Skip to main content

Newcoin For Apps

Newcoin is an algorithmic points protocol allowing application developers to add powerful machine learning capabilities for predictive filtering with a few lines of code.

We are the algorithm layer of the composable web. Think of it as the same kind of powerful algorithms web2 has, but privacy-preserving, open, and interoperable. You can query smart contract tables to read numbers that tell your application if something or someone is adding value to the Web, and you can turn your app into an intelligent system without having to do all the machine learning, data science, computation, data processing etc., and just query those endpoints for each blockchain wallet, for free.

What makes products like Google, Gmail, TikTok or ChatGPT so powerful is their ability to learn and adapt based on the accumulation of feedback data and the discovery of patterns in user preference. What if we could achieve a synergy between all apps, systems, and networks where user preferences, what we like, value, would be shared in a synergistic way.

In order to achieve this synergy, Newcoin encodes all the human and algorithmic points available from social graphs and blockchains and aggregates them into Base Points which are on-chain values that can be universally queryable by smart contracts or any other system. The sum of all the points constitutes a cross-platform Agent Graph, where each agent owns their relationships to all the other nodes of the graph.

Newcoin Network

With Newcoin, you can add a layer of intelligence to your application:

  • Trust anonymous Internet users
  • Find out if a blockchain wallet or DiD contributes value within a network
  • Sort and prioritize content feeds
  • Mitigate noise, spam, and sybil attacks
  • Subsidize storage and blockspace to true value creators
  • Discern bot activity from real value creation
  • Airdrop tokens to the most relevant network participants
  • Retroactively reward public goods
  • Automate and augment data curation and human feedback for reinforcement learning

Most of the Internet is composed of weighted directed graphs representing interaction histories: the likes of your social media post, the followers count of an influencer, the amount of links that a web page receives on Google Search, the feedback you give to ChatGPT after an inference. Those numbers are used by machine learning to predict the next piece of content, who to tag, the next token in a sentence, the next video to play, should the app flag this tweet or DM... This prediction capability is currently siloed within separate systems, when in fact a data synergy between them could produce a deep, cross-platform intelligence network effect. With Newcoin, we are providing a computing substrate for all networks to communicate and achieve synergistic intelligence with a shared data vocabulary.

When you interact with people on social media, you give them feedback, in the form of likes, upvotes, downvotes etc. When you train a model, it generates weights, which are parameters that help improve the model. All those points/scores are statistical models for predictive filtering that enhance user experience, improves discoverability, generative AI, and mitigates spam. Social media algorithms and LLMs are able to predict the "next thing" that should come by analyzing vast amounts of feedback patterns. The bigger the dataset, the more accurate it becomes, and we can't achieve big datasets with closed and fragmented competing systems.

Unlike siloed systems where your like of a ChatGPT inference belongs to OpenAI and a like on TikTok belongs to ByteDance, we are building an open graph where all this feedback is owned by the users, who can cross-pollinate different networks, from AI to social, to smart contracts, and let machine learning algorithms use this Agent Graph to generate synthetic feedback which come from finding patterns.

Imagine if you could own all your Twitter and Instagram likes and present them to a machine learning algorithm that would generate recommendations on Lens or Farcaster.

Imagine if all the likes, follows, and transactions as a music artist would increase your royalties when you contribute sounds to a music machine learning model.

Imagine if all the statements we make about what we like and value would be standardized and shared between various systems to better to compound intelligence across systems instead of trying to lock it within silos.

Newcoin encodes and computes all the signals available across various platforms and turns them into endpoints that can be queried by any system within the composable web.

To build with Newcoin, you have two main components:

IPSPBase Points