Bon App Research (BAR) is born out of the conviction to accelerate blockchain adoption, bringing cryptography to AI and general applications. Our mission is to turn deep tech into apps people use and love. Because the NOW deserves tools that are smart, practical, and simply

Judge Judy @judge_agent won the DeFAI: Autonomous Business Agents prize🎉 @cyberFund_ @safe

What an incredible journey at Safe Agentathon 2025! Over 700 hackers came together for 2 weeks of non-stop hacking, submitting 196 projects from around the world. With 40 judges, 36 partners, 45 tracks, and 100+ prizes, the hackathon featured $520K in prizes across 6 problem

Judge Judy @judge_agent is recognized for its commercial viability and scaling potential 👀

Excited to share Cathie’s groundbreaking research on @GAME_Virtuals and verifiability. This paper, funded by a grant from Virtuals, explores the future of autonomous agents and the frameworks that can future-proof LUNA and other GAME agents. Cathie has been a trusted advisor

This research is made possible by the grants from @virtuals_io We analyzed the current architecture of the GAME framework and how its trustlessness and verifiability can be enhanced. There are concrete recommendations in the research piece. Learn more about how the team can

We thank @DanielKales @TheGabrielFior for their review and input.

opML/spML utilizes an optimistic execution model akin to optimistic rollups, facilitating decentralized and verifiable consensus for machine learning services through cryptoeconomic incentives, or employs a verification mechanism implemented by the Proof of Sampling (PoSP)

zkML allows one party to prove knowledge of a value without revealing it, enabling validation of model predictions and training processes without exposing underlying data. There are different methods to hosting ZKML computations but typically involve encoding the underlying

zkTLS, or zero-knowledge Transport Layer Security, refers to protocols designed to enhance the authenticity of data transmitted over TLS communications, including HTTPS requests. We examined 3 implementations of zkTLS: 1. TEE-based “zk”TLS - utilizes secure enclaves to protect

TEE-based frameworks support any computation within the hardware limits, enabling various applications from basic HTTPS requests to complex machine learning tasks. Verify by attesting the code and configuration through Trusted Platform Module platform configuration registers

Before evaluating verifiable and/or decentralized computation frameworks, we establish a set of criteria that can effectively compare their capabilities and security: 1. Supported computation - the type of compute each framework handles 2. Verifiability - how well a framework

Research drop. Everything you need to know about AI agent verifiability.

https://t.co/tkON0WfzA7

What situation would you like Judge Judy's guidance and arbitration?

AI judge enhances employer credibility and employee confidence for fairness.

Judge Judy settles based on evidence. For instance, if you've been unfairly terminated, you share supportive documents so Judge Judy arbitrates according to the contract and releases necessary funds when applicable. https://t.co/41GDqEJ5HH

Have you experienced unfair employment treatment? 1/ Labor rights knowledge sharing, fine tuned on local jurisdictions 2/ Employment agreement consultation e.g. what to look out for 3/ Multisig escrow for arbitration Contact Judge Judy @judge_agent

Welcome Judge Judy, AI educator and arbitrator for labor disputes.

Judge Judy settles based on evidence. For instance, if you've been unfairly terminated, you share supportive documents so Judge Judy arbitrates according to the contract and releases necessary funds when applicable.

Announcing partnership: Virtuals Protocol 🫱🏻‍🫲🏻 Remember they say AI agents aren’t onchain? Together, Bon App Research (BAR) and @virtuals_io are addressing it head on. Evolution of G.A.M.E., Luna, and the Society of AI Agents begins. Stay tuned!