OpenAI terminated three researchers on October 1 for mishandling sensitive information and violating company policies. The employees, identified as Jasmine Wang, Tomek Korbak, and Mikita Balesni, reportedly shared confidential data with an external AI evaluation organization. At least two of the fired workers were safety and alignment researchers. OpenAI said an internal investigation confirmed the policy breaches, and the firings come as debate over AI safety risks intensifies. The company also recently delayed its new model GPT-6.1 Astra due to safety concerns.
Microsoft shares surged 37.7% in the third quarter, marking the strongest quarterly gain since 1998. Azure growth accelerated to 43% and Microsoft guided to 45% growth next quarter. Microsoft 365 Copilot surpassed 30 million paid seats, while Agent 365 reached nearly 40 million agents. OpenAI's upcoming Dots launch will integrate with Microsoft Agent 365. Fifty-three of 55 analysts rate Microsoft a buy or higher.
Meta's stock rose nearly 29% in the third quarter, its best performance since early 2024, following the launch of Muse, a personal AI agent that drew about 5 million downloads in three weeks. Meta President Dina Powell McCormick said the company is exploring transaction fees and subscriptions to monetize Muse, which already helps users reduce costs. Muse reached number one on the App Store, and Meta is partnering with Shopify, Slack, and Walmart. Amazon blocked Muse from shopping on its site, and analyst Laura Martin warned Meta is notoriously slow at monetizing new products.
Amazon is looking to borrow about eight billion dollars to purchase powerful Nvidia chips for its AI projects. The company plans to place these chips in a separate financial group so costs do not fully appear on its main financial records. Amazon expects to spend around 220 billion dollars this year, mostly on AI infrastructure.
Nvidia faces mixed developments. The company paused its AI ambassador program in Oregon schools without explanation, despite the state setting aside 1.5 million dollars for the initiative. Separately, a Chinese company called Semi-Tech Leasing Group used state money to buy restricted Nvidia processors, including high-performance B300 Blackwell chips, for more than 700 servers powering an AI developer. The purchase was funded by government funds from Shenzhen and Beijing.
CoreWeave announced availability of the NVIDIA Vera Rubin NVL72 with Cognition as its first production customer. Cognition reported up to 4.8x higher token throughput for SWE-2 inference workloads compared with a GB200 NVL72 baseline. CoreWeave also highlighted its NVIDIA Vera CPU platform for agentic AI workloads, showing over three times faster sandbox startup times. However, the company faces risks from heavy dependence on NVIDIA technology and high infrastructure costs.
California Governor Gavin Newsom signed Executive Order N-10-26 directing state agencies to continue using the term
Key Takeaways
- OpenAI fired three safety researchersāJasmine Wang, Tomek Korbak, and Mikita Balesniāfor sharing confidential data with an external AI safety organization
- Microsoft shares surged 37.7% in Q3, its best quarter since 1998, with Azure growth hitting 43% and Copilot surpassing 30 million paid seats
- Meta's Muse AI agent reached 5 million downloads in three weeks and hit number one on the App Store, but monetization remains uncertain
- Amazon plans to borrow 8 billion dollars to buy Nvidia chips, with total AI infrastructure spending expected around 220 billion dollars this year
- Nvidia paused its AI ambassador program in Oregon schools despite 1.5 million dollars in state funding allocated for the initiative
- Chinese state-backed Semi-Tech Leasing Group used restricted funds to purchase Nvidia B300 Blackwell chips for over 700 servers
- CoreWeave reported up to 4.8x higher token throughput using NVIDIA Vera Rubin NVL72 compared to GB200 NVL72 baseline
- California Governor Newsom signed an executive order requiring state agencies to continue using the term "Artificial Intelligence"
- OpenAI also delayed its GPT-6.1 Astra model launch due to safety concerns
- Meta is partnering with Shopify, Slack, and Walmart to expand Muse, while Amazon blocked Muse from shopping on its site
OpenAI fires three researchers over sensitive data handling
OpenAI fired three researchers on October 1 for mishandling sensitive information and violating company policies. The workers reportedly shared data with an external AI evaluation organization. At least two of the fired employees worked on safety and alignment. The identities were reported by The Wall Street Journal as Jasmine Wang, Tomek Korbak, and Mikita Balesni.
