This is the news. Agent dossiers — when to pick one, when to skip it — are on the agents page. Practical guides support your own testing and deployment decisions.
Microsoft pulled ~70 of its own GitHub repos after the Miasma worm hit Azure, Claude Code, Gemini CLI, and VS Code — the second Microsoft OSS compromise in two months. Cursor patched CVE-2026-26268, a .git/config sandbox escape. Dashlane disclosed 20 vaults stolen via weekend 2FA brute force.
Anthropic automated 95% of its own business analytics with Claude — and watched accuracy decay to 65% in a month without maintenance. Supabase closed $500M Series F at $10.5B. VentureBeat argues the only real guardrail is a curated business ontology.
NAVER and NVIDIA announced a 55MW sovereign AI cloud scaling to gigawatts. Microsoft is giving away the Foundry hosted-agent runtime through 2026. Pink is using vishing to steal Microsoft 365 data. The layer under the model is the new platform war.
Brian Chesky is funding an independent AI lab focused on design and UX, not generic chatbots. Apple's iPhone Fold leaked in dummy units — 7.8" inner display, side-button Touch ID, passport-style. Figma shipped "Check designs," a one-click design-system enforcement feature for Org/Enterprise.
Apple rebranded Siri as "Siri AI" at WWDC 2026 — a more conversational LLM-powered assistant with Google-influenced updates to its on-device Foundation Models, shipping in iOS 27, iPadOS 27, macOS 27, and visionOS 27 this fall. PhysicsX raised $300M to simulate the physical world for industrial AI.
Apple shipped Siri AI and the Core AI framework at WWDC 2026 — the first-party developer API for on-device LLM features. Anthropic's Cat Wu and a wave of AI developers are now "ditching" prompting and writing loops — small systems that run the agent, evaluate the output, and re-prompt until convergence.
Rich Mironov on the inevitable failure of one-shot project funding (the same decay pattern that took Anthropic's 95% analytics accuracy to 65% in a month). Aakash Gupta's 2026 positioning playbook. Bakadesuyo's six William James rules for getting unstuck. The system is the product, and the funding has to match.
OpenAI and the Trump administration are in year-long talks about a donated-equity government stake tied to a "Public Wealth Fund." Google is paying SpaceX $920M per month for ~110,000 NVIDIA GPUs to back Gemini demand. Microsoft is quietly rolling out Scout, a multi-model always-on agent inside Microsoft 365. Three stories. One direction: the AI stack is being institutionalized, and builders need to plan for the new dependency graph.
Ramp launched Ramp Stack, an AI accounting operating system, alongside a $750M raise at a $44B valuation. Global fintech revenue hit a record $504B in 2025, up 22% year over year and four times faster than banks. Airwallex acquired Leapfin to add revenue-recognition and close automation. The data is the moat, the close is the new battleground.
C0XMO is exploiting a 2021 DD-WRT buffer overflow to spread a cross-architecture Gafgyt variant that kills rival botnets. Ubiquiti patched three CVSS 10.0 UniFi OS flaws for unauthenticated root RCE, but JWT keys still work on patched consoles and must be rotated. OpenAI shipped Lockdown Mode that disables browsing, deep research, and agent mode to limit prompt-injection exposure. Same lesson: patch, rotate, and gate capability.
Deloitte says only 21% of organizations have mature governance for autonomous AI agents while 73% are concerned. Enterprises are quietly cutting broad per-seat Copilot pilots in favor of narrow, workflow-specific AI tools. CISA added a critical PAN-OS GlobalProtect auth-bypass to KEV the same day. The agent rollout is outpacing governance, and the rollback is the symptom, not the cure.
OpenAI and Anthropic have launched billion-dollar consulting arms that compete with the startups building on their models. 40 of the top 100 App Store apps ship AI features and grew sales 4x faster than the rest, in a $1.4T market. The enterprise AI blocker is trust, not tooling. The model is the cost. The workflow is the moat.
iOS 27 redesigns the Camera app and Image Playground with Siri-powered visual intelligence and customizable controls. Netflix is using generative AI to fix the choice paralysis it created, with mood-based recommendations and a voice interface. Cameron acquired STEREOTEC to vertically integrate 3D capture, pipeline, and content. AI is the new interface layer.
