The Stack — May 02, 2026

Today's technology briefing highlights significant developments across AI/ML, industry dynamics, infrastructure, and the developer world.

AI/ML

OpenAI's GPT-5.5 and Anthropic's Mythos Preview models were evaluated in cybersecurity tests by the UK's AI Security Institute. GPT-5.5 slightly outperformed Mythos Preview in expert-level Capture the Flag challenges, though both models faced difficulties in certain simulations, such as "Cooling Tower." OpenAI also announced a restricted release of its cybersecurity tool, Cyber, to vetted professionals, aiming to expand access through the Trusted Access for Cyber (TAC) program with the U.S. government.

Industry

Legal AI startup Legora, backed by Nvidia's NVentures, has reached a $5.6 billion valuation after a recent funding round, positioning it against U.S. rival Harvey. Meanwhile, Anthropic is reportedly seeking a valuation of over $900 billion in its upcoming funding round, potentially surpassing OpenAI's valuation. Apple has reported increased Mac sales driven by AI workload demands, with the Mac mini becoming a top-seller in China.

Infrastructure

AWS is dealing with extensive recovery efforts after Iranian drone strikes damaged data centers in the UAE and Bahrain. The restoration of services is expected to take several months, with AWS advising customers to migrate resources to other regions. Additionally, Ubuntu and Canonical have been targeted by a DDoS attack attributed to a group sympathetic to the Iranian government, disrupting web services and OS updates.

Developer World

A new tool called "Understand Anything" has been introduced, allowing developers to convert codebases into interactive knowledge graphs, facilitating easier navigation and comprehension of complex code structures. In the realm of safer systems programming, the release of Lib0xc offers C standard library-adjacent APIs designed to enhance safety in systems programming by focusing on fixed-size data structures.

These updates underscore ongoing efforts to integrate AI responsibly, improve software engineering practices, and address cybersecurity challenges across the tech landscape.