China now has an open-weight model that can find software vulnerabilities and create attacks for anybody to use.

AI coding agents boost code output by 180% but shipping rises only 30%, MIT finds. Why private data access beats benchmark ...

DeepReinforce today released Ornith-1.0, a family of open-source coding models built around a mechanism most RL-trained agents avoid: the model itself writes the training harness that guides its own ...

When an agent does something, the whole company should learn from it, so that every developer gets access to the shared ...

The deal adds a missing piece to the company's vertically integrated model at the software application layer.

AI coding benchmark MirrorCode published its full results June 26, showing Claude Opus 4.7 autonomously rebuilt a 60,000-line interpreter and scored 56% overall — completing tasks that take human ...

As organisations ramp up the use of AI coding agents in software development, they may find costs increase significantly if such tools are overused.

As enterprises embrace agentic AI and vibe coding, Secure Code Warrior CEO and co-founder Pieter Danhieux warns that ...

Booz Allen report warns Chinese AI models like DeepSeek and Qwen may produce more vulnerable code for U.S. government users, ...

Gemini 3.5 Flash is shockingly fast at generating code and spinning up agents, but that speed comes at a cost: sloppy execution, ignored instructions, and frequent mistakes that break real workflows.

China is exerting increasing pressure on U.S. AI companies. Researchers claim that Zhipu AI’s latest model, GLM-5.2, can ...

Google DeepMind is reportedly reorganizing its AI coding team to boost Gemini's coding performance through midtraining ...