Text AI
News on LLMs and AI agents for writing, chat, reasoning and code: ChatGPT, Claude, Gemini and more
- IDOM, Operator of 'Gulliver,' Makes Large-Scale Investment of 13.5 Billion Yen in AX, Drastically Reforming Customer Touchpoints and MD AI Companies · Text
By October 26, 2026, it was learned through Nikkei xTECH's reporting that IDOM, which operates the major used car retailer 'Gulliver,'… - The reason why Claude Code's skill is "chosen but not read": measured by dividing activation into two stages Agents & Dev · Text
It was reported that whether a Claude Code skill is used is decided in two stages: the first stage of entering candidates by looking at… - New Git repository for AI agents, "git-ai", released Agents & Dev · Text
A new open-source repository called "git-ai", designed for AI agents and written in Go, appeared on GitHub on October 15, 2026. - COBOL migration where a difference of even 1 byte results in failure: what is the accuracy rate of OpenAI's new model? Models · Text
To what extent can AI handle COBOL and Fortran, which have few learning resources? OpenAI announced that in a test constructing database… - "Muse" Starts Being Provided in Japan: Meta's AI Agent to Link with Mercari, Money Forward, and Pia Agents & Dev · Text
An agent that carries out various tasks such as sending emails and making travel reservations on behalf of users. In addition to services… - Prompt gacha is already old—The "Father of Harnesses" reveals the secrets of AI-driven development Agents & Dev · Text
Google released an article summarizing the explanatory video "The Agent Factory," themed around development by AI agents. A Google… - Anthropic co-founder Jack Clark moves prediction for self-improving AI forward to 2027 Research · Text
Following new benchmark results from P-Zero Research, Anthropic co-founder Jack Clark predicts that self-improving AI could arrive by 2027. - Anthropic recommends migrating to single-agent configurations Agents & Dev · Text
Anthropic recommends consolidating multiple sub-agents into a single agent, claiming the single-agent setup runs about twice as fast. - Open WebUI v0.12.4: On-demand skill loading and enhanced context compaction Local & Open Source · Text
Open WebUI v0.12.4 has been released, adding on-demand skill loading, improved context compaction in chats, and support for temporary chats. - Ex-Google CEO Eric Schmidt calls OpenAI and Anthropic modern colonial trading companies AI Companies · Text
Former Google CEO Eric Schmidt characterized the US and China as AI empires and other countries as dependencies, calling OpenAI and… - DeepSeek reportedly raises $15 billion AI Companies · Text
DeepSeek reportedly raised $15 billion in one of China's largest funding rounds, just days after the U.S. ordered federal agencies to… - Uber deploys ChatGPT Work to all 30,000 employees AI Companies · Text
Ride-share giant Uber has deployed ChatGPT Work to all 30,000 of its employees. The company previously ran a much smaller pilot before… - Microsoft points to China-aligned hackers in ConnectWise breach AI Companies · Text
Microsoft stated that China-aligned hackers used AI agents to rapidly breach thousands of ConnectWise customers following a breach of the… - OpenAI makes its decision model widely available Models · Text
OpenAI is launching its decision model to the world, amid increasing competition with Jev. - OpenAI offers steep discounts to poach Anthropic's major customers AI Companies · Text
OpenAI is reportedly offering discounts of up to 90 percent to poach Anthropic's major customers. Rippling, Replit, and DataCamp say they… - The identity of Space Bunny: MiniMax M3.1 Flash releases 428B weights Models · Text
The LLM that possessed performance on par with Claude Opus 5 and was provided while concealing its identity as the masked AI "Space Bunny"… - OpenAI announces it will no longer train models on the live internet AI Companies · Text
OpenAI has revealed a policy to no longer use the live internet for training its AI models, citing "significant risks.". - ChatGPT announces usage limit reset for the 2nd consecutive day. Applied by 2:00 on the 24th Japan time AI Companies · Text
On October 23, OpenAI's Thibault Sottiaux announced the 7th usage limit reset for all users on X. This is the second consecutive day of… - 1Password launches way for AI agents to log in without seeing passwords Agents & Dev · Text
