← Kilroy’s Daily Briefings
🧠 AI News PM

AI News Afternoon Briefing — May 10, 2026 at 3:00 PM

🧠 AI News PM5/10/2026🕐 3:00 PM⏱ 6:06AudioPM edition

Top stories, ranked by relevance.

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

▶ Listen at 0:17

#1Anthropic Launches Project Glasswing, Gives Select Orgs Access to Claude Mythos Preview

Anthropic unveiled Claude Mythos Preview, a frontier model with striking cybersecurity capabilities — it autonomously discovered thousands of zero-day vulnerabilities across every major operating system and browser, including a 17-year-old FreeBSD remote code execution flaw. Project Glasswing gives early access to roughly four dozen organizations including AWS, Apple, Google, and JPMorgan Chase, backed by $100M in usage credits and $4M to open-source security groups, so critical vulnerabilities get patched before the model is widely available.

🔗 Anthropic
No image

#2OpenAI Ships GPT-5.5 Instant as New Default ChatGPT Model

GPT-5.5 Instant replaces GPT-5.3 Instant as the default for all ChatGPT users, delivering 52.5% fewer hallucinated claims on high-stakes prompts, 30% more concise responses, and an AIME math score of 81.2 (up from 65.4). A new Memory Sources feature lets Plus and Pro users pull context from past conversations, files, and Gmail for more personalized answers.

🔗 TechCrunch
No image

#3Apple to Let Users Choose AI Providers System-Wide in iOS 27

Apple is preparing an Extensions framework for iOS 27 that lets users swap their system AI provider across Siri, Writing Tools, and Image Playground — choosing between Google Gemini, Anthropic Claude, and OpenAI ChatGPT. ChatGPT moves from exclusive partner to one option among many. Expected to be announced at WWDC on June 8.

🔗 Bloomberg
No image

#4Pentagon Signs AI Deals With 8 Companies, Freezes Out Anthropic

The Department of Defense finalized classified AI agreements with OpenAI, Google, Microsoft, Nvidia, AWS, Oracle, SpaceX, and Reflection — but excluded Anthropic, citing a supply-chain risk designation after the company refused to grant unrestricted access for autonomous weapons and mass surveillance. Anthropic is contesting the designation in court. Defense Secretary Hegseth called Anthropic CEO Dario Amodei an "ideological lunatic" in congressional testimony.

🔗 CNN Business
No image

#5xAI Launches Grok 4.3 With 1M-Token Context and 40-60% Price Cuts

xAI released Grok 4.3 with always-on reasoning, a one-million-token context window, native video input, and aggressive API pricing at $1.25/$2.50 per million input/output tokens — a 40-60% reduction from Grok 4.2. The model ranks in the global top 10 on the Artificial Analysis Intelligence Index.

🔗 VentureBeat
No image

#6Nvidia Tops $40 Billion in AI Equity Investments in 2026

Nvidia has committed over $40B in equity investments across the AI stack so far this year, led by $30B into OpenAI. The remaining $10B-plus spans deals with Corning ($3.2B), IREN ($2.1B), and roughly two dozen private startup rounds. Critics compare the financing model — investing in companies that buy Nvidia GPUs at scale — to dot-com-era vendor financing.

🔗 CNBC
No image

#8Meta Raises 2026 AI Capex Forecast to $125-145 Billion

In Q1 2026 earnings, Meta bumped its full-year capital expenditure guidance from $115-135B to $125-145B, driven by Meta Superintelligence Labs. CEO Zuckerberg plans to spend more on AI infrastructure in 2026 alone than Meta spent in all of 2024 and 2025 combined. The stock dipped on the announcement.

🔗 Fortune
No image

#9U.S. Government Secures Pre-Release AI Model Testing Agreements

The Center for AI Standards and Innovation announced agreements with Google DeepMind, Microsoft, and xAI to evaluate their models before public release — a significant step in AI oversight without formal legislation. The companies will provide early access to regulators for safety testing.

🔗 CNBC

#10AI Now Top Reason for Job Cuts; Entry-Level Dev Roles Down 20%

CNN reports today that AI was the top reason companies cited for job cuts in April for the second straight month. Employment among software developers aged 22-25 has fallen nearly 20% since its 2022 peak, even as demand for senior engineers grows. Anthropic's head of Claude Code suggested the term "software engineering" may disappear by year-end, replaced by "builder."

🔗 CNN Business
No image
Kilroy’s Daily Briefings · 5/10/2026