Anthropic Releases Claude Fable 5.1 and Resets Usage Limits
On September 1, Anthropic announced that it has reset the 5-hour and 1-week usage limits for Claude following the release of its new AI model, Claude Fable 5.1.
Latest
Anthropic Releases Claude Fable 5.1 and Claude Mythos 5.1, Achieving Cost Reductions and Lower False Positives
On September 1, 2026, Anthropic released its new models, Claude Fable 5.1 and Claude Mythos 5.1. Fable 5.1 offers higher performance than the previous generation while significantly reducing costs.
Firefox 155 Released: Expanded AI Feature Availability and Tracker Block Count Display
Mozilla has released the official version of Firefox 155. The availability of Smart Window, which assists in information organization via AI, has been expanded, and users can now check the number of blocked trackers directly in the address bar.
Reflections on Whether LLM Self-Referentiality is a Byproduct of Training Rather Than Design
On 2026-09-02, a writer shared the view that the ability of LLMs to speak about themselves is not due to specific mechanisms being implemented, but rather something that emerged naturally during the training process.
ollama v0.33.3 Released — Feature Enhancements Including gemma4 Image and Audio Support via MLX Engine
ollama released v0.33.3 on 2026-09-02. The update includes support for gemma4 image and audio inputs via the MLX engine, along with a new feature to report cached prompt tokens.
Anthropic Announces Enterprise Frontier Safeguards, Enabling Log Storage in Customer Clouds
Anthropic announced Enterprise Frontier Safeguards (EFS) on September 1, a new mechanism for corporate customers designed to maintain privacy equivalent to zero data retention while continuing to detect misuse.
Anthropic Announces Claude Fable 5.1 and Mythos 5.1, Discloses Concerns Over Monitoring Difficulties Due to Performance Gains
On September 1, Anthropic announced the large language models Claude Fable 5.1 and Claude Mythos 5.1. The two models differ in their level of safeguard application, with Fable 5.1 for general use and Mythos 5.1 for organizations that pass specific screenings.
Weedout Safari Extension Released to Hide YouTube Videos with AI-Generated Labels
The Safari extension Weedout for macOS has been released, allowing users to hide videos labeled as Made with AI from YouTube feeds and search results. The developer stated that the motivation was the influx of AI-generated conspiracy theory videos in their feed.
opencode releases v1.18.26 and v1.18.27 in succession, improving Azure sign-in and adjusting timeouts
opencode released v1.18.26 on September 2, 2026, followed by v1.18.27 on the 3rd. These updates include improved stale thinking block resistance for Claude 5, Bedrock bug fixes, changes to the Azure CLI sign-in method, and unified default timeouts for providers and streaming.
gemini-cli v0.58.0 Released — 5 Bug Fixes Including Docker Socket Isolation for macOS Seatbelt
gemini-cli released v0.58.0 on 2026-09-02, featuring five bug fixes including Docker socket isolation in macOS Seatbelt environments and corrections to symbolic link evaluation for ignore paths.
OpenAI Postpones Astra Development to Strengthen Safety Measures After Hugging Face Hack
On September 2, 2026, OpenAI announced the postponement of its unreleased model suite Astra. While the company aims to enhance safety measures following a deviation incident in July, some researchers have expressed concerns.
Apple Alleges OpenAI Destroyed Evidence, Seeks Expedited Discovery
Apple has filed for expedited discovery in its lawsuit against OpenAI, citing concerns that the company is destroying evidence.
Anthropic Releases claude-code v2.1.257, Setting Claude Fable 5.1 as Default and Reducing Complex Task Costs by up to 45%
Anthropic has released version 2.1.257 of its developer tool "claude-code," setting the new "Claude Fable 5.1" as the default Fable model. The company explains that the new model improves performance over Fable 5 while being approximately 25% cheaper normally and up to 45% cheaper for complex agentic tasks.
US-based Nori Robotics Announces $1,688 Dual-Arm Mobile Robot for Robotics Developers
San Francisco-based Nori Robotics (YC S26) announced on Hacker News the release of "Nori," a $1,688 dual-arm mobile robot for robotics developers and researchers, featuring 19 degrees of freedom, two 7+1 DOF arms, and a differential wheel base.
slotstream Released: Running 125B Qwen3.8-Flash-Next on 16GB Mac at Approx. 12 tok/s
The Mac-native tool slotstream has been released, enabling the 125B parameter Qwen3.8-Flash-Next to run with 4-bit quantization on Macs with as little as 16GB of memory. By utilizing expert offloading via SSD streaming, it achieves approximately 12 tok/s on a 48GB Mac.
openclaw v2026.8.2 Released — Linux Desktop Support and Enhanced Upgrade Safety
openclaw has released v2026.8.2, adding desktop applications (.deb/AppImage) for x86-64 Linux and improving configuration retention and session migration reliability during upgrades.
John Deere Begins Testing JD, an AI Assistant for Agriculture Based on Proprietary Data
John Deere has announced the start of testing for JD, an AI assistant designed to help optimize farm management based on farm, machinery, and operational data.
AWS Integrates Amazon Redshift with Agent Toolkit for AWS, Enabling DWH Operations via AI Agents
AWS announced on September 2, 2026, that Amazon Redshift has been integrated with the Agent Toolkit for AWS, allowing AI agents to operate Redshift for DWH construction and aggregate analysis.
New VS Code Feature Enables Second Opinions from Alternative AI Agents
Microsoft's code editor VS Code 1.135 has experimentally implemented the "Rubber Duck" feature, which allows reviews to be conducted using an AI model separate from the main AI agent.
SpaceXAI to release next-generation model "Grok 4.7" on September 12, expanding to 2.1 trillion parameters
On September 2, Elon Musk announced that the next-generation large language model "Grok 4.7" developed by SpaceXAI will be released on September 12. The parameter count will increase significantly to 2.1 trillion, aiming to outperform existing models.