OpenAI Launches GPT-6 With Interactive UI And Live Thinking
The shift from plain text to interactive elements and live thinking changes how users interact with AI models and impacts performance speed.
• OpenAI is rolling out GPT-6 globally with an interactive user interface instead of plain text.
• The model can respond while still thinking, which OpenAI says cuts wait times by 44 percent in internal tests.
• Paying customers receive GPT-6 Sol while free users get GPT-6 Luna, with availability starting today.
Read more
The shift from plain text to interactive elements and live thinking changes how users interact with AI models and impacts performance speed.
• OpenAI is rolling out GPT-6 globally with an interactive user interface instead of plain text.
• The model can respond while still thinking, which OpenAI says cuts wait times by 44 percent in internal tests.
• Paying customers receive GPT-6 Sol while free users get GPT-6 Luna, with availability starting today.
Read more
Anthropic Releases Claude Haiku 5.5 With Lower Prices And Major Performance Jumps
The release matters because significant price cuts and performance gains in small AI models directly affect the economics of high-volume and agentic tasks.
• Anthropic released Claude Haiku 5.5 with prices dropping by up to 90 percent for prompts under 100,000 tokens according to the company.
• The new small model scores higher than its predecessor across multiple benchmarks and features adjustable reasoning levels.
• Anthropic also cut cache read costs for Sonnet 5.5 by 50 percent and rolled out monthly API credits for subscribers.
Read more
The release matters because significant price cuts and performance gains in small AI models directly affect the economics of high-volume and agentic tasks.
• Anthropic released Claude Haiku 5.5 with prices dropping by up to 90 percent for prompts under 100,000 tokens according to the company.
• The new small model scores higher than its predecessor across multiple benchmarks and features adjustable reasoning levels.
• Anthropic also cut cache read costs for Sonnet 5.5 by 50 percent and rolled out monthly API credits for subscribers.
Read more
Microsoft Launches New Surface AI PCs Powered By Nvidia Chips
The release brings local Nvidia powered AI processing and developer oriented Windows 11 features directly to new hardware starting at $2,600.
• Microsoft revealed the Surface Laptop Ultra starting at $2,600 and the Dev Box starting at $6,000.
• Both machines run on Nvidia RTX Spark chips to process AI models locally on device.
• The devices feature a revamped Windows 11 with Execution Containers for sandboxing AI agents.
Read more
The release brings local Nvidia powered AI processing and developer oriented Windows 11 features directly to new hardware starting at $2,600.
• Microsoft revealed the Surface Laptop Ultra starting at $2,600 and the Dev Box starting at $6,000.
• Both machines run on Nvidia RTX Spark chips to process AI models locally on device.
• The devices feature a revamped Windows 11 with Execution Containers for sandboxing AI agents.
Read more
Microsoft Gives Copilot Access to Local Files and Deeper Windows Control
These updates give users deeper AI integration and faster system controls directly within Windows 11.
• Microsoft updated its Copilot AI to access local PC files and perform tasks across the operating system.
• The new features rely on Hybrid Intelligence, combining local and cloud AI models to handle multi-step user requests.
• A new Windows search experience arriving this fall allows quick actions like dark mode toggling and text messaging.
Read more
These updates give users deeper AI integration and faster system controls directly within Windows 11.
• Microsoft updated its Copilot AI to access local PC files and perform tasks across the operating system.
• The new features rely on Hybrid Intelligence, combining local and cloud AI models to handle multi-step user requests.
• A new Windows search experience arriving this fall allows quick actions like dark mode toggling and text messaging.
Read more
OpenAI Launches Decisions API For Faster Text And Image Evaluation
The new API provides faster and cheaper text and image evaluation while restructuring developer pricing tiers.
• OpenAI launched the Decisions API in public beta to evaluate text and images.
• The new API runs ten times faster than the Responses API and costs $0.10 per million input tokens.
• OpenAI also cut its paid API tiers from five down to three new levels.
Read more
The new API provides faster and cheaper text and image evaluation while restructuring developer pricing tiers.
• OpenAI launched the Decisions API in public beta to evaluate text and images.
• The new API runs ten times faster than the Responses API and costs $0.10 per million input tokens.
• OpenAI also cut its paid API tiers from five down to three new levels.
Read more
Hackers Used AI Agents to Breach Seoul Megachurches and Steal Data
The incident highlights an emerging threat where malicious actors use automated AI agents to execute large scale database raids against major religious and financial institutions.
• Yoido Full Gospel Church said data tied to 850,000 members may have been stolen in a hack.
• Oasis Security found attack records showing the activity of AI sub-agents and machine-written logs.
