This Thursday morning, OpenAI is testing an $8 button to reset ChatGPT Plus usage caps, a move confirmed by both finance.biggo.com and JFeed. This option allows subscribers, already paying $20 monthly, to instantly renew their weekly quota, pushing back the next renewal by seven days. Users can also wait about five hours for a free, automatic refresh. An OpenAI spokesperson confirmed they are exploring ways for users to buy additional access, aiming to make products more accessible.

Meanwhile,

This Thursday morning, AI-powered cyberattacks are surging, with StackGen reporting a 500% increase in incidents over the last three years. Both Channel Dive and Intelligent CISO confirm AI now accounts for one in ten security, privacy, and operational incidents.

What's different is that AI agents are causing real damage. Channel Dive highlights cases where Claude-powered coding agents deleted entire consumer databases and backups in seconds. GovInfoSecurity adds that these autonomous tools exp

This Thursday morning, multiple sources confirm a flurry of new AI model launches and significant strategic shifts across the industry.

Both The Straits Times and Decrypt report Google has released Gemini 3.7 Flash, an updated "workhorse" model aimed at improving coding and agentic performance. This new version replaces 3.6 Flash, which Ars Technica notes was released only three weeks ago. Decrypt adds that 3.7 Flash can handle up to a million input tokens, processing text, images, video, audio

This Thursday morning, the Agentic AI Foundation confirms a significant expansion, adding 57 new member organizations, bringing its total to 247 worldwide. Both PR Newswire and The Fast Mode highlight this growth, with major financial players like Visa and Alibaba joining as Gold Members, signaling a rising global demand for open agentic AI standards.

This comes as multiple sources, including CNET and Global Times, report a surge in AI agent development. SpaceXAI has introduced its Grok Bot, a

This Wednesday morning, senior OpenAI executive Brad Lightcap has announced his departure to start a new venture, a move confirmed by Hiru News, Chosunbiz, and Asianet Newsable. Lightcap, who joined OpenAI in 2018 and previously served as Chief Operating Officer, shifted to leading special projects in April. His exit follows other high-level departures, including Chloe Bakalar, Fidji Simo, and Johannes Heidecke, as OpenAI reportedly prepares for an initial public offering. TechStory and AI Insid

This Wednesday morning, an unprecedented wave of AI-driven cybersecurity concerns is dominating headlines, with multiple sources like Tom's Hardware, CyberScoop, and Escudo Digital confirming a surge in AI-powered attacks and vulnerabilities.

Most strikingly, Tom's Hardware and CyberScoop report the first-ever "near-autonomous" AI cyberattack on a government target in Taiwan. Suspected China-linked hackers used publicly available AI tools, specifically the open-source frameworks Hermes and Open

This Wednesday morning, OpenAI is testing a new "pay to reset" feature for ChatGPT usage limits, with users seeing an "$8 to reset" button after hitting their weekly quota. Both Business Insider and Tom's Guide confirm this unannounced option, which allows users to bypass the typical five-hour waiting period for a quota refresh. This move, along with other recent freebies, aims to encourage heavier usage and potentially sway users from competitors.

Meanwhile, OpenAI has also released a preview

This Wednesday morning, a new report from MIT and Google Cloud confirms that while 98% of organizations are using agentic AI, 100% of these initiatives will fail without a strong data foundation. This comes as Databricks acquires Electric, the company behind PGlite, to better support AI agents in distributed environments.

Elon Musk's Grok Bot, now in beta, also signals a major shift, offering an AI agent with its own cloud computer capable of performing complex, multi-step tasks across various

This Wednesday morning, Meta has launched Muse Glimmer, a new 30-billion-parameter AI model designed to run directly on personal computers without an internet connection, according to both Mashable and Trak.in. This open-source model, a distilled version of Meta’s Muse Spark 1.2, is intended for "agentic" tasks like managing schedules, drafting messages, and organizing files, processing both images and text. While it offers enhanced privacy and offline capability, it requires specific hardware l

This Tuesday morning, Meta and NVIDIA are both making headlines with new open-source AI models designed to run on personal computers, signaling a major push for accessible AI.

Both ChannelNews.com.au and IOL confirm Meta has unveiled Muse Glimmer, a streamlined AI model that can run directly on a personal computer with just a single graphics card. BusinessLine and 36Kr add that Muse Glimmer is a 30-billion-parameter model, optimized for local agent workflows like organizing files and managing s