ScholarEvolve uses published research to improve AI agents, raising Qwen's AppWorld task completion from 69% to 81.4% without retraining the model.
A 450-task benchmark built from real CVEs shows even the best AI coding agent writes code that's both working and secure only ...
A school math teacher refers to a dry textbook in response to a simple question, "What is a logarithm?" A familiar plot ...
Ads are what make it possible to produce new content on Phoronix each and every day over the past 20 years. Please support ...
Twine is a free, open-source dev tool for designing interactive fiction and text-based games. Turns out it's perfect for ...
A new Python package called PoolParty streamlines the design of DNA sequence libraries for high-throughput biological ...
UCSB and Microsoft's ScholarEvolve generates code from academic papers to improve an agent's harness without retraining model weights. It raised Qwen3.5-27B's task completion rate to 81.4%. Here is ...
GPT-6 Astra edited Qwen’s weights and raised MATH-500 from 7.9% to 48.1% in author-run tests. How weight-space programming works—and its limits.
Hackers are spreading a Python-based malware kit called ArmorSteal through fake software download pages. The kit steals ...
Researchers at San Diego State University have developed a software tool called CodeJeNN that automatically converts neural ...
On October 21, 2026, a compromised LiteLLM 1.107.0 release was published to the Python Package Index (PyPI) and distributed through official container images. The malicious package added files that ...
Build an AI code review pipeline on GitHub Actions with Semgrep, reviewdog and a Python model reviewer. 12 steps, tested code ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results