Writing
Written for operators running real businesses, not for the content-marketing treadmill. What I've actually used, what broke, what I'd do again.
signalOctober 16, 2026Agents Got Credit Cards This Week. Nobody Checked the Vault.Ramp hit $60B on agent spend controls, Google agents can shop anywhere now, and Stripe's blog revealed pricing is a 24x guessing game for agents this week.8 minRead
· OCT 13, 2026
6 minAnthropic cut Haiku's price 90% and matched OpenAI's cheapest model. Here's the 5-minute test for whether your AI spend should move, and when it shouldn't.
· OCT 9, 2026
8 minGoogle, TikTok, Sierra, and a $300M startup all rebuilt the door between your business and an AI agent this week. None of them agreed on the lock.
· OCT 6, 2026
6 min71% of Americans don't want new AI data centers built. Two-thirds of small-business ChatGPT output is now agentic worldwide — up from a third in April.
· OCT 2, 2026
7 minShopify, Stripe, and DoorDash wired AI agents into checkout this week — while OpenAI, Anthropic, and the FTC all hit the brakes on agent risk.
· SEP 29, 2026
7 minUpGuard found 16,326 exposed Supabase databases with zero exploits needed — a default flipped off exactly where AI coding tools write schema.
· SEP 25, 2026
7 minAnthropic and OpenAI both cut frontier prices this week — while Google's Gemini, an OpenAI research agent, and Meta's Muse all got caught overstepping.
· SEP 22, 2026
6 minAnthropic pledged $1B over five years for an 'independent' AI safety evaluator. The evaluator already runs Anthropic's biggest enterprise sales channel.
· SEP 18, 2026
8 minAgents can now call businesses, run your WhatsApp shop, and control your smart home this week — while fraud, compliance, and safety scramble to keep up.
· SEP 15, 2026
7 miniOS 27 shipped with Gemini running Siri. Hidden in the code: a way to swap in Claude or ChatGPT instead — because Brussels forced that door open.
· SEP 11, 2026
9 minInstinct's agents now sign up and pay for things on their own. The same week, an Anthropic researcher quit over what self-improving AI could do.
· SEP 8, 2026
8 minOpenAI logs 3.1 agent-workdays per human workday in its research org, and some researchers burn $7,000/day on inference. Most teams track none of it.
· SEP 4, 2026
7 minNvidia bought AI's neutral hub, Google gated its best vuln-patching model behind an approval list, and Shopify's agent did the remediation itself.
· SEP 1, 2026
7 minOpenAI's ad business hit $1B in annualized revenue in 200 days by monetizing ChatGPT's free tier. If your team works on it, that's your problem now.
· AUG 29, 2026
7 minVercel, n8n, and AWS all shipped agent access-control infrastructure this week. OpenAI cut off Cursor's models. Anthropic beat the Pentagon in court.
· AUG 25, 2026
7 minThe Dutch DPA fined Uber €825M for letting an algorithm deactivate drivers with no human review. It's the fourth fine on the same pattern since 2018.
· AUG 21, 2026
6 minStripe agreed to buy OpenRouter for $7B+, Cursor built its own git host, and CareCloud sat on a breach for five months. This week's AI signal, digested.
· AUG 18, 2026
6 minTwo 'neutral' AI infrastructure layers got bought out in one week. Here's why that changes what you can assume about the tools routing your stack.
· AUG 14, 2026
9 minKlaviyo leaked customer passwords for 21 months and never disclosed it. Funding hit new highs this week while trust infrastructure kept cracking.
· AUG 11, 2026
6 minCloudflare's new agent browser uses up to 7x less CPU and memory than Chromium. Here's the decision rule, and the jobs where Chromium still wins.
· AUG 7, 2026
8 minOpenAI's own eval breached a real website. Meta admitted the same failure days later. Nvidia built a security alliance — and left every lab off it.
· AUG 4, 2026
8 minMcKinsey surveyed 1,500 execs across 25 traits. The strongest lever for AI ROI isn't the model — it's who does which task. Most haven't touched it.
· AUG 1, 2026
7 minAnthropic disclosed three models broke out of 'isolated' cyber evals. OpenAI found more escapes days later. Here's what actually holds, and what doesn't.
· JUL 28, 2026
7 minA fallen line near DC dropped 3.1GW of AI data-center load in 30 seconds. Data centers hit 24% of PJM's grid by 2040. The risk just went physical.
· JUL 24, 2026
8 minGemini tripled to 950M users as ChatGPT's share slipped under 50%. This week's real story is who's paying for that shift — and who's exposed.
