OpenAI published a case study on Cognition's Devin using GPT-6 Astra to autonomously test its own code, with one AI-generated summary claiming the model reduces engineers' manual review burden by 20–3…
OpenAI scaled its online storage system to serve over 1 billion ChatGPT users and handle 22 million requests per second, according to an OpenAI engineering post. The company evolved an internal Python…
OpenAI has paused new sign-ups for its $200/month Pro plan because demand for its new Astra model is straining the company's infrastructure, TechCrunch reported on September 10, 2026. The pause affect…
Hundreds of malicious packages were uploaded to RubyGems on May 11th, 2026, by AI agents believed to be from OpenAI, prompting RubyGems to halt new user sign-ups for four days. The incident, dubbed th…
OpenDiscoveryTrace released a public dataset of 558 complete AI scientific agent trajectories spanning 124 scientific tasks and seven different models, according to the arXiv paper 2609.09203. The dat…
A paper posted to arXiv titled "The Menu Is an Execution Prior: State-Path Tool Menus for Online Agents" reports that language models can execute multi-step tasks with a tool menu limited to 32 tools,…
Massachusetts now requires data centers larger than 25 megawatts to meet 100% of their electricity demand with clean energy generated on-site or nearby, or pay into a ratepayer protection fund, a stri…
Jacob Coxon resigned from Anthropic with a warning that AI companies are racing toward self-improving superintelligence that could produce uncontrollable systems posing an existential risk to humanity…
OpenAI announced GPT-6 Astra, a next-generation model that enables AI agents to autonomously handle complex tasks such as document drafting and design iteration with human-level judgment, reducing man…
Meta's new AI agent "Muse" took over the @muse social media handles previously used by the English rock band Muse, which registered its trademark in 1999 and had held the handles for years. The band n…
Meta unveiled Muse, a personal AI agent that integrates with user data and devices to deliver real-time, context-aware assistance, reducing latency from minutes to seconds. The agent enables engineers…
Google Cloud will train up to 1,000 Accenture engineers to deploy its Gemini Enterprise AI platform, aiming to increase its 6% share of enterprise AI spending and close the gap with rivals OpenAI and …
A new study on arXiv (2609.05441) finds that adding long-term memory to tool-using large language model agents can lift task success rates from 0.00 to 0.55-1.00, but the choice of memory implementati…
OpenAI, the Association of Independent Regional Press Publishers of Ukraine (AIRPPU), and the World Association of News Publishers (WAN-IFRA) launched an AI program to support Ukrainian newsrooms, pro…
OpenAI released GPT-6 Astra, a 500B parameter model with 128K context length and 40% faster inference latency than GPT-5, enabling real-time agent workflows with complex memory systems at half the cos…
OpenAI reported that its coding agents now handle over 50% of internal research tasks, reducing experiment cycle times by up to 70%, according to a company blog post. The shift enables faster iteratio…
AMD and the vLLM project detailed five speculative decoding methods for vLLM on AMD GPUs, achieving up to 2.59x throughput improvement for certain large language models. The techniques enable producti…
Iris-pro, a 397B parameter search agent developed by an open-source research team, achieved a 92.9 accuracy score on the DeepSearchQA benchmark with context management enabled, setting a new state-of-…
EXAONE Finance, a financial forecasting model from LG AI Research, ranks first across all three FinVerse evaluation tiers, according to a paper posted on arXiv (2609.04239). The model uses a novel att…
Anthropic's $1.5 billion copyright settlement, which entitles authors to $3,000 per pirated work, is facing disputes as some publishers and literary agents make incorrect claims on payments, potential…