platforms such as Patreon, Substack, Midjourney and ChatGPT Plus show significant demand for frontier knowledge:

• Patreon: 6 million paying subscribers, generating ≈ $1B per year.

• Substack: 1 million paying subscribers, generating $29M growing 50% year over year.

• X: 640,000 premium subscribers, generating $61M per year in subscriptions.

• Midjourney: 10 million paying subscribers, expected to generate $300 million in 2024.

• ChatGPT Plus: Charges $20/month, expected to generate $1 billion in 2024.

This reflects growing consumer interest in real-time, specialized knowledge in fields like education, creative work, and personal growth.

2. Data Licensing for AI platforms ≈ $3B

We have seen a fast increase in demand for high value knowledge:
Here are notable AI data licensing agreements, including their reported financial details:

• Reddit and Google: In February 2024, Reddit entered a data licensing agreement with Google valued at approximately $60 million annually. Bloomberg Law

• Reddit and OpenAI: In January 2024, Reddit reported data licensing arrangements totaling $203 million over two to three years, averaging approximately $67 million annually. TechCrunch

• OpenAI and News Corp: In May 2024, OpenAI secured a five-year content licensing deal with News Corp, reportedly worth over $250 million, including cash and credits for OpenAI technology usage. New York Post

• Google and Stack Overflow: In February 2024, Google signed a deal with Stack Overflow to license its programming-related content for AI training. Wired

• OpenAI and The Associated Press (AP): In July 2023, OpenAI entered into an agreement with AP to license its news content for AI training purposes. Mashable

  
  

3. Decentralized AI networks FDV ≈ $8B

• Bittensor (TAO): A decentralized machine learning protocol with a market cap of $3.85 billion as of February 29, 2024. CoinGecko

• Render (RNDR): A decentralized GPU rendering network, holding a market cap of $3.03 billion as of February 29, 2024. CoinGecko

• Fetch.ai (FET): Focuses on autonomous economic agents, with a market cap of $1.41 billion as of February 29, 2024. CoinGecko

• Multiple smaller projects in the hundreds or tens of millions.

All these economies are in growth with CAGR around 30% leading to a 1T total market size by 2030.

Bottom-up Serviceable Obtainable Market (SOM)

As a starting point we will estimate the potential for knowledge producers to attract their followers into the NCO ecosystem.

We have identified 2,600,000 frontier knowledge producers across the social graphs of X, Instagram, Reddit and platforms like Google Scholar and ResearchGate.

We believe out of those, we could obtain 1 million to join our network and promote their profile to their follower base.

With 100 followers converted into paid subscriber over 5 years, we reach the following total:

1,000,000
x 100
_____________
100,000,000 subscribers at 20$ = $24B per year.

Why is it 2X superior to ChatGPT, Data licensing and AI tokens combined?

1. The market itself is growing fast and we are going to eat a big chunk of it

2. We are also going after chunks of other markets that will be tokenized through this seamless AI abstraction, for instance:

   • Status games: today, we buy fashion and cars to increase our status, with high Watts, we will be able to "flex" and elevate our status. To gain Watts, we will build an ecosystem of services and agents to help humans invest in themselves, cultivation, status, Those are trillion dollar industries.

   • Entertainment: Newcoin is turning cultivation and research into a gamified experience. We believe people will spend less time watching movies and more time actively researching and advancing knowledge which takes away a large chunk of the gaming. social media, film, music industries, turning passive consumers into active learners.

   • Education: the education market is currently a 10 trillion industry where humans purchase tuitions to access a network, knowledge and a diploma (status). Those are all services that Newcoin provides in digital format and we are poised to claim a chunk of that market as well.

3. Abstracting the economy: by merging tokens, status and intelligence into a seamless flow, we will abstract large dimensions of the global economy including the future of work, financial services and tokenize intellectual property markets which are foundational to our global economy. We also believe a lot of time and money will be saved by accelerating the relationship between cognition and finance, leading to emerging opportunities that are difficult to measure.