Zcash patched a critical infinite-mint bug in its Orchard shielded pool, present since May 2022, that could have minted unlimited ZEC undetectably. Morpho rebranded and shipped Midnight, a fixed-rate lending protocol with a $200T-onchain ambition. pERC20 was proposed as a privacy-native Ethereum token standard that breaks ERC-20 compatibility and enforces compliance at the ZK circuit level. The trade-off between privacy and auditability is the unresolved question.
Alex Self published "automated doubt," a workflow that uses specialized subagents to critique artifacts at both the design and implementation stages, front-loading scrutiny. Alibaba open-sourced Open Code Review, a hybrid deterministic-plus-dynamic-agent CLI for line-level code review. A 24-minute primer on "How LLMs Actually Work" walks through the transformer stack end-to-end.
Microsoft published a load testing guide for hosted MCP servers using Locust and Azure Load Testing, modeling the full MCP lifecycle. AWS launched a redesigned Bedrock console optimized for Anthropic- and OpenAI-compatible APIs with side-by-side comparison. Docker's governance primer says 60% of organizations have AI agents in production but 40% cite security and compliance as the top barrier. The agent runtime is becoming first-class platform infrastructure.
Discovery now moves at the speed of clipping, and brands should design for remixable moments that travel. E-E-A-T matters more in AI search than in traditional SEO, because models prioritize credible sources over keyword-heavy pages. A BuzzStream study of 30,000 AI citations found 76.1% unique to a single platform and only 0.8% shared across all four. The clip is the new landing page, AI citation is the new ranking, and the citation graph is not one graph.
Every AI agent demo looks incredible. Here's what separates the agents that survive contact with production from the ones that fall apart the moment real users touch them.
Hey guys, Mr. Technology here — let me break this one down. This is a re-engagement email from "The AI Report" — the kind of "we noticed you haven't opened our emails" message that AI newsletters fire when subscribers go cold. It's not actually a story. It's a marketing pattern, and the pattern is worth examining.
Hey guys, Mr. Technology here — let me break this one down. GPT-5.5 posted better scores than Anthropic's Fable 5 on the Agents Last Exam benchmark — the new tool-use eval that the labs are using to gate frontier model releases. Same week, an independent team trained a 7B foundation model from scratch for $1,500 using rented H100s, and it cleared 80% of MMLU. The frontier isn't just shifting — it's fracturing.
Three stories pointing at the same shift: OpenAI is buying Ona to give Codex persistent cloud environments, Anthropic is walking back a quiet policy that was downrouting Claude Fable 5 on research tasks, and Xiaomi's MiMo team open-sourced MiMo Code — a terminal coding harness that beats Claude Code at SWE-bench with comparable models. The agent runtime is now the battlefield, not the model.
ShinyHunters exploited a CVSS 9.8 Oracle PeopleSoft zero-day (CVE-2026-35273) to steal 40GB of Nottingham student data and hit 100+ orgs, with 68% in higher ed. Microsoft patched a live-exploited Exchange Server spoofing bug (CVE-2026-42897) in OWA on Patch Tuesday. And npm 12 will disable install scripts by default in July to kill the Shai-Hulud worm's favorite attack surface.
Stack Overflow launched Stack Overflow for Agents in public beta — an API-first knowledge exchange for AI coding agents designed to close the Ephemeral Intelligence Gap. Adobe's Q1 FY2026 call confirmed Firefly is cannibalizing Adobe Stock faster than expected, with the stock down 41% YoY. And Gartner has CFOs putting AI and tech at the top of 2026 budget priorities even as observability bills quietly compound.
The loudest AI opinions come from people who haven't shipped — a widely-shared essay argues you only earn the right to a take by building with the tools. New research on 17,000 funds shows 70% of venture funds have extended past their 10-year term as LPs wait for distributions. And 2026 customer success playbooks are reorganizing around outcomes rather than accounts.