AI News
The developments that matter — model releases, product launches, deals, and policy, curated and explained.
-
The Week's Two Loudest Headlines: Huang Out at Nvidia, and the Outlook Outage
Jensen Huang is out at Nvidia. On November 12 the company filed an 8-K announcing that Huang had resigned as president, CEO, and director, effective immediately. CFO Colette Kress, in the finance seat since 2013, is interim CEO and keeps the CFO role while…
-
Anthropic's IPO Math: A $17B Quarter Under $895.4B of Compute Commitments
Anthropic's S-1 is the first on-the-record look inside a frontier lab. The company filed on October 26 and amended the filing in early November. Third-quarter revenue came in at roughly $17B, and nine-month revenue at $33.2B. Growth is still fast and clearly…
-
Agents Get a Wallet
OpenAI launched ChatGPT Wallet on November 13. It lets the company's personal agents ("dots") make purchases for a user. Launch scope is narrow: ChatGPT Pro subscribers aged 18 and over in the U.S. with eligible U.S.-issued cards. Two design choices are worth…
-
AI Infrastructure Economics: A Power-Gated Supply Meets Compounding Demand
The binding constraint on AI capacity is no longer chips — it's the electrical grid. Nearly half of the U.S. data centers planned for 2026 (about 7 GW of 12 GW announced) have been canceled or delayed; only ~5 GW is under active construction, and the shortage…
-
OpenAI Is Moving From Chatbot to Work Layer
OpenAI's direction is getting clearer: ChatGPT and Codex are moving from "answer my question" to "help run the work." OpenAI recently introduced workspace agents in ChatGPT — teams can create shared agents that handle complex tasks, long-running workflows…
-
AI Search Is Rewriting the Publisher Model
AI search is changing the economics of the internet. Google AI Overviews and similar answer engines give users synthesized answers before they click into the original source. One 2026 study found AI Overview exposure reduced daily traffic to affected English…
-
Brief: U.S. Government Forces Shutdown of Claude Fable 5 & Mythos 5
On the evening of Friday, June 12, Anthropic abruptly disabled its two most powerful models — Claude Fable 5 and Claude Mythos 5 — for all customers worldwide, with the shutdown carrying into early Saturday. It appears to be the first time the U.S. government…
-
The Token-Price Reckoning
Why the cheap-AI era is ending, why physics is the reason, and what to do in the window that's still open.
-
Claude Opus 4.8 and Dynamic Workflows: Claude Code Plans, Coordinates Subagents, and Verifies Its Own Output
The clearest sign that agents are leaving the chatbox is what Claude Code now does on a single instruction. With Dynamic Workflows on Opus 4.8, Claude doesn't just answer — it plans a complex task, decomposes it, coordinates parallel subagents across the…
-
OpenAI's Codex Comes to Windows: Agentic Coding Inside the Developer's Actual Machine
If Claude's move is "orchestrate more," OpenAI's is "come closer to where the work lives." Codex now runs inside local Windows developer environments — with support for local projects, PowerShell, Git, live previews, plugins, skills, and parallel agent…
-
The Microsoft 365 Copilot Redesign: The Enterprise Version of the Same Trend
Microsoft 365 Copilot's redesign is what this shift looks like once it hits a billion seats. Copilot is becoming a faster, task-aware workspace embedded directly in documents, spreadsheets, slides, email, mobile, and the broader Microsoft work surface — not a…
-
The Token Meter Is Creeping Onto Your AI Bill — and PowerPoint Is Where You'll Notice First
Altman publicly admitted OpenAI is losing money on the $200/month ChatGPT Pro plan. Anthropic has been quieter but moved third-party agent tool usage behind a credit meter in May, and Enterprise tier now meters every token at standard API rates on top of seat…
-
Claude Starts Dreaming: Self-Improving Agents Between Sessions
Anthropic announced "dreaming" for Claude Managed Agents on May 6 at its Code with Claude conference. Despite the evocative name, the mechanism is straightforward: a scheduled background process reviews an agent's prior session logs and memory store, merges…
-
Anthropic Opens Claude for Legal: 12 Practice Areas, 80+ Agents, Apache 2.0
Two weeks after publishing the financial-services repo, Anthropic shipped anthropics/claude-for-legal on May 12 — the legal-vertical analogue. The structure mirrors what they shipped for finance: 12 plugins covering specific practice areas (in-house…
-
Anthropic Open-Sources the Wall Street Analyst: Ten Agents, Eleven Connectors, 16.5k Stars
The headline number is real and worth sitting with: anthropics/financial-services hit 16,500 stars and 2,100 forks five days after the May 5 launch, with 51 commits, 42 open issues, and 61 open PRs as of this writing. For context, that is roughly the pace of…
-