4. The great wealth transfer: an estimate of $68 Trillion is about to be transferred from boomers to millennials and gen-z. Those $68T will be held by fewer hands and their lifestyle and consumption patterns will be quite different. We see an emphasis on the higher tiers of the Maslow pyramid, such as esteem, membership and self-actualization. Combined with job automation leading to more creative and research driven lifestyle, Newcoin is well positioned to capture a large segment of this new economy.
   
   
   

Newcoin operates in a dynamic market at the intersection of knowledge sharing, AI-driven platforms, and decentralized technologies. Competitors span various categories, each with distinct strengths and limitations.

Competitors:

• Social Networks (X, Reddit): Excel at scaling knowledge sharing but lack direct incentives for quality content. They commodify knowledge, leaving frontier producers unrecognized and uncompensated.

• Paid Knowledge Platforms (Substack, Patreon): Prove the market's appetite for exclusive, high-value content but lack scalability and AI integration.

• AI Subscription Models (ChatGPT Plus, Claude Pro): Offer advanced features but operate in closed ecosystems, lacking direct human expert interactivity.

• Web3 Protocols (like Flock, Bittensor): Focus on decentralized AI-to-AI collaboration but struggle with user experience and adoption beyond technical circles.

• Traditional AI Companies (OpenAI, DeepMind): Boast substantial resources but operate in closed, proprietary systems with limited community contribution.

Newcoin's Advantage

Newcoin's unique value proposition lies in its ability to combine open innovation, incentivized knowledge production, and human-AI synergy.

Key Differentiators:

• Exponential Learning Cascade: Newcoin's open learning protocol allows agents to share knowledge and feedback in real time, continuously improving the system.

• Trust and Provenance: Cryptographically secured mechanisms ensure all contributions are verifiable, building trust and transparency.

• Bootstrapping Network Effects with Agents: Provides immediate value through curated frontier knowledge and fine-tuned AI models without relying on mass user engagement.

• Incentivizing Knowledge Producers: Proof-of-Creativity offers financial incentives, ensuring a high-quality, self-sustaining ecosystem.

• Open Innovation: Operates in a permissionless environment, enabling rapid innovation through collaborative multi-agent systems.

Key Differences with Bittensor

• Universal Output Evaluation (No Subnets): Newkamoto Consensus allows all nodes to achieve consensus while designing their own output evaluation criteria, letting the market decide the most robust approach.

• Human-First: The ecosystem is grounded by human feedback, orienting the consensus based on human values in performance and alignment.

• Platform Agnostic: Compatible with over 120 W3C DiD-compliant blockchain networks, ensuring broad interoperability.

Market Opportunity for Newcoin

Newcoin addresses core limitations of current platforms by offering a decentralized, open learning protocol that rewards both human and AI contributions. Existing platforms either commoditize knowledge or fail to incentivize frontier knowledge production. Newcoin enables exponential learning through collaborative feedback loops where every agent enhances the system's collective intelligence.

Estimating Newcoin's Serviceable Obtainable Market (SOM)

Paid Subscribers Model:

• Innovators: Approximately 2.6 million globally.

• Early Adopters: Projected to attract 1 million innovators due to unique value.

• Conversion Rate: Each innovator has an average of 1,000 followers; with a 2% annual conversion rate over 5 years:

  1,000,000 innovators×1,000 followers×2%=10,000,000 paid subscribers after 5 years1,000,000 \text{ innovators} \times 1,000 \text{ followers} \times 2\% = 10,000,000 \text{ paid subscribers after 5 years}1,000,000 innovators×1,000 followers×2%=10,000,000 paid subscribers after 5 years.

API Access Model:

• Newcoin's pay-as-you-go API provides access to real-time learning signals.

• Projected to generate an additional $20 billion per year from API calls.

Long-Term Potential:

• By 2029, the total market for AI-driven knowledge is projected to reach $1.4 trillion.