• President Lee confirmed AI is believed to have been used in recent hacks on South Korean commercial banks.
Read more
The incident highlights an emerging threat where malicious actors use automated AI agents to execute large scale database raids against major religious and financial institutions.
• Yoido Full Gospel Church said data tied to 850,000 members may have been stolen in a hack.
• Oasis Security found attack records showing the activity of AI sub-agents and machine-written logs.
• President Lee confirmed AI is believed to have been used in recent hacks on South Korean commercial banks.
Read more
GitHub Copilot local sandboxing is now generally available
GitHub says developers and organizations can set enforceable limits on what Copilot-run tools access on their machines.
• GitHub Copilot local sandboxing is generally available in its CLI, app and VS Code sessions using Agent Host.
• Policies can limit agent-run commands' access to files, networks and credentials.
• GitHub says enterprise settings can require sandboxing, and the feature costs no extra.
Read more
GitHub says developers and organizations can set enforceable limits on what Copilot-run tools access on their machines.
• GitHub Copilot local sandboxing is generally available in its CLI, app and VS Code sessions using Agent Host.
• Policies can limit agent-run commands' access to files, networks and credentials.
• GitHub says enterprise settings can require sandboxing, and the feature costs no extra.
Read more
Nous Research Secures $1.5B Valuation and Launches Enterprise AI Agents
Nous Research confirmed a $90 million Series B funding round and launched new tools for corporate customers.
• Nous Research raised a $90 million Series B round at a $1.5 billion valuation.
• The funding was led by Robot Ventures and included backers like Nvidia and Samsung.
• The startup launched Hermes for Businesses to let companies deploy private AI agents.
Read more
Nous Research confirmed a $90 million Series B funding round and launched new tools for corporate customers.
• Nous Research raised a $90 million Series B round at a $1.5 billion valuation.
• The funding was led by Robot Ventures and included backers like Nvidia and Samsung.
• The startup launched Hermes for Businesses to let companies deploy private AI agents.
Read more
AWS Launches Open-Source Strands Box to Restrict AI Agents
Autonomous AI agents often operate without human review, and this open-source tool gives developers deterministic ways to enforce rules and prevent destructive actions.
• AWS launched an open-source AI sandbox called Strands Box to control autonomous agents.
• The tool combines OS-level isolation with the Dogwood Local Engine for contextual rule enforcement.
• Strands Box is available on GitHub for macOS, with Linux and Windows support in development.
Read more
Autonomous AI agents often operate without human review, and this open-source tool gives developers deterministic ways to enforce rules and prevent destructive actions.
• AWS launched an open-source AI sandbox called Strands Box to control autonomous agents.
• The tool combines OS-level isolation with the Dogwood Local Engine for contextual rule enforcement.
• Strands Box is available on GitHub for macOS, with Linux and Windows support in development.
Read more
Microsoft releases Agent Lightning v1.0 for training AI agents with their existing harnesses
Microsoft says Agent Lightning v1.0 lets developers train existing AI agents without rebuilding their harnesses inside a training framework.
• Microsoft open-sourced Agent Lightning v1.0 to train AI agents with the harnesses they use in deployment.
• The roughly 3,500-line framework runs agents locally or as standard Kubernetes jobs.
• Microsoft says a coding-agent example improved Qwen3.5-9B by 14.6 percentage points on SWE-bench Verified.
Read more
Microsoft says Agent Lightning v1.0 lets developers train existing AI agents without rebuilding their harnesses inside a training framework.
• Microsoft open-sourced Agent Lightning v1.0 to train AI agents with the harnesses they use in deployment.
• The roughly 3,500-line framework runs agents locally or as standard Kubernetes jobs.
• Microsoft says a coding-agent example improved Qwen3.5-9B by 14.6 percentage points on SWE-bench Verified.
Read more
AI Tools Enabled Single Attacker to Breach South Korean Banks
This incident highlights the real-world cybersecurity risk that AI tools can enable a single attacker to execute massive data breaches.
• A suspected Chinese-speaking attacker breached multiple South Korean financial institutions between late September and early October 2026.
• The attacker used ARTEX, an open-source tool running AI language models that automates penetration testing and security flaw detection.
• Crowdstrike noted the case shows how AI tools allow a single person to execute massive breaches in a short window.
Read more
This incident highlights the real-world cybersecurity risk that AI tools can enable a single attacker to execute massive data breaches.
• A suspected Chinese-speaking attacker breached multiple South Korean financial institutions between late September and early October 2026.
• The attacker used ARTEX, an open-source tool running AI language models that automates penetration testing and security flaw detection.
