← Kilroy’s Daily Briefings
🧠 AI News PM

AI News Afternoon Briefing — Wednesday, May 13, 2026 at 3:00 PM

🧠 AI News PM5/13/2026🕐 3:00 PM⏱ 7:28AudioPM edition

Top stories, ranked by relevance.

Story cards stay below the sticky dock while audio, chapters, date, and brief navigation remain accessible.

▶ Listen at 0:24

#1Google Launches Gemini Intelligence, Makes Android an AI Operating System

At today's Android Show, Google unveiled Gemini Intelligence — a system-level AI layer for Android 17 that lets Gemini move across apps, understand on-screen context, and complete multi-step tasks like booking reservations, building shopping carts, and filling forms. Chrome on Android gets Auto Browse, which can navigate websites and take actions on your behalf. Rollout begins this summer on Samsung Galaxy and Pixel devices, expanding to watches, cars, glasses, and laptops later this year.

🔗 Google Blog
No image

#2Palo Alto Networks Warns AI-Driven Cyberattacks Will Be the "New Norm" Within Months

Palo Alto Networks CTO Lee Klarich says organizations have a "narrow three-to-five-month window" to get ahead of AI-powered exploits before they become routine. The company's own testing found that frontier AI models accomplished the equivalent of a full year of penetration testing in under three weeks, discovering 75 previously unknown flaws. Anthropic's Mythos and OpenAI's GPT-5.5-Cyber are specifically named as models lowering the barrier for attackers.

🔗 CNBC
No image

#3Google Catches First-Ever AI-Generated Zero-Day Exploit in the Wild

Google's Threat Intelligence Group identified and thwarted the first known case of attackers using AI to develop a zero-day exploit — a 2FA bypass in a popular open-source admin tool. Researchers identified the AI origin through hallucinated CVSS scores, educational docstrings, and textbook-style Python formatting. The attackers planned a mass exploitation campaign; Google says its proactive discovery likely prevented it.

🔗 Bloomberg / CNBC
No image

#4OpenAI Launches $4B Deployment Company, Acquires Tomoro

OpenAI launched a standalone business unit called the Deployment Company, backed by $4 billion from TPG, Advent, Bain Capital, Brookfield, SoftBank, Goldman Sachs, and consulting giants McKinsey, Bain, and Capgemini. Its first move: acquiring applied AI firm Tomoro, bringing 150 forward-deployed engineers on day one. The entity is valued at $14 billion and signals OpenAI's pivot from research lab to full-stack enterprise AI company.

🔗 OpenAI / Axios

#5Google and SpaceX in Talks to Put AI Data Centers in Orbit

The Wall Street Journal reports Google and SpaceX are in active discussions to build data centers in space under Google's Project Suncatcher, aiming for a test deployment by 2027. The deal would pair SpaceX's launch capacity with Google's compute needs as AI workloads strain terrestrial infrastructure. This follows Anthropic's deal last week to lease capacity at SpaceX's Memphis data center. SpaceX's upcoming IPO at a $1.75 trillion valuation positions orbital compute as a core revenue pitch.

🔗 TechCrunch / Bloomberg
No image

#6Alibaba Reports 38% AI Cloud Revenue Surge Despite 84% Earnings Drop

Alibaba's Cloud Intelligence Group posted 38% year-over-year revenue growth to 41.6 billion yuan, with AI-related product revenue hitting triple-digit growth for the eleventh consecutive quarter. But adjusted EBITDA cratered 84% as the company pours capital into chips, data centers, and its Qwen model family. CEO Eddie Wu says annualized AI recurring revenue will surpass 30 billion yuan by year-end. U.S.-traded shares jumped 7%.

🔗 CNBC / Washington Post
No image

#7Anthropic Restricts Mythos to Project Glasswing Consortium

Anthropic confirmed it will not make Claude Mythos Preview generally available, instead granting monitored access to over 40 organizations through Project Glasswing, an industry consortium focused on finding vulnerabilities in critical software. The UK's AI Security Institute found Mythos succeeded at expert-level hacking tasks 73% of the time and completed a 32-step simulated corporate network attack in 3 of 10 attempts. Anthropic has not yet offered access to EU regulators.

🔗 Anthropic / The Decoder
No image

#8Musk Requires Wall Street Banks to Buy Grok for SpaceX IPO Access

Bloomberg reports Elon Musk is requiring banks vying to underwrite SpaceX's record-setting IPO to purchase Grok enterprise subscriptions. Morgan Stanley and Apollo Global Management have begun testing Grok internally. Some banks have agreed to spend tens of millions annually on the chatbot. SpaceX confidentially filed its S-1 in April, targeting a June listing at $1.75 trillion.

🔗 Bloomberg
No image

#9NVIDIA Nemotron 3 Nano Omni: Open-Source Multimodal Model Unifies Vision, Audio, and Text

NVIDIA's Nemotron 3 Nano Omni is now available on OCI Enterprise AI — a fully open-source 30B-parameter model that reasons across video, audio, images, and text in a single inference call. It delivers 9x higher throughput than comparable open omni models by collapsing four separate model endpoints into one, directly reducing latency and cost for multimodal agent workflows.

🔗 NVIDIA Blog
No image

#10Apple Camera AirPods for Siri AI Reach Advanced Production Testing

Apple's AirPods with built-in infrared cameras have entered advanced mass-production testing, per Mark Gurman. The low-resolution cameras are designed to give Siri visual awareness of a user's surroundings — not to take photos. Launch was pushed from H1 2026 pending a Siri upgrade now powered by Google Gemini, expected in September. An LED will illuminate when visual data is being captured.

🔗 MacRumors / 9to5Mac
No image
Kilroy’s Daily Briefings · 5/13/2026