From: thepipeline_xyz
AI is playing an increasingly crucial role in the verification and computation of data, particularly within the blockchain ecosystem, to enable more sophisticated and data-driven applications.
Challenges with Onchain Data
Currently, onchain data is limited, primarily consisting of transactions, addresses, and balances. This scarcity restricts developers from building advanced data-driven or consumer applications. Bringing data from Web2 to Web3 often involves complex cryptographic algorithms, which can be a barrier for developers [02:22:15].
Primus: A Cryptographic Layer for Secure Data Intelligence
Primus (formerly “P” as in “P-dot-one”) aims to address these challenges by building a cryptographic layer for secure data intelligence [02:23:03]. The project connects authenticated, high-value data from Web2 to Web3 and AI applications in a trustless and secure manner across different platforms [02:23:09].
Primus offers two core capabilities:
- Data verification [02:23:22]
- Data computation [02:23:27]
Core Technologies
The foundational technologies behind Primus include:
- MPC TLS (Multi-Party Computation Transport Layer Security) [02:23:30]
- ZK TLS (Zero-Knowledge Transport Layer Security) [02:23:36]
- ZK FHE (Zero-Knowledge Fully Homomorphic Encryption) [02:23:36]
ZK TLS allows for the verification of data flow from Web2 to Web3. Primus’s protocol is noted to be significantly faster than existing solutions like TLS Notary, boasting a 10x speed improvement [02:24:00]. ZK FHE enables verifiable computation on encrypted data, with Primus claiming breakthroughs that improve processing time by more than 300 times compared to alternatives like Zama’s [02:24:12].
Functional Focus Areas
Primus concentrates on four key functional areas to enable data-driven and consumer applications throughout the data lifecycle:
- Onchain data verification [02:24:32]
- Confidentiality [02:24:34]
- Enterprise-level solutions [02:24:34]
- Confidential AI [02:24:36]
Short-Term and Long-Term Goals
In the short term, Primus is focused on providing data verification capabilities through a standard protocol for developers [02:24:47]. This enables the creation of applications such as:
- Digital goods marketplaces on Web3 [02:24:57]
- Verifiable data marketplaces [02:25:00]
- Prediction markets [02:25:03]
Primus plans to open-source its development to foster community growth and is open to co-hosting hackathons to drive innovative applications [02:25:11]. Long-term, the project aims to deploy standalone solutions like FHE payments and FHE VMs for confidential AI [02:25:26].
The company plans to provide SDKs to support various use cases, including onchain data on-ramping and data marketplaces [02:27:46]. This is particularly relevant for applications needing authenticated data, such as Ever Network’s decentralized healthcare data platform [02:28:10]. The long-term vision involves focusing more on ZK FHE to compute over encrypted data [02:28:29].
Other Applications
The broader application of AI in data extends to creating AI agents for the future, leveraging authorized onchain data and decentralized data protocols [02:16:45]. This indicates a future where AI plays a role in managing and utilizing verifiable, real-world data within blockchain environments.
For instance, Totem, an AI deep-in network, aims to unlock the value of daily conversations [03:39:58]. Their device captures voice data, converts it into real-time transcription, translation, and summarization, then encrypts it using Zero-Knowledge Proofs to create anonymized datasets. These datasets can be uploaded onchain via Monad’s fast throughput and exchanged for tokens, enabling users to monetize their voice data [03:41:11]. This demonstrates AI’s role in processing and verifying real-world data for privacy-preserving monetization.