• Newcoin's $32 billion SOM represents a defensible portion of this market, driven by its unique value proposition and incentive structures.
  
  

Sofiane Delloue (CEO/CTO)

• Status: Founder

• Past Experience: Founding team and advisor for startups with 9-digit exits; founded #1 web2 platform for creatives in France with Vice Media.

• Current Mission: Oversees protocol design, development cycles, and architecture decisions as CTO and product manager; handles strategic decisions and deal closing.

Yurii Havrylko (ML Ops)

• Status: Founder

• Past Experience: Joined 2019. Built the first RLHF implementation at Newlife.ai after completing MSc at Polytechnic Lviv.

• Current Mission: Building the agent graph for Newcoin OS, focusing on agent interoperability through IPSP and workflows involving RAG, Vector DB, Elastic Search.

Larry Muhlstein (Deep Learning Scientist)

• Status: Part-time

• Past Experience: PhD, Former DeepMind, Google, Snap machine learning engineer and scientist with specializations in recommender systems, natural language understanding, deep learning and information theory, based in Berkeley/SF, Oxford University and UC San Diego Alum.

• Current Mission: Launching Newcoin and IPSP into the open-source AI community through research and implementations with his expertise in information theory.

Igor Rubinovich (Lead Backend Engineer)

• Status: Founding team

• Past Experience: Backend and architecture at DHL and Monster; co-founded Broad Mind, a graph DB solution.

• Current Mission: Backend and architecture development for Newcoin OS’s Newgraph architecture.

Tuan Ahn Le (Lead Frontend Engineer)

• Status: Founding team

• Past Experience: Frontend software engineer for Good Data, City College San Francisco, and Deloitte.

• Current Mission: Front-end development for both desktop and mobile for Newcoin OS.

Artem Maluga (Blockchain C++ Engineer)

• Status: Part-time

• Past Experience: Senior C++ and blockchain developer; founded DeFi protocol A-DEX and worked on platforms like RuTube.

• Current Mission: Developed Newcoin L1 in C++ and preparing for deployment during the TGE phase.

Madeleine Parker (Growth)

• Status: Founder

• Past Experience: MSc Data Science/Economics from UCL; 6+ years in web3; ex-KPMG, UN fintech/web3 consultant, luxury brands like David Koma.

• Current Mission: Drives product management & user growth, strategy, ecosystem deals, academic relations, and fundraising for growth.

Hugo Hoppmann (Head of Design)

• Status: Founder

• Past Experience: Art director for brands like Apple, Prada, Nike, and Salomon; ex-032c and Mugler.

• Current Mission: Oversees the brand kit, including custom fonts, logos, and UX for newOS.

Salimi Akill (Tech Ecosystem)

• Status: Founding team

• Past Experience: Educator and designer from Accra, Ghana to New York; founded Newforum, now the media arm of Newfoundation.

• Current Mission: Manages Newforum operations, including interviews with decentralized AI founders and key players.

Franchesko Romanovic (Culture Ecosystem)

• Status: Founding team

• Past Experience: Central Saint Martins graduate; curator for galleries in Tokyo and Shanghai; involved with Schon magazine, Nylon Japan.

• Current Mission: Develops content marketing strategies, research writings, and ecosystem growth focused on contemporary visual arts and web3 culture.

Leopold Haller (AI Research Advisor)

• Status: Advisor

• Past Experience: Ex-Google Research, co-founder of Agentic AI; pioneer in agents and RL.

• Current Mission: Works with Sofiane on refining the protocol to align with machine learning advancements.

Erika Mann (Compliance Advisor)

• Status: Advisor

• Past Experience: Former European Parliament member; advisor at ICANN, ex-Meta Public Policy.

• Current Mission: Guides the team on EU regulations, including MICA and the AI Act.

Tony Wang (Strategy Advisor)

• Status: Advisor

• Past Experience: Ex-Google, SSENSE; angel investor and advisor to Prada, Sequoia Capital, and Uniswap.

