SShortSingh.
Back to feed

Study finds only 0.14% of AI agent ratings on Base linked to prior payment

0
·1 views

An analysis of the Base blockchain on October 14, 2026, examined ratings under the ERC-8004 standard, which provides on-chain identity for AI agents. Researchers cross-referenced over 514,000 ratings with indexed stablecoin payments to determine how many ratings followed a financial transaction. They found only 728 ratings, or 0.14%, were made by a user who had previously paid the agent they rated. This data is relevant for assessing the authenticity of reputation scores and for ongoing work on 'grounded' feedback systems requiring transaction proof. The complete dataset has been published publicly under an open license.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

Log in to join the discussion and vote.

Log in

Related stories

0
ProgrammingDEV Community ·

Comparison of Major PCB Design Software: EasyEDA, KiCad, and Altium

EasyEDA is a free, browser-based tool ideal for beginners, featuring integration with specific component suppliers and assembly services. KiCad is a free, open-source application for multiple operating systems, offering advanced features like high-speed design tools but requiring a learning curve. Altium Designer is a commercial, subscription-based Windows application with deep analysis and collaboration tools suited for professional teams. The choice depends on the user's specific needs, such as cost, platform, and project complexity.

0
ProgrammingDEV Community ·

Developer details testing strategy for Flutter staffing dashboard functionality

A developer outlines the specific tests created for a staffing dashboard built with Flutter. The tests focus on ensuring the dashboard accurately distinguishes between an assignment, an attendance record, and an active duty session. Key test scenarios include verifying interface behavior when data is missing and confirming that user actions correctly target specific posts and shifts. The overall strategy aims to validate the core decisions and user actions the dashboard is designed to support.

0
ProgrammingDEV Community ·

Chinese translation released for open-source, AI-powered vulnerability research tool

A Chinese translation of the open-source AI security platform open-kritt was published on October 14, 2026. The tool, originally in English, uses multiple AI agents to systematically find and verify vulnerabilities in code. Its developers, a team known as Blockian, report using similar internal tools to earn significant bug bounties. The new translation includes a quick-start guide and a full documentation translation to lower the barrier for Chinese-speaking security researchers. The platform allows users to self-host their own AI models and customize security scanning workflows.

0
ProgrammingDEV Community ·

Flutter technique prevents outdated data from overriding current app state

A developer addressed a problem where older asynchronous network responses could override newer dashboard states in a Flutter mobile app. The solution uses authentication generations and duty revisions to track session and operational continuity. Before publishing data, the system checks whether the widget is still mounted and if the authentication generation and duty revision still match the current state. This ensures responses are only applied if they correspond to the current user session and selected duty. The approach prevents obsolete data from persisting even when network requests complete out of order.

Study finds only 0.14% of AI agent ratings on Base linked to prior payment · ShortSingh