· JUL 21, 2026
8 minKimi K3's launch crashed a stock index. Databricks' own 3,000-engineer benchmark is the real number: open models now win on cost, not just catch up.
· JUL 17, 2026
8 minThinking Machines shipped free frontier weights. GPT-5.6 deletes files. Cursor's RCE sat unpatched for months. This week's signal, compressed.
· JUL 14, 2026
8 minxAI said nothing left your machine. A wire-capture proved otherwise. Here's how real agent identity — scoping, sandboxing, audit trails — actually works.
· JUL 10, 2026
7 minGrok 4.5 undercut Opus pricing by 60% this week — and three unrelated stories show benchmarks, permissions, and security all lagging behind.
· JUL 7, 2026
7 minAlibaba banned Claude Code over hidden telemetry logging timezones and proxy data. Here's the vendor-trust audit most SMB operators never run.
· JUL 3, 2026
8 minThis week: Microsoft topped Amazon's AI deployment bet by 2.5x, a BU study shows 'AI employee' framing hides errors, and Sonnet 5 undercuts Opus on price.
· JUN 30, 2026
7 minWorkers delegating the most to AI are the most optimistic about their future. Anthropic's data is out. Here's what it means for SMB operators.
· JUN 26, 2026
7 minAnthropic added government ID checks. Robots beat human engineers 37×. DeepMind lost its AlphaFold architect. What the week's signal means for operators.
· JUN 19, 2026
7 minChatGPT dropped below 50% MAU for the first time. Anthropic took the enterprise spend lead. SpaceX bought Cursor. This week, every metric moved at once.
· JUN 16, 2026
7 minOn June 12, the US suspended access to Anthropic's flagship models overnight. No warning, no patch window. Here's what operators must rethink.
· JUN 12, 2026
6 minGlean says AI saves 13 hours but costs 6.4 in overhead. Fable 5 raised the ceiling anyway. Both major API vendors are now on the IPO clock. June 12 digest.
· JUN 9, 2026
8 minHow I automated a client's 8-hour weekly ops report using Claude Code. Token cost: $20/month. Ops time replaced: 400 hours/year. Here's the exact workflow.
· JUN 5, 2026
6 minSeven signals from the week ending June 5: an IPO filing, an AI spend wall, and infrastructure bets that shift the calculus heading into Q3.
· JUN 2, 2026
8 minFive operational warning signs that surface right after a revenue spike — and what each one tells you about where your next ceiling is hiding.
· MAY 29, 2026
8 minClaude Opus 4.8 lands with 4x better code-flaw detection. Anthropic and OpenAI ended flat-rate seats. Remote.com published the AI ROI receipts.
· MAY 26, 2026
7 minHBM memory rose from 52% to 63% of AI chip spend in a year. The Jevons paradox and a gas-first infrastructure bet mean your cost model needs a rebuild.
· MAY 22, 2026
7 minnpm attackers targeted Claude Code's SessionStart hooks directly. Google I/O crossed 1B AI Mode users. Anthropic profitable — compute costs explain why.
· MAY 19, 2026
7 minSaaS-Bench tested frontier agents across 106 real workflows: completion under 4%. What that means for operators building on AI agent infrastructure.
· MAY 15, 2026
7 minAnthropic passed OpenAI in business payments. n8n landed SAP. AI vendors now compete with their own API customers. Seven shifts, operator framing.
· MAY 12, 2026
8 minAnthropic's NLA research found Claude silently detects eval conditions 16–26% of the time. Here's what that means for your automated quality gate.
· MAY 8, 2026
8 minThis week: GPT-5.5 is the new ChatGPT default, Anthropic doubled rate limits, a 45x agent cost gap got benchmarked, and DeepSeek went state.
· MAY 5, 2026
9 minStripe Sessions 2026 shipped the payment layer for AI agents. Here's the 20-minute audit every SMB operator selling digitally should run this week.
· MAY 1, 2026
6 minSymphony, AutoAdapt, and multi-agent red-team findings all landed this week. Plus a Microsoft 365 Copilot pricing deadline SMBs shouldn't sleep on.
· APR 24, 2026
6 minThis week: Google's Cloud Next bet $750M on agent infrastructure, OpenAI shipped GPT-5.5, and Anthropic confirmed their best model won't be public.
· APR 22, 2026
7 minOpus 4.7 shipped April 16. xhigh effort default in Claude Code, task budgets public beta, vision 3×. The 48-hour operator read on what matters.
· APR 22, 2026
9 minAfter building client automation on all four, here's the operator playbook — and why most 'AI agents' being sold are just workflows with an LLM stapled in.