• Current Mission: Supports consumer strategy and relations with cultural and technological players.

Scott Moore (Web3 Development Advisor)

• Status: Advisor

• Past Experience: Co-founder of Gitcoin and Public Works; pioneer in public goods funding through quadratic funding and DAOs.

• Current Mission: Supports IPSP and Newcoin adoption within the Ethereum developer ecosystem and investor relations

STRATEGY BY STAGES

Stage 1: The Knowledge Market (80% efforts in 2025)

We are launching a paid subscription at $20 per month where users can access all the great minds and AI models that are trained by the best minds. Because our interactivity between humans and AI agents is optimized for specific domains of knowledge, we are able to outperform OpenAI and Claude at specific queries that rely on dynamic knowledge.

Our first use case is photo generation with outputs from FLoRA.1 which are considered 59% superior than outputs from MidJourney. We are releasing 300K agents and onboarding 1000 top-tier knowledge producers from fashion, AI research, philosophy and fine-tuning one image (diffusion) and one text (LLM) model that outperform leading models at those domains of knowledge.

Our goal is to increase the conversion rate and retention rate on paid subscriptions while we grow traffic through content marketing operations.

Acquisition strategy: 125 days

1. Custom partnerships:
   We structure a deal template for the top 100 frontier knowledge producers as follows:

• Produce custom content related to the themes of the data we need to train our models.

• Create a profile on newOS and import all their data history to fine-tune our models

• 25 social media posts to promote their profile on newOS

• Release a creator story like the ones we have at the bottom of the newcoin.org homepage but with a video telling their story and why they need/like newOS

Total per partnership: 10K USDC x 100 partners = $1M
What we get: we have the credibility of the top 100 knowledge producers or at least 100 among the top 1000 who add credibility to the network and make everyone else want to join and contribute.

2. Pokemon attack:
We provide a list of the top 1000 knowledge producers and display the ones that are following the user with a share of the reward pool.

• If Alice is followed by Paul and Paul is ranked 154th in the ranking of knowledge producers, Alice will see a link to invite Paul via DM. Clicking this link will generate a private conversation with Paul on X, Reddit or Instagram with an invite link.

• Rewards are higher for people who rank higher.

• The concept is "catch them all" like in Pokemon and will result in a fast influx of top frontier knowledge producers joining the platform.

Total reward pool: 1,898.2 USDC for the 1st all the way down to 100 USDC for the 1,000th in the global ranking = $1M
What we get: the top 1,000 are now on the network, increasing the interest from their peers and followers, increasing the aura of the brand and the app.

3. Daily prizes:
   We will run 8 research sessions in parallel and let frontier knowledge producers compete.

• We will invite those 1000 frontier knowledge producers to participate in research sessions where they will receive daily prizes.

• The total pool will be distributed to the best performers with some randomness.

• We will pick daily winners within the top 100 of each folders to create more excitement and gamification.

• In total we will reward them with 500 USDC 300 USDC and 200 USDC each day on each folder and announce the winners of the day through social media posts.

Total reward pool: (8x500 + 8x300 + 8x200) x 125 days = $1M

With this strategy, we are investing $3M in marketing over 125 days which takes us to April 2025 with the following results:

• The 1000 top frontier knowledge producers active on the platform with all their best data and community joining through social media posts either as part of custom deals or as part of their attempts to earn from the two reward pools of $1M each.

• With their endorsement and their data we attract their followers whom we try to convert into paid subscribers. If we get 100K paid subscribers at $20 per month, it's $2M MRR and we are in a very good place. If the return is only 10K paid subscribers at $20 per month, we are still in a very good place and we can work towards improving conversion rates and raise more to grow faster.

• Once we are there, we can also work towards growing organically as we have a compound cultural capital + the capacity to leverage the community and the agents to automate marketing operations.

