Open Source Ia

Anthropic released a report arguing that the open-source model GLM-5.3 poses a major cyberattack risk, with its core concern being the model's potential for misuse. The report fuels the ongoing debate about the benefits of open models versus their potential security risks.

AI Safety • MIT Technology Review

A new open-source project called openTPU aims to build an AI accelerator designed by AI agents. The project, which runs on an FPGA card, is a learning initiative to explore how far AI can go in hardware design and whether it can build the chip that runs its own inference.

AI Infrastructure • Hacker News

According to CrowdStrike, a suspected Chinese-speaking attacker hacked multiple South Korean financial institutions using AI-powered hacking tools. At Shinhan Bank alone, more than 25,000 customer records were stolen, with the attacker using the open-source tool ARTEX that leverages AI models like DeepSeek and GLM-5.

AI Safety • The Decoder