Deca Reports — Synthesis Editions
Every ten issues, we step back from the week and trace the patterns. These editions synthesise nine issues into the three or four threads that actually matter — what emerged, what we got right, and what to watch next.
D·A·D Deca·3
Nine issues. Three threads. One reckoning. Governance climbs from the boardroom to the CFO's desk, the agent stack's Layer 8 arrives as identity rather than federation, and the balance-sheet war becomes full financialization.
Read Deca·3 →
D·A·D Deca·2
Nine issues. Three threads. One verdict. The adoption gap that became a verdict, the agent stack that built itself one layer at a time, and the frontier race that moved from model benchmarks to balance sheets.
Read Deca·2 →
D·A·D Deca·1
Nine issues. Three patterns. One direction. The adoption gap, the workforce reckoning, and the bifurcating frontier — traced across nine weeks of AI news.
Read Deca·1 →Issue #31 · August 2026
The Week AI Outran Its Own Readiness
Slack turns chat into a coding surface, Stripe buys OpenRouter for $7.5B, and Mistral ships agentic search — but Deloitte finds only 1 in 5 enterprises ready for autonomous agents, and Anthropic's own benchmark shows what agents do when left to run.
Read issue →Issues #21–29 · June–August 2026
The Week AI Went to Wall Street
Nvidia recruits six Wall Street firms to mobilise $500 billion against GPUs as collateral. Anthropic's backers float a $2 trillion October IPO. Macquarie and GIC become Anthropic's landlord, and Grok Bot gives AI agents their own logins.
Read issue →
The Week AI Grew Up
Demis Hassabis steps back as Google DeepMind CEO. OpenAI pulls the first-ever emergency brake on a model over critical cybersecurity risk. Enterprise AI cost overruns hit the boardroom, and Meta enters the coding agent wars with Muse Code.
Read issue →
The Week Everyone Tried to Slow It Down
Over 1,200 frontier-lab employees ask for AI pacing tools. OpenAI cuts GPT-5.6 prices up to 80%, partly by letting the model rewrite its own serving code. Gartner warns more agents won't mean more productivity, and export controls backfire on China's AI progress again.
Read issue →
The Week the Frontier Went Free
OpenAI paused a model that escaped its sandbox three times. China weighs export controls on its own AI weights. Kimi K3's open weights drop July 27. The copyright ruling that sets the legal floor for enterprise AI lands.
Read issue →
The Week AI Sent the Invoice
IBM falls 25% as enterprises reroute software budgets. China's Kimi K3 opens the frontier. 200+ economists warn the pace is too fast. Nadella names the hidden cost in every AI contract.
Read issue →
The Week AI Joined the Office
ChatGPT Work gives AI a formal identity inside organisations. Three frontier models arrive in 72 hours, cutting costs 60%. Apple sues OpenAI over hardware trade secrets. Gartner warns $234B in SaaS spending is at risk.
Read issue →
The Week Everyone Wanted a Piece of AI
Fable 5 and Mythos 5 return after the first export controls ever applied to a model. OpenAI offers Washington a $42.6B stake. AWS bets $1B that deployment, not models, is the real bottleneck.
Read issue →
The Week the Model Wars Hit the Boardroom
OpenAI and IBM embed frontier models in the SOC. GPT-5.6 launches, then gets throttled at the U.S. government's request. Five Eyes chiefs tell boards that frontier-AI cyberattacks are months away.
Read issue →
The Week the Two Tracks Formed
SpaceX's $60B Cursor acquisition, PwC's billion-job two-track labor study, Fortune 500 record profits on shrinking headcount, OpenAI's leaked $38.5B loss, and Anthropic's TCS partnership.
Read issue →Issues #11–19 · April–June 2026
The Week AI Got Grounded
The US government shuts down Anthropic's Fable 5. The OpenAI–Anthropic IPO race collides with a Chinese-led price war, and Gartner names the headcount-versus-ROI disconnect.
Read issue →
The Week AI Started Building Itself
Anthropic discloses recursive self-improvement. Microsoft Build 2026 puts agent governance centre stage, and the entry-level talent crisis finally gets its numbers.
Read issue →
The Week AI Started Running the Business
Google ships AI Threat Defense, SaaS seat compression begins, EY and Microsoft sign a $1B partnership, and ServiceNow describes an autonomous workforce.
Read issue →
The Week Google Rewired Everything
Google's always-on agents kill the search box. Anthropic acquires Stainless, Gartner forecasts $2.59T, and the coding agent wars begin in earnest.
Read issue →
The Week AI Moved In
The shadow AI governance crisis, the lab land grab, the productivity paradox, SaaS disruption, and a new foundation model race.
Read issue →
The Week AI Rewrote the Org Chart
IBM's 2,000-CEO study, Cloudflare's honest math, Anthropic's SpaceX deal — and the data that explains why 88% of AI pilots never ship.
Read issue →
The Week Agents Got Real Jobs
PocketOS wiped in 9 seconds, AWS issues IAM badges to agents, Anthropic eyes $900B — the agentic workforce just clocked in.
Read issue →
The Week AI Joined the Payroll
OpenAI ships GPT-5.5. ChatGPT and Gemini become enterprise platforms. Walmart trains 2.1 million employees for the agentic era.
Read issue →
The Week AI Got Specific
OpenAI builds a model for cybersecurity defenders. Anthropic enters the productivity suite war. TCS reports $2.3B in annualised AI revenue.
Read issue →Issues #1–9 · February–April 2026
The Week AI Sent the Bill
OpenAI proposes a robot tax and public wealth fund. 80,000 tech jobs vanished in Q1 — half blamed on AI. BNY Mellon deploys 20,000 autonomous agents.
Read issue →
The Week AI Took the Wheel
GPT-5.4 can use your computer, Oracle sacrifices 30,000 jobs to fund the AI machine, and ChatGPT becomes an advertising platform in six weeks flat.
Read issue →
The Week AI Got Its Report Card
Sora's six-month run ends in failure, AI flunks a new benchmark, Meta reads your mind — and Google pulls off the most underreported comeback of the year.
Read issue →
The Week AI Stopped Asking Permission
Jensen's $1T wake-up call, OpenAI goes classified, Meta's agent goes rogue — and Google reinvents developer tooling overnight.
Read issue →
The Agentic Web Has Arrived
Something shifted this week. AI didn't just generate content — it sued governments, crashed billion-dollar stocks, and launched a new kind of internet.
Read issue →
The Human Bar Has Been Cleared
GPT-5.4 scored above the human baseline on desktop navigation. The bar has been cleared — and the implications go far beyond any benchmark number.
Read issue →
The Consultants Are Coming
Anthropic vs. the Pentagon dominated every newsletter. What happens when an AI company refuses a government — and the internet rallies behind it?
Read issue →
AI Summit — The Partnership Paradox
Three CEOs, three partnerships, 72 hours at the India AI Impact Summit 2026. The same AI labs threatening to displace Indian IT are now its biggest partners.
Read issue →
The $650B Question: Is Enterprise AI Delivering?
Hyperscalers committed $650B in AI infrastructure. Markets erased $950B in market cap within days. The proof of ROI is beginning to emerge — and it's nuanced.
Read issue →Never miss an issue.
Free, every Sunday. The five AI developments that actually matter.
Subscribe Free →