software-security

LIVE PULSE

4.9 Anthropic CEO Amodei calls for slower AI development and shared safety rules — 11 src2.6 Agility Robotics unveils Digit 5 humanoid for warehouses and factories — 2 src2.4 Apple ships rebuilt Siri with Google Gemini, but not in the EU — 2 src2.2 Siri AI in macOS 27 Golden Gate: FAQ, Germany availability, privacy questions — 2 src1.7 Sam Altman says OpenAI will not go public in 2026, citing AI safety concerns — 5 src1.4 OpenAI contractors review real ChatGPT conversations to rate responses, report says — 2 src1.3 Anthropic data retention policy prompts firms to limit Claude use for sensitive work — 1 src1.3 CoMem Paper Proposes Shared and Individual Memory Design for LLM Multi-Agent Systems — 1 src1.3 Paper proposes evolving context parameterization for large language models — 1 src1.3 Fine-Tuning Vision-Language Models with Listener Gaze for Referring Expressions — 1 src4.9 Anthropic CEO Amodei calls for slower AI development and shared safety rules — 11 src2.6 Agility Robotics unveils Digit 5 humanoid for warehouses and factories — 2 src2.4 Apple ships rebuilt Siri with Google Gemini, but not in the EU — 2 src2.2 Siri AI in macOS 27 Golden Gate: FAQ, Germany availability, privacy questions — 2 src1.7 Sam Altman says OpenAI will not go public in 2026, citing AI safety concerns — 5 src1.4 OpenAI contractors review real ChatGPT conversations to rate responses, report says — 2 src1.3 Anthropic data retention policy prompts firms to limit Claude use for sensitive work — 1 src1.3 CoMem Paper Proposes Shared and Individual Memory Design for LLM Multi-Agent Systems — 1 src1.3 Paper proposes evolving context parameterization for large language models — 1 src1.3 Fine-Tuning Vision-Language Models with Listener Gaze for Referring Expressions — 1 src

topic2 events

papersSEP 12 04:00 UTC

Study Measures Gap Between Static Analysis Passes and Dynamic Failures in Python

A new arXiv paper examines cases where Python code passes static security checks but fails when actually executed, a discrepancy the authors call the static-pass dynamic-fail gap. The work focuses on security-sensitive code as well as code produced by large language models, arguing that static analysis alone is insufficient as a security gate. The authors frame the measurement as a step beyond reliance on static guarantees for evaluating generated software.

static-analysis/llm-generated-code/security/python

papersSEP 10 04:00 UTC

Paper studies measurable independence when AI models build, defend and test software

A new arXiv paper examines the situation where a single family of generative AI systems takes on three roles in software work: authoring application code, protecting and monitoring it, and hunting it for exploitable weaknesses. The authors introduce the notions of measurable independence and bounded autonomy to evaluate how separate these roles really are and how much freedom such systems should be granted. The work pushes back against the assumption that complete autonomy for these multi-role models is the right default.

research/agents/safety/cybersecurity