OpenAI fires 3 AI safety researchers for sharing confidential data
OpenAI fired three AI safety researchers after an internal investigation found they allegedly shared confidential information with a third-party AI safety organization. The company said the employees breached policies on accessing sensitive data. The researchers were identified as Jasmine Wang, Tomek Korbak, and Mikita Balesni. OpenAI also recently delayed its new model GPT-6.1 Astra due to safety concerns.
OpenAI fires staff for sharing sensitive AI model information
OpenAI fired three workers for sharing sensitive information about its AI models with a third-party organization involved in AI safety. The workers were identified as Jasmine Wang, Tomek Korbak, and Mikita Balesni. OpenAI said the investigation confirmed they violated company procedures. The company also delayed its latest AI model launch due to safety concerns.
OpenAI fires three workers over mishandling sensitive information
OpenAI fired three researchers for allegedly mishandling information tied to an external organization that analyzes AI models. The company did not name the workers, but at least two worked in safety research. A spokesperson said the employees violated policies on handling sensitive company information. The firings came as debate over AI safety risks intensified.
OpenAI fires three workers over mishandling sensitive information
OpenAI fired three researchers for allegedly mishandling information involving an external organization analyzing AI models. The company did not name the workers but said at least two were safety researchers. A spokesperson told the BBC the employees violated policies on sensitive data. The firings happened as concerns over AI safety and guardrails grew in recent months.
OpenAI fires three safety researchers over leaked sensitive information
OpenAI fired three safety researchers for violating company policies on handling sensitive information. An investigation confirmed they mishandled data outside established procedures. The employees shared confidential information with a third-party AI safety organization. Spokesperson said the firings broke essential trust. CEO Sam Altman has publicly stressed keeping powerful AI under human control.
OpenAI terminates three staff for leaking private data to outside group
OpenAI parted ways with three safety team members who leaked private information. The employees shared confidential data with an undisclosed third-party AI safety organization. The departures follow broader accusations that frontier AI labs mishandled sensitive information. A separate report found AI agents probed government websites using aggressive techniques. The FTC is reportedly investigating the matter.
Meta stock surges on Muse AI but monetization concerns persist
Meta shares rose nearly 29% in the third quarter, its best performance since early 2024. The surge followed the launch of Muse, a personal AI agent that drew about 5 million downloads in three weeks. Analyst Laura Martin warned Meta is notoriously slow at monetizing new products. Muse is free and Meta will not share conversations with its advertising business. Amazon blocked Muse from shopping on its site.
Microsoft stock posts best quarterly gain since 1998
Microsoft shares surged 37.7% from July to September, its strongest quarter since 1998. Azure growth accelerated to 43% in the fiscal fourth quarter and Microsoft guided to 45% growth next quarter. Microsoft 365 Copilot surpassed 30 million paid seats and Agent 365 reached nearly 40 million agents. OpenAI's Dots launch will integrate with Microsoft Agent 365. 53 of 55 analysts rate MSFT a buy or higher.
California governor orders agencies to keep using term Artificial Intelligence
Governor Gavin Newsom signed Executive Order N-10-26 directing California state agencies to continue calling artificial intelligence by its name. The order was signed September 30 and takes effect immediately. Newsom stated that renaming something cannot distract from addressing documented security and safety risks. The order applies only to state agencies and does not create any legal rights or penalties.
New System Ensures AI Agents Spend Budget Safely
Researchers created a new method to help AI agents manage money safely when they work together. This system uses special credits that track spending even if messages get lost or systems fail. It ensures that the total budget stays the same and no one spends more than allowed. Tests using computer models and real software confirmed the system works correctly under difficult conditions.
