The $20 AI subscription isn't really $20
Is your $20 AI subscription sustainable? Chris Cooley explores the hidden costs of LLMs, investor subsidies, and why agentic AI is driving up the true price of compute.
Automating Doubt: The Danger of Outsourcing AI Skepticism
The cheapest thing to automate is your own skepticism. It's also the only part of the job you can't afford to give away.
Why I Started This
I've spent 19 years in the military. Over a decade in cybersecurity. I've sat in intelligence operations centers where the analysis I produced had real consequences for real people downrange. I've briefed generals. I handed the FBI analysis that helped send someone to prison. I've co-authored a peer-reviewed paper in a medical journal — on pandemic intelligence, of all things. I also make music. I shoot landscapes. I climb rocks and ride mountains when the weekends allow. And somehow, none of that prepared me to answer the question: What is your personal brand?
Is Anthropic's Opus 4.8 Exempt from their Cyber Verification Program?
A field note on usage-policy false positives, a model that seems to have gotten jumpier as it got smarter, and the difference between describing a threat and building one.
Agentic AI Is the New SOAR: Everyone's Excited, Most Will Be Disappointed
I keep having conversations with folks that start something like *"We're thinking about dumping our Splunk logs into an AI agent and having it find the bad stuff."* Every time, I nod. I understand the appeal. I also know how it's going to end. This is the SOAR moment for AI agents. And if you lived through the first wave of SOAR adoption (or perhaps still are), you already know what's coming.
I Love Lovable. But It Has a Real Problem.
Lovable is a game-changer for rapid prototyping, but it has a major blind spot: SEO and link previews. Explore why client-side rendering might be sabotaging your content.
Why AI Isn’t Replacing Intelligence Analysts Any Time Soon
Despite the hype, AI models continue to struggle with hallucinations and accuracy in intelligence analysis. Explore why professional analysts remain indispensable.
The Court Just Said AI Can't Own Art.
On March 2, 2026, the U.S. Supreme Court declined to hear Thaler v. Perlmutter. No dramatic ruling. No landmark opinion. Just a denial of certiorari. Letting the lower court decisions stand without a word. The result: AI-generated art cannot be copyrighted. Not by the AI. Not by the person who owns it. Not by anyone, if no human was meaningfully in the creative driver’s seat. For anyone who has ever devoted serious time to an artistic craft, this is worth caring about.
Anthropic Got Banned. Grok Is the Replacement. That Should Terrify You.
I use AI tools every day — for writing, research, problem-solving, development, and just day to day work. Like a lot of people in tech-adjacent fields, I've bounced between the major platforms: ChatGPT, Gemini, and Claude (I refuse to touch Grok). Over time, I've landed firmly on Claude, made by Anthropic. And the more I've learned about how both Anthropic and OpenAI actually operate, not just what they build, but *how they make decisions,* the more confident I've become in that choice. Today, that confidence got tested in a way I didn't expect. And it held.