← Kilroy’s Daily Briefings
🧠 AI News PM

AI News Afternoon Briefing — Thursday, May 21, 2026 at 3:00 PM

🧠 AI News PM5/21/2026🕐 3:00 PM⏱ 6:51AudioPM edition

Top stories, ranked by relevance.

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

▶ Listen at 0:19

#1Google Launches Gemini 3.5 Flash — Frontier Intelligence at Flash Speed

Google's marquee I/O 2026 model drop is now generally available. Gemini 3.5 Flash outperforms the much larger Gemini 3.1 Pro across nearly all benchmarks while running four times faster, and a further-optimized variant inside Antigravity 2.0 clocks twelve times faster at equivalent quality. Google also debuted Managed Agents in the Gemini API, letting developers spin up a reasoning, tool-using agent in a single call. Internal Antigravity token volume has surged from 500 billion to over 3 trillion per day since March.

🔗 Google AI Blog

#2Google Overhauls Search With Always-On AI Agents

Google is entering what it calls "the era of Search agents." A new seamless AI Search experience — blending traditional results, AI Overviews, and conversational AI Mode — is live worldwide today. Coming this summer: "information agents" that run 24/7 in the background, reasoning across the web to surface what matters to you, replacing the aging Google Alerts paradigm. Generative UI powered by Antigravity will build custom interactive responses on the fly, free for all users.

🔗 Google Blog

#3Trump Delays AI Executive Order on Pre-Release Model Reviews

Hours before a planned signing, the White House postponed an executive order that would have created a voluntary 90-day government review of frontier AI models before public release. Trump cited concerns about overregulation and falling behind China. The framework had been negotiated with OpenAI, Anthropic, and other major labs, and follows last month's Mythos cybersecurity scare that pushed pre-deployment testing into the spotlight.

🔗 Washington Post
No image

#4Hark Raises $700M Series A at $6B Valuation

Brett Adcock's stealth AI lab Hark closed one of the largest Series A rounds ever — $700 million at a $6 billion post-money valuation — for a "universal AI interface" combining custom foundation models with dedicated hardware. Backed by AMD Ventures, ARK Invest, Qualcomm Ventures, Salesforce Ventures, and others, the 70-person team plans to ship its first AI models this summer with AI-native hardware to follow.

🔗 TechCrunch
No image

#5Meta Cuts 8,000 Jobs, Redirects Up to $135B to AI Infrastructure

Meta began executing the first wave of its planned 10% workforce reduction, cutting 8,000 roles and canceling 6,000 open requisitions. The restructuring funnels an estimated $7–8 billion in annual savings toward AI infrastructure, with 2026 capex guided between $115B and $135B — a 73% jump over 2025. The cuts hit recruiting, content moderation, and non-AI product teams hardest, reorganizing survivors into AI-focused "pods." More rounds expected in August and fall.

🔗 TechRepublic
No image

#6Spotify and Universal Music Strike Landmark AI Covers Deal

Announced at Spotify's Investor Day, the first-of-its-kind licensing agreement lets Premium subscribers use generative AI to create covers and remixes of Universal Music artists' songs. Artists opt in and share revenue. The deal puts Spotify in direct competition with AI music startups like Suno and Udio while establishing a legitimate royalty framework for AI-generated derivative works. Pricing and launch date TBA.

🔗 TechCrunch
No image

#7Musk Loses OpenAI Lawsuit — Jury Rejects All Claims in Under Two Hours

A unanimous nine-member federal jury dismissed every claim in Elon Musk's $134 billion lawsuit against OpenAI and Sam Altman, finding all were barred by the statute of limitations. Deliberations lasted roughly 90 minutes after a three-week trial. Musk called it "a calendar technicality" and vowed to appeal to the Ninth Circuit, but for now OpenAI's for-profit conversion path faces one fewer obstacle.

🔗 NPR
No image

#8Pope Leo XIV to Present First-Ever AI Encyclical Alongside Anthropic Co-Founder

Pope Leo XIV's first encyclical, "Magnifica Humanitas," will be presented May 25 at the Vatican's Synod Hall with Anthropic co-founder Christopher Olah among the presenters. The document addresses AI's ethical challenges and the protection of human dignity. Leo has likened AI's rise to industrialization — "one of the defining moral tests of the modern age." VP Vance said he is "looking forward to reading" it.

🔗 National Catholic Reporter
No image

#9EU Bans AI Nudifier Apps, Delays High-Risk AI Act Rules

The European Parliament and Council reached a deal banning AI "nudification" apps — tools that generate non-consensual explicit imagery — with fines up to 35 million euros or 7% of global turnover, effective December 2026. The move was partly catalyzed by AI-generated images of Italian PM Meloni. In the same omnibus, however, the EU delayed high-risk AI compliance rules to December 2027 and August 2028, giving industry more breathing room.

🔗 Euronews
No image

#10Standard Chartered to Cut 7,000 Jobs to Fund AI Transformation

StanChart CEO Bill Winters announced plans to eliminate over 7,000 roles over four years, cutting 15% of corporate functions by 2030, explicitly calling it "replacing lower-value human capital" with AI investment. The remark triggered swift backlash — including from a former head of state — forcing Winters into damage control. Back offices in Chennai, Bengaluru, Kuala Lumpur, and Warsaw bear the brunt.

🔗 Bloomberg
No image
Kilroy’s Daily Briefings · 5/21/2026