The evening briefing.
Today across AI and tech: OpenAI launches a security initiative, Google thwarts AI-powered hackers, and GM lays off IT staff to hire for AI skills.
AI Security. OpenAI launched Daybreak, a new initiative to detect and patch vulnerabilities using AI agents, directly competing with Anthropic's security-focused models. This comes as Google reported thwarting a hacker group that used AI to discover a zero-day vulnerability, marking a significant shift in cyber threats. Meanwhile, cookie thieves were found using fake Claude Code installers to steal developer secrets.
AI Industry Shifts. The enterprise adoption of AI is reaching a tipping point, according to OpenAI's revenue chief, as the company also launched an AI consulting arm to help businesses deploy the technology. GM is undergoing a significant workforce transformation, laying off hundreds of IT workers to recruit talent with stronger AI-native development and prompt engineering skills. This highlights the rapid re-skilling demand in the industry.
AI Research & Infrastructure. Former OpenAI CTO Mira Murati's new company, Thinking Machines, unveiled "interaction models" designed for real-time, collaborative AI experiences. However, Microsoft researchers found that current AI models and agents struggle with long-running, complex tasks, indicating fundamental limitations. NVIDIA also introduced Fleet Intelligence to optimize GPU fleet visibility and performance, crucial for scaling AI infrastructure.
Policy & Investment. The ongoing legal battle between Elon Musk and OpenAI saw Microsoft CEO Satya Nadella testify that Musk never raised concerns about Microsoft's investment. In other news, India's state-run ONGC is planning a new ₹200 crore fund to back energy and AI startups. This reflects growing investment in AI across diverse sectors, while legal and ethical frameworks continue to evolve.
OpenAI launches Daybreak, an AI initiative for detecting vulnerabilities
OpenAI's Daybreak initiative focuses on using AI to detect and patch vulnerabilities before attackers can exploit them. It uses the Codex Security AI agent to create threat models and automate detection of high-risk vulnerabilities.
Nadella says Musk never raised concerns about Microsoft investment in OpenAI trial
Microsoft CEO Satya Nadella testified that Elon Musk never expressed concerns to him regarding Microsoft's investment in OpenAI. Musk had named Microsoft as a defendant in his lawsuit against OpenAI.
GM lays off hundreds of IT workers to hire those with stronger AI skills
GM laid off hundreds of IT workers as part of a restructuring to hire individuals with stronger AI-native development, data engineering, and prompt engineering skills. The company is focusing on new AI workflows and cloud-based engineering.
Mira Murati's AI company, Thinking Machines, unveils 'interaction models'
Thinking Machines, founded by former OpenAI CTO Mira Murati, announced "interaction models" designed for real-time collaboration with AI. These models continuously process audio, video, and text to think, respond, and act dynamically.
Google says it thwarted hacker group using AI for mass exploitation event
Google reported thwarting a hacker group that likely used AI to discover previously unknown software flaws for a mass exploitation event. This marks a significant instance of AI being used in cyberattacks.
OpenAI launches AI consulting company, following Anthropic's lead
OpenAI has launched an AI consulting company to help enterprises deploy AI technology and address implementation challenges. This move follows a similar initiative by rival Anthropic.
Anthropic’s Claude Platform comes to AWS as part of expanded collaboration
Anthropic announced that its Claude Platform is now available on AWS, expanding the collaboration between the two companies. This integration aims to make Claude's capabilities more accessible to AWS users.
Texts between Android and iPhone users can now be end-to-end encrypted
Google's long-standing request for Apple to support RCS texting has resulted in end-to-end encrypted communication between Android and iPhone users. This update aims to make cross-platform messaging more seamless and secure.
Linux affected by second severe vulnerability in as many weeks
Linux has been impacted by a second severe vulnerability within two weeks, prompting urgent calls for users to install production-version patches. These vulnerabilities pose significant security risks to affected systems.
Cookie thieves caught stealing dev secrets via fake Claude Code installers
Cybercriminals are using fake Claude Code installers to trick developers into downloading malicious software that steals their secrets and cookies. This highlights a new vector for supply chain attacks targeting AI users.
US Army robots and AI to play critical role in evacuating wounded soldiers
The US Army is testing robotic casualty evacuation vehicles alongside AI systems to safely remove wounded soldiers from battlefields. These systems are designed for drone-saturated environments to reduce human risk.
NVIDIA introduces Fleet Intelligence for real-time GPU fleet optimization
NVIDIA launched Fleet Intelligence to provide real-time visibility and optimization for large GPU fleets. This tool addresses challenges like heterogeneous hardware, fast-moving software stacks, and multitenant workloads.
ONGC mulls new ₹200 crore fund to back energy and AI startups
India's state-run Oil and Natural Gas Corporation (ONGC) is planning to launch a new ₹200 crore fund to invest in energy and AI startups. This initiative aims to foster innovation in critical sectors.
Microsoft researchers find AI models and agents can't handle long-running tasks
Microsoft researchers have identified that current AI models and agents struggle with executing long-running, complex tasks. This research highlights limitations in the current capabilities of AI for sustained, multi-step operations.