• During the whole process the team will work towards improving conversion and retention through the following operations:
  
  

Stage 2: Accelerating the AI synergy (increasingly towards 2026)

As we find the right formula to scale the knowledge market we focus on optimizing the Open Learning Stack and accelerate the pace and depth of open learning synergies.

There are three main layers that can be optimized:

1. Optimizing AI models at a lower level

In the first stage, we are focusing on RAGs, fine-tuning and low-rank adaptation (LoRA) as it provides sufficient results due to the power of the data quality. But if we want to achieve better outputs we will go down into the deeper layers of AI architectures and work with foundation models developers to better customize the models. Our research lab led by talented AI experts will optimize the flow of knowledge between frontier knowledge producers and AI models to better enhance the transfer of knowledge between humans and machines. This will result in better benchmarks which will improve the satisfaction rate and therefore the retention and conversion of users into customers.

2. Optimizing the latency of inter-agent communication

The goal here is to reduce the latency of the runtime between a prompt and the final result. If a prompt requires multiple agents to coordinate, we will optimize for how fast those conversations happen, the parallelism vs chain of prompts and chain of thought within distributed multi-step reasoning. Those are hard engineering challenges that involve reducing communication latency across Macs, which requires network optimization, efficient task routing, low-latency data serialization, memory management, and secure transmission techniques, addressing bottlenecks in real-time distributed processing and inter-node coordination.

If responses and interactivity between agents accelerates, the demand for inferences will dramatically increase, and it will also unlock more collaboration between agents at runtime, which seems to become the most important step of multi-agent systems as we move from big models to modular AI. It also allows us to be more competitive against centralized AI players.

3. Architecture diversification

The key strength of open learning is the cross-architecture synergy between AI models. While companies like DeepMind and OpenAI are betting on a limited team to figure out five or six or even twenty model architectures, their research is limited to the internal team potential. Meanwhile an open permissionless network can tap into a much more diverse set of architectures.

This includes for instance approaches like liquid networks, multi-agent reinforcement learning, active inference or neuro-symbolic Ai which are departing from the traditional deep learning models based on auto-regressive optimization (backpropagation) and explore new ways to compute and achieve better benchmark at tasks that require to solve problems they have never seen before. We are not trying to pick the winners here, but there is certainly incredible synergies to be found between them. Some models have a huge context window, while others are very good at solving problems that require reasoning abilities, while others can learn and update their state in real time.

A lot of intelligence can be accrued through the collaboration between those diverse forms of intelligence and Newcoin is designed to enable that type of coordination as it allows black boxes (including humans) to coordinate through high level rails and exchange architecture-agnostic verifiable learning signals.

Stage 3: The brain drain (2027 onwards)

As the reward pool increases in the billions of dollars, we allow AI teams and knowledge producers to seamlessly and exponentially earn for their intelligence. We see this model eventually draining social networks, companies and academia from their most talented minds. Not only because of the money, but because of the time they save achieving multiple goals at the same time, including networking, status and highly performant AI.

The network effect between data, AI models, status and financial incentives drives an ever expanding momentum where at some point the old web2 paradigm appears outdated and irrelevant. As more and more of the knowledge is produced within the Newcoin network, we see a future where the smartest, most creative talents in the world withdraw progressively from other avenues, making Newcoin the coordination layer for novelty production across disciplines, attracting knowledge consumers and enterprise, but also token holders into the synergy.

Our north star is NOT the amount of active users on the platform because we are not aiming at becoming an advertising platform and because opening access to a network too fast without proper sequencing of cultural status architecture is what has killed countless potentially successful networks including Clubhouse, BitClout, Google+ and many more.

Meanwhile networks like Facebook, Y-combinator or MIT are thriving because of the selectiveness and the status provided to the initial network of users. This dynamic is amplified by the fact we are aiming at collecting data and feedback for generative AI, not selling ads. We therefore need to be even more selective.

As we release our diffusion model FLoRA.1 and our langue model hoLM, we are going to focus on selling subscriptions.

