Anthropic admits to a month-long degradation in Claude’s output due to reasoning “effort” tradeoffs, cache bugs, and verbosity prompts. As Opus 4.7 rolls out with mixed developer reviews, the company is doubling down on “frontier safety” with the restricted Mythos model while leaning on high-profile advisors like Rishi Sunak to navigate global AI regulation.
Checkmarx Acquisition of Tromzo Accelerates Plan to Apply AI to Application Security
Checkmarx accelerates its AI-driven DevSecOps strategy after acquiring Tromzo, integrating AI agents to automate application security across the software development lifecycle.