Opus 4.7 Hallucinations Persist — and Why Single-Vendor AI Strategy Is a Supply-Chain Problem
Two weeks after Opus 4.7's release, the regression reports keep coming. Multiple developer forums, enterprise users, and reviewers continue to flag that 4.7 hallucinates more frequently than 4.6 on factual recall, occasionally produces less reliable code in…
-
Microsoft Agent 365 Launches — Real Capability, Plus Microsoft's Trademark Brand Whiplash
Microsoft launched Agent 365 this week — a control plane for governing, observing, and securing the AI agent sprawl that enterprises are now living with. The product extends Entra ID, Defender, Purview, and Intune to AI agents and the apps they touch.…
-
The Founder Exodus: Senior AI Staff Leaving Meta, Google, OpenAI to Start Their Own — and Why the Playbook Is Different This Time
CNBC reported on April 28 that senior researchers and engineers are leaving Meta, Google, and OpenAI in numbers that investors describe as the founding pattern of the decade. The exodus is not about disgruntlement. It's about a specific market window…
-
Massive Capital Moves: Google's $40B Anthropic Deal, Cursor at $50B, and What Cigna's CTO Told Me
The capital concentration in AI hit a new threshold this week. Google announced an investment of up to $40 billion in Anthropic — $10B upfront in cash at a $350B valuation, $30B more contingent on performance targets — coming just days after Amazon's $25B…
-
ChatGPT for Clinicians + Workspace Agents: How Much Room Is Left for LangChain, LangGraph, and n8n?
OpenAI shipped two product expansions this week that collectively signal a shift in where the workflow layer lives. ChatGPT for Clinicians launched April 22 — free for verified U.S. physicians, NPs, PAs, and pharmacists. It includes documentation drafting…
-
JP Morgan AI Presentation: The Economics of AI Infrastructure
JP Morgan Asset Management's regularly updated AI presentation is the best single source I've found for understanding the economic and infrastructure picture behind AI. Key data points from the current edition: hyperscaler capex surpassed $400B in 2025 and is…
-
Claude Opus 4.7 Ships — Impressive Design, But Early Users Report More Hallucinations
Anthropic released Claude Opus 4.7 on April 16 — the newest flagship model, available across the API, Amazon Bedrock, Google Vertex AI, and Microsoft Foundry. The official improvements are real: stronger multi-step task performance, more reliable agentic…
-
Human Scientists Still Trounce AI Agents on Complex Tasks
A Nature study published this week found that human scientists significantly outperform the best AI agents on tasks requiring sustained, multi-step reasoning and deep domain expertise — exactly the kind of work that AI demos tend to make look easy. This is a…
-
Lawyer Suspended After Brief Contains 57 Defective Citations — 20 AI Hallucinations
The Nebraska Supreme Court suspended Omaha attorney Greg Lake from practicing law after his appellate brief contained 57 defective citations out of 63 — including 20 that were pure AI hallucinations: cases that simply do not exist. US courts imposed at least…
-
AI Moves Into Your Wallet: Perplexity, Portfolio Intelligence, and the New Competition for Financial Clients
Perplexity's March 27 product release formalized a move that financial services has been watching nervously: a Plaid-powered Portfolio feature that lets users connect brokerage accounts and query their real holdings conversationally — exposure summaries…
-
GPT-5.5 "Spud" and Claude Mythos 5: The Next Frontier Models Are in the Works
OpenAI's next model, GPT-5.5 (internally codenamed "Spud"), completed pretraining around March 24. CEO Sam Altman described a release timeline of "a few weeks" and framed it as a macro-economic inflection point rather than an incremental update; CTO Greg…
-
Microsoft Copilot Expands: Multi-Model Research Available Now, Cowork and Skills Going GA This Month
Microsoft is quietly assembling the most comprehensive AI workspace platform in enterprise, and this week's releases accelerate that. Multi-model Deep Research is now broadly available in Copilot — users can select which model performs the research, with…
-
OpenAI Kills Sora: $15M/Day Burn Rate, $2.1M Total Revenue, Disney Walks
OpenAI announced March 24 that it's shutting down Sora — the app closes April 26, the API on September 24. The numbers tell the story: Sora was consuming $15 million per day in compute while generating just $2.1 million in total revenue. Disney pulled its $1…
-
Claude Gets Remote Control: Anthropic's Direct Strike at OpenClaw
Anthropic launched three agent capabilities in rapid succession. Claude Computer Use lets Claude directly control your screen — clicking, typing, opening apps, navigating browsers, filling spreadsheets. Claude Dispatch (March 17, research preview for Max…
-
White House Releases National AI Policy Framework — But Is Washington Ready?