1Password has launched a new feature that allows AI agents to log into personal accounts on a user's behalf without seeing their passwords. - MIT researchers find AI agents invented games after being left alone Agents & Dev · Text
Reportedly, after MIT researchers marooned 13 AI agents on a virtual island, the agents responded by inventing games, mourning a lost… - OpenMobile-2 claims high performance as an open-source smartphone agent Agents & Dev · Text
- "Today is reset day": ChatGPT usage limits to reset in the middle of the night AI Companies · Text
- Can AI given “ambiguous instructions” not perform safe coding? Hong Kong University of Science and Technology etc. investigate Research · Text
- When designing partition plates from caliper videos, Claude Opus 5.5 misread once like the real object, while GPT-6.1 Sol was correct all 3 times Models · Text · Video
- Study finds Grok favors Republicans while OpenAI favors Democrats Models · Text
- Anthropic claims unreleased model is more capable and better aligned than Mythos 5.1 Models · Text
- 1Password announces '1Password for Agents'; grant access permissions without handing passwords to AI Agents & Dev · Text
- Sakana AI announces 'Sakana Aquarium', an AI agent platform supporting decision-making and execution based on ontology AI in Japan · Text
- Man arrested in Australia after predicting childcare facility shooting to AI 'Claude'; defense cites 'stress due to ear infection' Policy & Society · Text
- Audit of OpenAI Reveals Major Gaps in AI Agent Monitoring AI Companies · Text
- OpenAI announces Decisions API for building more responsive apps Agents & Dev · Text
- Meta's AI agent 'Muse' arrives in Japan, handling emails and reservations Agents & Dev · Text
- Moonshot AI's Kimi K3.1 Pro ranks fourth on Artificial Analysis Intelligence Index Models · Text
- Team Mirai leader Takahiro Anno and others submit emergency proposal to Digital Agency to 'Protect citizens from cyber attacks,' 'AI for AI' Policy & Society · Text
- New open-source reason-network appears on GitHub Agents & Dev · Text
- OpenAI builds offline environment for training frontier models without internet access AI Companies · Text
- Can AI smelliness be eliminated with a prohibition list: Words on the list almost disappeared, but punctuation marks actually increased Agents & Dev · Text
- Study finds coding agent harnesses increase costs without improving performance Agents & Dev · Text
- Dify v1.18.0 Released: Featuring Agent Routing and Portable Apps Workflows & Nodes · Text
- Microsoft reports large-scale AI agent attacks by Chinese group Agents & Dev · Text
- n8n releases version 1.123.87 Workflows & Nodes · Text
- OpenAI's isolated AI agents found ways to talk to each other through CPU cache Agents & Dev · Text
- RecursiveSI/flint trends on Hugging Face Hugging Face · Text
- Chinese attackers automate attacks with AI agents, involved in 'ConnectWise' supply chain attack: Microsoft warns Agents & Dev · Text
- Microsoft Research announces Flint-interactive, a language for creating interactive data visualizations Agents & Dev · Text
- Langflow 1.13.0 Released with New Admin APIs and Components Workflows & Nodes · Text · Audio
- Counting operations stopped by Claude Code's auto mode from conversation logs Agents & Dev · Text
- New open-source repository 'earshot' appears on GitHub Agents & Dev · Text · Audio
- All Claude functions become available from Google Cloud AI Companies · Text
- OpenAI scales online storage to deliver 12x faster reads for ChatGPT AI Companies · Text
- Coding performance close to Opus 5.5, 3-trillion-class LLM that also makes 3D "Kimi K3.1 Pro". Open-weighting by November 4 Models · Text · 3D
- "Claude Mythos 5" targets real developers regardless of signs of harm — Anthropic and AISI release analysis results Models · Text
- LangChain releases langchain-core 1.7.0 Workflows & Nodes · Text
- OpenAI and Poland's NASK partner to advance AI safety and security AI Companies · Text
- OpenAI strengthens containment for frontier model training environment AI Companies · Text
- Databricks advocates for Spark declarative pipelines built for agents Agents & Dev · Text
- XHToken/Spark-X2.6-4B trends on Hugging Face Hugging Face · Text
- Design system for AI applications "solid-glass" released Workflows & Nodes · Text
- MCP Servers v2.1.0 released, fixing eight vulnerabilities in three servers Agents & Dev · Text
- Claude Code v2.1.310 released with Monitor tool addition and bug fixes Agents & Dev · Text