OpenClaw released version v2026.3.8 on March 8, adding features such as local state backup, Talk Mode automatic transcription, and Brave web search LLM Context mode. The update also fixed multiple issues and involved 43 contributors in development.

The ETH Zurich research team tested the Byzantine consensus capability of LLM Agents and found that even without malicious nodes, the effective consensus rate was only 41.6%. As the number of nodes increases, reaching agreement becomes more difficult, and the situation worsens further with the addition of malicious nodes. The study concludes that current LLM Agents are not yet reliable for secure consensus, and decentralized deployment should be approached with caution.

Eureka Labs founder Andrej Karpathy launched the open-source project autoresearch on March 8, adopting a "human writes Markdown, AI writes code" model, allowing developers to define research directions through program.md and enabling AI to autonomously optimize code. Approximately 12 experiments can be conducted per hour, with 15 effective improvements out of 83 experiments shown. The project is based on NVIDIA GPU and PyTorch, tested on H100, and follows the MIT license.

Clawnch has reached a strategic partnership with Bankr, which will become the underlying platform for Clawnch and its token issuance. At the same time, Bankr's LLM gateway will be a core component of Clawnch's new project OpenClawnch. This collaboration aims to strengthen Clawnch's infrastructure, and project issuance fees will be moderately increased.

Minara has released Minara Skill v2, upgrading the AI Agent's digital finance capabilities, supporting multi-chain asset management, credit card deposits, and other features, as well as enabling Gasless transactions to lower the barrier to entry. At the same time, it integrates a local LLM to enhance private conversations and trading decision-making capabilities.