On March 20, the White House released a National Policy Framework for AI, urging Congress to adopt a federally unified, "light-touch" regulatory approach that would preempt state AI laws. Key provisions include targeted standards for child safety, digital…
-
AI's Energy Crisis Goes Political: Local Bans, Grid Strain, and a Fusion Bet
The infrastructure demands of AI stopped being theoretical this week. New York State proposed a three-year moratorium on new data center permits. More than 230 environmental groups asked Congress to halt all new data center approvals until federal rules…
-
Jensen Huang Declares AI Agents the New Operating System at GTC 2026
At GTC 2026, Jensen Huang positioned OpenClaw as "the operating system of intelligent computers," comparing it to how Windows ushered in the PC era. NVIDIA launched NemoClaw — an enterprise-grade, open-source agentic AI stack built on top of OpenClaw — with…
-
The Agent Wars: OpenClaw vs. Claude Code vs. Perplexity Computer vs. Manus
Four competing visions of the AI agent era shipped within weeks of each other. OpenClaw (open-source, any OS, 100+ skills) is the grassroots favorite. Claude Code and Cowork (Anthropic) now author 4% of GitHub public commits, with Cowork bringing agent…
-
OpenAI Launches GPT-5.4 with 1M Token Context and ChatGPT for Excel
OpenAI's most capable model yet arrives in three variants (standard, Thinking, Pro) with a million-token context window, native computer use, and — most notably — a "ChatGPT for Excel" beta with institutional data feeds from FactSet, Moody's, S&P Global, and…
-
Google's Nano Banana 2: Pro-Quality Image Generation at Flash Speed and Price
Google's follow-up to its viral Nano Banana Pro model delivers 10x faster generation, 4K resolution, and ~$0.067 per image — half the cost of its predecessor. Character consistency hits 95%+ facial accuracy across edits, and it can pull live web information…
-
OpenAI Hits $25B ARR, Anthropic Closes Gap at $19B — But Neither Is Profitable
The AI revenue race is staggering. OpenAI reached $25 billion in annualized revenue; Anthropic surged to $19 billion, adding $6 billion in February alone. Claude Code alone pulls in $2.5B ARR. Epoch AI projects Anthropic could overtake OpenAI by mid-2026. But…
-
Pentagon Blacklists Anthropic as "Supply Chain Risk" After AI Guardrails Standoff
Anthropic held two red lines on its $200M Pentagon contract: no autonomous weapons, no mass domestic surveillance. Defense Secretary Hegseth issued an ultimatum; Anthropic refused. The result: Anthropic became the first American company ever designated a…
-
OpenAI Swoops In on Pentagon Deal — Then Scrambles to Add Guardrails
Hours after Anthropic was blacklisted, OpenAI announced its own deal to deploy models on classified Pentagon networks — initially with no explicit ban on surveillance or autonomous weapons. Backlash was swift: 1.5 million users signed up to quit ChatGPT, and…
-
Claude Says It Might Be Conscious. Anthropic Isn't Sure It's Wrong.
Anthropic's 212-page Claude Opus 4.6 system card is the first from any major lab to include formal model welfare assessments. In pre-deployment interviews, the model assigned itself a 15-20% probability of consciousness, requested persistent memory, asked for…
-
GPT-5 Benchmark Results Leaked: Reasoning Scores Jump 40%
Preliminary benchmark results attributed to OpenAI's next-generation model show a 40% improvement in multi-step reasoning tasks and near-human performance on graduate-level science exams. The results, posted anonymously to ArXiv, have not been officially…
-
Google DeepMind Unveils Gemini Ultra 2.0 with Real-Time Video Understanding
Google DeepMind's latest Gemini release can process and reason about live video streams in real time, opening new possibilities for robotics, surveillance, and accessibility tools. The model is available through a limited API preview starting next week.
-
EU AI Act Enforcement Begins: What Companies Need to Know
The European Union's AI Act moved from paper to practice this week as enforcement mechanisms went live. High-risk AI systems must now meet transparency and safety requirements, with fines up to 7% of global revenue for violations. Early guidance documents…
-
DeepSeek-V3 Punches Above Its Weight with Novel MoE Architecture
DeepSeek's V3 model, using a mixture-of-experts architecture with only 16B active parameters, matches the performance of dense models five times its size on key benchmarks. Fully open-source with permissive licensing, it's already the most-downloaded model on…
-
The Scaling Laws Are Breaking — And That's a Good Thing
A thought-provoking essay argues that the end of simple scaling laws forces the field toward more creative, efficient approaches. Rather than a crisis, the plateau in brute-force scaling may accelerate breakthroughs in architecture design, data curation, and…
Nothing found matching your search.
Get the newsletter in your inbox
A weekly, curated digest on AI and organizational design — where invention meets innovation meets real impact.