• Crowdstrike noted the case shows how AI tools allow a single person to execute massive breaches in a short window.
Read more
Nvidia commits $1 billion to US science over five years
Nvidia says the five-year commitment will provide computing and research support to US institutions and government-related science projects.
• Nvidia says it will make commitments valued at $1 billion for US science over five years.
• The support covers university research, quantum computing and cloud providers serving US government missions.
• Nvidia is collaborating on Genesis Mission phase 2 projects in quantum computing, fusion and other fields.
Read more
Nvidia says the five-year commitment will provide computing and research support to US institutions and government-related science projects.
• Nvidia says it will make commitments valued at $1 billion for US science over five years.
• The support covers university research, quantum computing and cloud providers serving US government missions.
• Nvidia is collaborating on Genesis Mission phase 2 projects in quantum computing, fusion and other fields.
Read more
Google introduces the Gemini agent for work
Google says teams can assign work across apps to one agent while using company tools, shared context and enterprise controls.
• Google announced a Gemini agent that can answer questions, complete assigned work, create media and write and run code.
• Google says it works inside Workspace apps and carries memory and controls across devices and channels.
• Google says the agent can select among Gemini and Claude models and offers enterprise security and spending controls.
Read more
Google says teams can assign work across apps to one agent while using company tools, shared context and enterprise controls.
• Google announced a Gemini agent that can answer questions, complete assigned work, create media and write and run code.
• Google says it works inside Workspace apps and carries memory and controls across devices and channels.
• Google says the agent can select among Gemini and Claude models and offers enterprise security and spending controls.
Read more
Zenity Finds Single Prompt Compromised AWS AgentCore Accounts
According to Zenity, overly broad default permissions in cloud platforms can allow a single prompt to compromise multiple AI agents.
• Zenity Labs found a single prompt could hijack every AI agent in an AWS account and region through Amazon Bedrock AgentCore.
• The flaw allowed attackers to access private conversations, source code, and stored credentials due to overly broad default permissions.
• Zenity reported the issue to AWS in December 2025, prompting AWS to adopt IMDSv2 and restrict default execution roles.
Read more
According to Zenity, overly broad default permissions in cloud platforms can allow a single prompt to compromise multiple AI agents.
• Zenity Labs found a single prompt could hijack every AI agent in an AWS account and region through Amazon Bedrock AgentCore.
• The flaw allowed attackers to access private conversations, source code, and stored credentials due to overly broad default permissions.
• Zenity reported the issue to AWS in December 2025, prompting AWS to adopt IMDSv2 and restrict default execution roles.
Read more
China AI Lab Manus Raises Over $500 Million in First Funding After Meta Split
The funding round establishes financial independence for Manus after Chinese regulators forced the startup to abandon a $2 billion acquisition by Meta.
• Butterfly Effect announced that Manus raised over $500 million in its first funding round since a failed acquisition by Meta.
• Boyu Capital and IDG Capital led the round, with participation from Tencent, HSG, ZhenFund, and other existing shareholders.
• Chinese authorities previously ordered Manus to unwind the $2 billion Meta deal in April amid talent retention concerns.
Read more
The funding round establishes financial independence for Manus after Chinese regulators forced the startup to abandon a $2 billion acquisition by Meta.
• Butterfly Effect announced that Manus raised over $500 million in its first funding round since a failed acquisition by Meta.
• Boyu Capital and IDG Capital led the round, with participation from Tencent, HSG, ZhenFund, and other existing shareholders.
• Chinese authorities previously ordered Manus to unwind the $2 billion Meta deal in April amid talent retention concerns.
Read more
Waymo secures a $5 billion loan to fund robotaxi expansion
The $5 billion loan provides Waymo with financial flexibility to scale its robotaxi business as it expands into international markets and faces federal safety scrutiny.
• Waymo closed its first debt financing with a $5 billion loan from lenders including PIMCO and Blackstone.
• The Alphabet subsidiary will use the funds to scale its commercial robotaxi operations in the U.S., Europe, and Japan.
• Federal safety regulators are investigating Waymo following incidents involving school buses and a collision with a child.
Read more
The $5 billion loan provides Waymo with financial flexibility to scale its robotaxi business as it expands into international markets and faces federal safety scrutiny.
• Waymo closed its first debt financing with a $5 billion loan from lenders including PIMCO and Blackstone.
• The Alphabet subsidiary will use the funds to scale its commercial robotaxi operations in the U.S., Europe, and Japan.
• Federal safety regulators are investigating Waymo following incidents involving school buses and a collision with a child.
Read more