Enable JavaScript for topic filters, creators and events.
Google shares how it uses agents to find security flaws
Google describes agents that scan code changes, examine possible vulnerabilities and propose fixes for human review.
Company announcement · 2026-09-18
Read original — Google shares how it uses agents to find security flawsAnthropic announces embedded evaluation with Accenture
Anthropic says Accenture's Faculty team will assess models and safeguards from inside the company. Funding and access arrangements are still developing.
Company announcement · 2026-09-18
Read original — Anthropic announces embedded evaluation with AccentureHow machine learning could improve flash-flood warnings
The Verge examines a system that combines satellite information with machine learning to help forecast dangerous flooding.
Reporting · 2026-09-18
Read original — How machine learning could improve flash-flood warningsAnthropic proposes ways to measure AI development
The proposed measures cover AI's role in research, oversight of agent actions and allocation of computing resources.
Company analysis · 2026-09-17
Read original — Anthropic proposes ways to measure AI developmentOpenAI links AI adoption to business outcomes
OpenAI describes analytics for understanding ChatGPT Work and Codex usage, spending and the work teams accomplish.
Company announcement · 2026-09-16
Read original — OpenAI links AI adoption to business outcomesIBM researchers ask whether an agent can repeat its success
An IBM Research post distinguishes average success from reliability across repeated attempts and describes a tool for finding inconsistent decisions.
Research / author report · 2026-09-15
Read original — IBM researchers ask whether an agent can repeat its successA visual workflow for image generation with Gradio
A Gradio walkthrough rebuilds image-generation features as connected steps, including image editing and image-to-video workflows.
Tutorial / background · 2026-09-10
Read original — A visual workflow for image generation with Gradio