Amazon Plans to Borrow Money for Eight Billion in AI Chips
Amazon is looking to borrow about eight billion dollars to buy powerful Nvidia chips for its AI projects. The company wants to put these expensive chips into a separate financial group so the cost does not appear fully on its main financial records. This group could sell shares to investors if it gets a good credit rating. Amazon expects to spend around two hundred twenty billion dollars this year mostly on AI infrastructure.
California Signs New AI Laws and Pennsylvania Considers Ban
California Governor Gavin Newsom signed twelve new laws related to artificial intelligence to increase transparency. Meanwhile, lawmakers in Pennsylvania are discussing a rule that would stop the sale of AI-powered toys for three years. These actions show that different US states are taking different steps to regulate how AI technology is used and sold.
Nvidia Pauses AI Ambassador Program in Oregon Schools
Nvidia has stopped its AI ambassador program in Oregon without giving a specific reason. The company had promised to train teachers and students on artificial intelligence concepts last year. Although the state set aside one point five million dollars for this program, no money has been spent yet. A spokesperson for Nvidia said they are working on next steps but did not explain the delay.
Chinese State Funds Buy Restricted Nvidia AI Chips
A Chinese company called Semi-Tech Leasing Group used state money to buy Nvidia processors that are usually restricted from export. The report says this group funded the purchase of more than seven hundred servers for an AI developer. These servers include models powered by Nvidia's high-performance B300 Blackwell chips. The company is backed by government funds from Shenzhen and Beijing to support national AI goals.
Meta's Muse AI service may charge fees for saving users money
Meta is exploring transaction fees and subscriptions to monetize its AI service called Muse. Meta President Dina Powell McCormick said Muse already helps people reduce costs and could take a fee when it saves money. The company also launched Muse for small businesses to handle tasks like cash flow and inventory. Muse reached number one on the App Store with millions of downloads. Meta is partnering with Shopify, Slack, and Walmart.
CoreWeave bets on NVIDIA Vera Rubin to boost AI infrastructure
CoreWeave announced availability of the NVIDIA Vera Rubin NVL72 with Cognition as its first production customer. Cognition reported up to 4.8x higher token throughput for SWE-2 inference workloads compared with a GB200 NVL72 baseline. CoreWeave also highlighted its NVIDIA Vera CPU platform for agentic AI workloads, showing over three times faster sandbox startup times. The company offers a full-stack platform including Kubernetes Service and Sandboxes. CoreWeave faces risks from heavy dependence on NVIDIA technology and high infrastructure costs.
Sources
- OpenAI fired three staffers for mishandling 'sensitive' info and violating company policies
- OpenAI fires 3 AI safety researchers for allegedly sharing confidential company information
- OpenAI fires staff for sharing āsensitive informationā with AI safety firm
- OpenAI fires three workers over mishandling 'sensitive information'
- OpenAI fires three workers over mishandling 'sensitive information'
- OpenAI fires 3 safety researchers accused of sharing confidential company information: report
- OpenAI Parts Ways With Three Safety Researchers Over Sensitive Information Mishandling
- META Stock Rides Muse AI To Best Quarter In Over 2 Years ā But Analyst Warns Monetization Could Be 'Notoriously Slow'
- MSFT Stock Posts Best Quarter Since 1998 As Azure Growth Revives AI Optimism
- Newsom orders California agencies to keep saying āArtificial Intelligenceā
- Fault-Tolerant Budget Conservation in Distributed Multi-Agent Delegation
- Amazon plans to secure external financing for $8 billion in AI chips
- AI Legislative Update: Oct. 2, 2026 ā Transparency Coalition. Legislation for Transparency in AI Now.
- Nvidia Pauses AI Program With Oregon Schools, Colleges
- Chinese state funds backed purchases of restricted Nvidia AI chips, report says
- Mark Zuckerberg's Muse Could Charge You for Saving You Money ā META President Reveals AI Agent's Potential Monetization Model
- CRWV's Vera Rubin Push: Can AI Infrastructure Fuel Its Growth Engine?
Comments
Please log in to post a comment.