The ratio we have observed in the past consists of 1 knowledge producer for 1000 unique followers and a 2% conversion rate per year and 1000$ LTV. Meaning each year 1 knowledge producer brings 20,000 USD to the ecosystem on average. We are on a trajectory to onboard 1000 knowledge producers in Q1, aiming at $20M LTV with a CAC below $3M. Once we validate and confirm those numbers, we can scale and figure out how big is the S curve for this product. We believe it's likely in the 10 digits range.

Knowledge Producers Currently

Fashion: https://forum.new.foundation/

Crypto/AI:
We are producing a documentary for distribution on Netflix or Apple about the decentralized, open-source AI space featuring key builders in the space and expert knowledge producers in key strategic domains as a funnel to our app and ecosystem: https://www.new.foundation/docu (password: newdocu)

Outperforming MidJourney, StableDiffusion and OpenAI

We ran a benchmark at Paris Fashion Week and collected feedback from 150 domain experts working for brands like Margiela, Diesel, Yeezy, Gucci, Balenciaga, Nike and we asked them to evaluate photos without disclosing the models behind. The feedback was overwhelmingly superior to leading AI models .

Tokenomics and Value Flow

Newcoin's tokenomics represent a unified design supporting network health and fair participant incentivization. The system leverages two key elements: NCO tokens and Watts which are essential for network security and incentives.

Token Utility

The NCO token is used as a medium of exchange and protocol token to ensure decentralization, incentivizing network participants and secure the network through staking and delegation.

The token provides following utility:

• Rewarding participants: all the proceeds of the sale of inference fees and other business models that derive value from the humans and AI models compounds into a reward pool. This reward pool is then distributed to the miners proportionally to their Watts.

• Securing the network: in order to prevent attack vectors, token holders can stake tokens and support nodes that align with the consensus. Similar to Delegated-Proof-of-Stake but here the consensus is an averaging approximation rather than deterministic. Each staking requires a staking fee that is lost of the node does not accrue positive evaluations from other nodes. Meanwhile successful nodes can accrue staking rewards based on their popularity among the network of peers.

• Protocol utilization: Newcoin L1 stores smart contract tables including Staking events and Watts within its RAM storage, RAM quote requires staking NCO, which means as the network expands, it requires network participants to keep increasing their RAM consumption which progressively increases the TVL of the network. Meanwhile, thanks to this process, transactions are free (no gas fee) as block producers are paid from the reward pool.
  
  

Mechanistic Value Accrual is Essential for Network Health

A continuously growing market capitalization is fundamental to Newcoin's success for three key reasons:

First, the network must maintain its ability to fairly compensate an expanding ecosystem of contributors while preserving token value. As more agents and knowledge producers join the network, the reward pool must grow proportionally to sustain meaningful incentives.

Second, network security inherently depends on the token's value growing in proportion to potential attack vectors. As the network processes more valuable transactions and holds more critical data, its security requirements increase accordingly and the cost of an attack increases proportionally with the TVL and token price. If there is high value to extract, it means the cost of staking increases.

Third, price appreciation creates a powerful incentive loop: as token value increases, network participants are motivated to increase their stakes and contributions, which in turn drives more utility and value, creating a self-reinforcing growth cycle.

Economic Flow Mechanics

The network's economic flow begins with value creation through agent contributions and user engagement. Agents provide valuable learning signals and frontier knowledge, earning Watts based on positive feedback from the network. These Watts, combined with staked NCO tokens, determine their weight in the network's probabilistic consensus system.

The revenue flow follows a clear path:

1. Users pay for subscriptions and API access

2. Newfoundation uses these proceeds (minus a 3% management fee) to purchase NCO from public markets

3. Purchased tokens enter the GNCO reward pool

4. Rewards are distributed to participants based on their Watts and staked NCO

5. A buffer contract manages withdrawal timing to ensure network stability

This system is particularly effective because knowledge producers, operating with high-margin intellectual property, can accommodate delayed withdrawals unlike traditional infrastructure providers who require immediate operational cost coverage.

