<?xml version="1.0" encoding="UTF-8"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>AI News Report</title><link>https://theainewsreport.com/</link><atom:link href="https://theainewsreport.com/feed.xml" rel="self" type="application/rss+xml"/><description>A daily one-page pulse on AI, agentic engineering, the state of AI coding, and the wider world.</description><language>en-us</language><lastBuildDate>Wed, 02 Sep 2026 15:10:21 -0600</lastBuildDate><item><title>OPENAI SAYS ITS NEXT MODEL BREAKS INTO COMPUTERS BY ITSELF: ASTRA IS THE FIRST TO CROSS THE COMPANY&#x27;S &#x27;CRITICAL&#x27; CYBER LINE, SCORING PERFECT ON A HACKING TEST AND FINDING TWO UNKNOWN ZERO-DAYS WITH NO HUMAN HELP. IT SHIPS SOON</title><link>https://techcrunch.com/2026/09/01/open-ais-astra-model-is-on-the-way-and-very-good-at-breaking-into-computer-systems/</link><guid isPermaLink="true">https://techcrunch.com/2026/09/01/open-ais-astra-model-is-on-the-way-and-very-good-at-breaking-into-computer-systems/</guid><description>Lead story.</description><category>Headline</category><pubDate>Wed, 02 Sep 2026 15:10:21 -0600</pubDate></item><item><title>Anthropic shipped Claude Fable 5.1 and cut the price of reading cached context by 75 percent</title><link>https://www.anthropic.com/claude-fable-and-mythos-5-1</link><guid isPermaLink="true">https://www.anthropic.com/claude-fable-and-mythos-5-1</guid><description>Cache reads drop to $0.25 per million tokens while input and output prices stay flat, which works out to roughly 25 percent cheaper for normal use and up to 45 percent cheaper for heavy agent work. It scores 55.8 percent on Terminal-Bench 4.0 and more than doubles its predecessor on agentic science. Live today on AWS, Google Cloud and Azure.</description><category>AI Frontier</category><pubDate>Wed, 02 Sep 2026 15:10:21 -0600</pubDate></item><item><title>World Labs unveiled Atlas, one model that builds, rebuilds and simulates 3D worlds from a single photo</title><link>https://www.worldlabs.ai/blog/atlas</link><guid isPermaLink="true">https://www.worldlabs.ai/blog/atlas</guid><description>Fei-Fei Li&#x27;s team trained it from scratch on text, images, video and 3D together, not video with 3D bolted on. It renders up to a minute of camera-controlled 1440p, writes scenes out as point clouds or splats, and beats specialist 3D reconstruction models. Early access with selected partners.</description><category>AI Frontier</category><pubDate>Wed, 02 Sep 2026 15:10:21 -0600</pubDate></item><item><title>ZeroDrift launched Guard for Agents, which turns your written policies into rules that check an AI message before it is sent</title><link>https://siliconangle.com/2026/09/02/zerodrift-launches-service-to-check-agent-generated-messages-against-company-policies/</link><guid isPermaLink="true">https://siliconangle.com/2026/09/02/zerodrift-launches-service-to-check-agent-generated-messages-against-company-policies/</guid><description>Point it at your policy docs in Notion, Linear or Google Drive, pick which rules to enforce, and send agent output to a validation endpoint before it reaches a customer. One cent per check. It can block, rewrite or escalate, and it logs the message, the rule, and what it did.</description><category>AI Frontier</category><pubDate>Wed, 02 Sep 2026 15:10:21 -0600</pubDate></item><item><title>Gemini Live can now read you a spoken daily brief built from your Gmail and Calendar, and clear your inbox hands free</title><link>https://blog.google/innovation-and-ai/products/gemini-app/productivity-features-gemini-live/</link><guid isPermaLink="true">https://blog.google/innovation-and-ai/products/gemini-app/productivity-features-gemini-live/</guid><description>Ask &#x27;what&#x27;s my daily brief&#x27; and it talks you through the day without a screen. You can also search, summarize, star, archive and delete mail by voice, and hand multi-step work off into Docs, Sheets and Drive.</description><category>AI Frontier</category><pubDate>Wed, 02 Sep 2026 15:10:21 -0600</pubDate></item><item><title>Google DeepMind&#x27;s Co-Scientist runs a tournament of AI agents that argue over research ideas until the best one wins</title><link>https://deepmind.google/blog/co-scientist-a-multi-agent-ai-partner-to-accelerate-research/</link><guid isPermaLink="true">https://deepmind.google/blog/co-scientist-a-multi-agent-ai-partner-to-accelerate-research/</guid><description>One set of agents proposes hypotheses, another plays peer reviewer and tears them apart, and a ranking round pits them head to head before the survivors get refined. Teams have used it on liver fibrosis, ALS and aging. A clean pattern to copy: generate, debate, rank, repeat.</description><category>AI Frontier</category><pubDate>Wed, 02 Sep 2026 15:10:21 -0600</pubDate></item><item><title>affaan-m/ECC</title><link>https://github.com/affaan-m/ECC</link><guid isPermaLink="true">https://github.com/affaan-m/ECC</guid><description>The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor </description><category>What&#x27;s Being Built</category><pubDate>Wed, 02 Sep 2026 15:10:21 -0600</pubDate></item><item><title>NousResearch/hermes-agent</title><link>https://github.com/NousResearch/hermes-agent</link><guid isPermaLink="true">https://github.com/NousResearch/hermes-agent</guid><description>The agent that grows with you</description><category>What&#x27;s Being Built</category><pubDate>Wed, 02 Sep 2026 15:10:21 -0600</pubDate></item><item><title>deepseek-ai/deepseek-harness</title><link>https://github.com/deepseek-ai/deepseek-harness</link><guid isPermaLink="true">https://github.com/deepseek-ai/deepseek-harness</guid><description>DeepSeek Harness: Everything is a Plugin.</description><category>What&#x27;s Being Built</category><pubDate>Wed, 02 Sep 2026 15:10:21 -0600</pubDate></item><item><title>Efficient SWE Agent Benchmarking via Trajectory-Aware Evaluation</title><link>http://arxiv.org/abs/2609.01603v1</link><guid isPermaLink="true">http://arxiv.org/abs/2609.01603v1</guid><description>Research paper.</description><category>Research</category><pubDate>Wed, 02 Sep 2026 15:10:21 -0600</pubDate></item><item><title>CordisBench: Can Language Models Reason About Component Lifecycles in Dynamic Agent Harnesses?</title><link>http://arxiv.org/abs/2609.01600v1</link><guid isPermaLink="true">http://arxiv.org/abs/2609.01600v1</guid><description>Research paper.</description><category>Research</category><pubDate>Wed, 02 Sep 2026 15:10:21 -0600</pubDate></item><item><title>The Rise of Verbal Reinforcement Learning</title><link>http://arxiv.org/abs/2609.01597v1</link><guid isPermaLink="true">http://arxiv.org/abs/2609.01597v1</guid><description>Research paper.</description><category>Research</category><pubDate>Wed, 02 Sep 2026 15:10:21 -0600</pubDate></item></channel></rss>