Farcaster Democratizing Ai via the Social Data Ledger

In an era where big tech monopolies own our digital lives, controlling and profiting from our data in ways we can hardly know, a hope for change shines. Enter Farcaster: a sufficiently decentralized social media platform changing the the landscape of social data. This post delves into how Farcaster's social data ledger (Hub) is not just an alternative but a revolutionary shift towards democratizing data and enabling a social dataDAO layer empowering users to monetize their digital footprint effectively.

It's no secret that platforms like Facebook and Twitter have turned user data into gold mines. From targeted advertisements to training sophisticated AI models, these tech giants extract immense value from our digital interactions. While Facebook might earn $200 per user annually through ads in the US, quantifying the value created from our data through AI models remains unclear. The control remains firmly in the hands of these platforms, with costly & limited API access for third-party applications further cementing their data monopolies. Twitter’s development of Grok AI based on our real-time data feed, is the new paradigm of the value generated by our social data & captured by big tech monopolies, without we the users have any say nor profits in it.

Farcaster stands out by storing data publicly, accessible to anyone running a Farcaster hub, free of charge. Its use of Postgres DB—an object-relational database—fits perfectly for scalable web & mobile applications, while users’ cryptographic signing of data to hubs ensures security and verifiability of our data. Imagine a social transaction on Farcaster not just as an online interaction but as a verifiable action on a decentralized social ledger. This approach enhances user experience by reducing the need for on-chain actions for every data interaction, simplifying web3 dApp development.

While Postgres excels in efficiency for web and mobile applications, leveraging this data for AI requires some work. Here, the concept of a Knowledge Graph (KG) comes into play. Unlike traditional DBs, a knowledge graph stores data in a machine-readable schema. Unlike relational databases, relations are in natural language, making the DB semantic. Here is a simple visualisation how a Postgres DB is different from a knowledge graph DB (image credit).

The term Knowledge Graph, was coined by Google in 2012 as a technology to enhance their search result & stayed as a monopolized asset under the control of Google, while many big tech monopolies built their own siloed knowledge graph DBs.

The key feature of KG that is making such DBs most relevant now, is the high efficiency being used as a knowledge base to query via LLM models. Which can solve the “hallucinations” problem of LLM models, by having a verifiable knowledge base.

The question then becomes: How can we develop a decentralized knowledge graph on Farcaster that benefits users directly? How can we bring governance & privacy to this asset? How this decentralized crowdsourced social data ledger can become the source of truth? How users can get paid when their data is being used?

The answer lies in the creation of data-driven communities like dataDAOs, which can govern and monetize this knowledge graph, ensuring that the value generated flows back to the contributors. This model not only challenges the existing data monopolies but also opens new avenues for user empowerment and data democratization.

In my next posts, I will delve deeply into how knowledge can be generated, crowdsourced, and curated through Farcaster's client. I'll explore how a community consisting of computational providers, data scientists specializing in knowledge extraction & data schema, and Large Language Model (LLM) services can collaborate with datalatte's dataDAO. We'll discuss how data is stored in a hybrid model of private and public knowledge and examine the privacy protocols that can be employed to ensure users' sovereignty and privacy. Additionally, I will explain how value is created and distributed among all parties involved, from the creation of knowledge to its consumption.

My name is Amir, and I am an engineer based in Berlin. After completing my PhD in chip design, I discovered my passion lies at the convergence of data, artificial intelligence (AI), and web3 technologies. Over the past three years, I have been engaged in research and development with datalatte, collaborating with partners such as Ocean Protocol (a decentralized data marketplace), OriginTrail (a decentralized knowledge graph), and Neuroweb.ai (Knowledge mining for AI).

Follow us:
datalatte Warpcast
dudeamir Warpcast