Joshua 1:6 Be strong and courageous, because you will lead these people to inherit the land I swore to their ancestors to give them.

What am I doing with my "time off"? Fine-tuning my first model on my @Nvidia DGX Spark... It's not just for fun, I'm creating something that gives me a competitive advantage in 2026. I did the math, and each iteration of what I'm doing would have cost $6K on OpenAI using GPT 4o

When Medicare payments to physicians fall behind the rate of inflation, everyone feels the strain. Learn how these cuts affect you.

I use AI extensively in my own writing. Not because I want it to “decide” what to say — but because I’m very clear about what it’s allowed to do, and what it isn’t. That distinction turns out to matter a lot more than people realize. When AI writes for me, it’s executing:

I can’t tell you how much these acknowledgments on the impact of my work means. This exactly why I created and published DAPM.

https://t.co/2ejp6i1MBL

Bear Down!!! I'm not going to lie... I turned off the game and tuned back in with 1.5 minutes left.

I feel old. What did the baby want me to get her?

"The question that actually determines success is simpler—and almost never asked: Are we ready to operate a private cloud as a product for the next 3–5 years?"

I asked @grok for a TL;DR Post for my "4+1: The Enterprise AI Field Manual – Why AI Platforms Fail — and How Enterprises Actually Fix Them" Enterprises don't buy "AI platforms"—they inherit responsibility for defensible judgment in autonomous systems. The core problem:

AI doesn't create decision authority problems. It exposes the ones you never documented. Before automation, a human reviewed invoices. Most approvals were routine. But occasionally, something felt off: "This vendor is new and the amount is high—I'm flagging this." "This pattern

Because I have no publishing strategy! My book is live on Amazon. 4+1: The Enterprise AI Field Manual: Why AI Platforms Fail — and How Enterprises Actually Fix Them

Two years later and I'm still right, which isn't the flex it sounds like. M3 Ultra. More unified memory. RDMA over Thunderbolt. Four-node clusters seeded to YouTubers. Apple keeps stacking hardware capabilities. The math: ~$26K for four Mac Studios. Roughly 50% of H200 compute

Your AI agent just auto-scaled your infrastructure to handle a traffic spike. Cost: $47,000 in 6 hours. Now: Three teams in a room pointing fingers. Platform team: "The agent did what it was configured to do." Finance: "Who approved this spend authority?" Product: "We needed

No more big content drops for the rest of the year, I promise!!! 🤞🏽

There are things that irritate me. But it's been about a month and a half since I lost my brother. I tell you what, those irritations don't seem so bad.

Most enterprise failures under automation aren’t execution problems. They’re authority problems. Systems start making decisions. No one explicitly owns them. Accountability shows up during the incident review. I’ve published a paper that explains this pattern and why it keeps

I tried to make this diagram work in the book. I really did. But sometimes the medium loses to the message. This diagram shows three real operating models for AI judgment (Layer 2C) that I keep seeing in enterprises: • Platform-led judgment • Governance-coupled judgment •

Most enterprise AI failures aren’t model failures. They’re responsibility failures. My upcoming book breaks down why “AI platforms” stall in production—and what actually has to be owned, governed, and defended when AI starts making decisions. More soon.

This a whole different level of mental and physical toughness. DO IT ANYWAY https://t.co/E9dKsrTlGa via @Casey

How in the world did Mermaid chart become a standard?