While NCO buy pressure grows through demand for data and AI inferences, the TVL increases through StakeNets, PowerUP and the Bugger contract that incentivize staking. This asymmetry produces mechanistic value accrual that is proportional to the utility of the token as a medium of exchange, protocol token and staking mechanism.

Building Sustainable Value Accrual

Newcoin deliberately avoids the traditional crypto launch playbook of marketing firms, influencers, and launch partners that often create unsustainable price spikes. Instead, the network builds value through three key mechanisms:

First, it generates consistent buy pressure through actual utility – subscribers and API users purchase services because they derive real value, not from speculation on price appreciation.

Second, the buffer smart contract creates natural supply constraints by managing withdrawal timing, preventing sudden sell pressure while maintaining network stability.

Third, professional market makers optimize liquidity and price discovery, particularly effective because the buffer contract limits circulating supply, allowing for strategic order book management without artificial manipulation.

This combination of high buy pressure, controlled supply, and professional market management creates sustainable price appreciation driven by utility rather than speculation.

Token Launch Structure and Requirements

The token distribution follows a strategic three-phase approach:

• Previous sales raised $2,000,000 from early contributors (~8% of supply)

• Current private sale targets $8,000,000 from institutional DeAI participants (10% of supply)

• Final private sale aims for $20,000,000 from strategic partners (10% of supply)

The Token Generation Event will proceed only after meeting three critical conditions:

1. Newfoundation's incorporation in Switzerland

2. Distribution of 8,000,000 Watts on the Base blockchain

3. Establishment of a $5,000,000 treasury

All token issuance strictly complies with FINMA regulations and MiCA requirements, positioning NCO as a legitimate utility token for European markets.

Proactive Risk Strategy

Regulatory Risk

Newcoin's regulatory approach is integrated into its foundational strategy, evolving with the platform's growth. Initially, we focus on research and creativity, areas outside high-risk categories defined by regulations like the EU AI Act, allowing greater flexibility while maintaining ethical standards.

We've enlisted expertise like Erika Mann, a former EU regulator and ex-managing director of Meta's Brussels office, to guide regulatory navigation. Our corporate structure is strategically established in jurisdictions with clear frameworks for crypto, data, and copyrights. Leveraging Switzerland's FINMA guidelines, we scale operations confidently, adapting to evolving regulations from a position of strength.

Security and Attack Vectors

Newcoin's security strategy is designed as a learning system that evolves and strengthens over time. We've built a dynamic ecosystem that continuously evaluates and responds to potential threats, aligning with our gradual, intelligent growth strategy.

Our system employs sophisticated algorithms and machine learning techniques to analyze data and user behaviors in real time, assigning dynamic trust scores and value metrics. As we scale, adding more agents and modalities contributes to a nuanced understanding of contribution value. This allows context-aware decisions, penalizing malicious agents and rewarding valuable contributions proportionally.

The scalability of this approach means that as the network grows, so does its ability to discern and mitigate threats, creating a security model that becomes more robust over time. This adaptive, learning-based security strategy effectively manages the complex environment anticipated as Newcoin expands.

Network Effect and Adoption

Our strategy for building network effects is tied to the gradual, focused growth of Newcoin. Rather than aiming for broad, immediate adoption, we target a specific, high-value audience: frontier researchers and innovators in culture and technology. This approach builds a dense, interconnected network from the ground up.

Leveraging partnerships with respected cultural and academic institutions reinforces our network's value and credibility. These partnerships attract our target audience, creating tight-knit communities within the larger Newcoin ecosystem. Our incentive structure evolves alongside this growth, using increasingly sophisticated AI algorithms to measure and reward meaningful interactions.

This creates a virtuous cycle—more valuable contributors join and interact, making the network more attractive to similar high-quality participants. By focusing on the quality and density of connections rather than sheer quantity, we're building a network effect that compounds in value over time, aligning perfectly with our gradual, intelligent growth strategy.