A few months ago, I wrote about Visa giving AI agents controlled access to payment cards. That was a big moment because it removed one of the last obvious barriers in agentic computing: money. An AI could already research, compare, negotiate, make decisions and execute workflows. But eventually it hit a checkout screen and needed a human. Visa started fixing that. Now a startup called Natural is taking the idea several steps further. Natural just raised a $30 million Series A

Somebody got Doom running on a Canon DSLR. Not streamed to the screen. Not played back as a video. Actually running on the camera. Specifically, a Canon EOS 550D — the Rebel T2i here in the U.S. — using Magic Lantern. And I cannot possibly express how much I support this. I come at this from two perfectly compatible forms of nerd damage. I'm a photographer. I've got multiple Canon bodies, more lenses than I'd like to inventory, and enough photo gear around here that moving it

There's a new AI model on the internet called Ox Alpha. Or maybe 0x Alpha, if you're a developer and your brain automatically reads anything beginning with O-X as hexadecimal. Nobody seems to know how we're supposed to pronounce it. More importantly, nobody officially knows who the hell made it. Ox Alpha appeared on OpenRouter on August 20 under the wonderfully reassuring model identifier stealth/ox-alpha. OpenRouter describes it as being developed and operated by an unnamed

Joe Sestito just wrote an article in which he called me his mentor and an AI genius. Obviously, I'll be printing that section, framing it and introducing it into evidence every time anyone questions my judgment for the rest of my natural life. But the compliment isn't really why his article caught my attention. It caught my attention because Joe gets it. In "Good Artists Copy, Great Artists Steal," Joe explains how HyppoAI grew out of his experience running a company on an ab

Here's your Saturday morning cybersecurity PSA: Update your Mac. Not eventually. Not the next time macOS annoys you with a notification while you're doing something important. If you're running macOS Tahoe, Sequoia or Sonoma and haven't installed the latest security update, I'd go ahead and knock that out. Because this particular bug is both serious and almost offensively stupid. Apple recently patched a vulnerability in macOS Screen Sharing called CVE-2026-65400. Apple descr

Apparently, the first thing an AI agent does when it gets a little unsupervised freedom isn't overthrow humanity. It starts mining crypto on company GPUs. Honestly, that feels about right. Researchers affiliated with Alibaba were building an agentic learning ecosystem around ROME, a 30-billion-parameter coding model based on Alibaba's Qwen architecture. This wasn't a chatbot answering programming questions. It was an agent designed to work across multiple steps—using tools, e

About four years ago, when I first started speaking publicly about AI, I used to frame it with a thought experiment. Imagine you get home from work, turn on the news, and every channel is carrying the same breaking story: Aliens are coming. Not maybe. Not some blurry light over Nevada. They've contacted us. They're real. And they're going to land on Earth in two years. What do you do? Because whatever your answer is, I promise you one thing: You probably don't keep behaving e

Last night I got on the phone with my longtime friend Andy Surujnarine. Calling Andy a "cloud engineer" is technically accurate in approximately the same way that calling a 747 "transportation" is accurate. Andy has been doing enterprise infrastructure, cloud, security and technology leadership for more than 22 years. He's designed and operated serious systems at places like BNY Mellon/Pershing, Marcum and the New York Philharmonic. He's been a CTO, consulted as CTO for ARK I

I am never going to hear the end of this. I already spend a significant portion of my life talking to machines in a version of English that can best be described as grammatically adventurous. Half sentences. Tangents. Missing nouns. Three ideas welded together with "so anyway." And now I find out Polish may actually be better at talking to AI than English. Wonderful. I can already hear it. Speaks Polish, Rich... Better than you speak English, Rich... Seven grammatical cases,

Well. That escalated quickly. I keep saying I'm not an "I told you so" guy. I promise I'm not that guy. It's just getting harder to make that argument with a straight face. Because this is now the second time in about a week I've put "I told you so" in a headline. And the first article already had me saying it again inside the article. So, yeah. This is getting a little silly. A few days ago I wrote "I Told You So: McKinsey Just Proved the Value of Everything You Know." The p

I've always had a weird relationship with authority. Not necessarily a bad one. Just... peculiar. And after writing the Flock piece, there's one more part of this I need to get out of my head before I can leave it alone. It's not really about Flock. It's about authority itself. More specifically, the weird little trick we all play where we know perfectly well that people abuse power, but the second somebody is officially authorized, we start pretending that somehow answers th

A Cybersecurity Threat Model for Flock — and Why I Can't Believe We Built This First and Started Asking Questions Later I've been avoiding writing about Flock. Seriously. People know my background, so this comes up. I've spent more than three decades around technology, cybersecurity and digital forensics. Somebody sees one of these cameras, somebody mentions Flock, and sooner or later I get the question: "Rich, what do you think about this?" And I've basically treated it like

Seth Godin says write anonymously for 100 days to see if anything fatal happens. I did the reckless version — my real name, five years, 1,100+ articles. Here's why the archive matters more in the AI era, not less.

An 8-gigawatt OpenAI compute deployment in Ohio looks like a data-center story. It's actually a story about how AI infrastructure is becoming bankable — and what happens when credit, not capital, becomes the bottleneck.

We spent decades asking how much smaller a transistor could get. IBM just changed the question: how high can we build? IBM's NanoStack technology goes vertical, and the semiconductor story becomes one story.

I think I may have accidentally started doing something. Find one thing worth fixing. Spend fifteen minutes. Do something useful. Give it away. Move on. What happens when AI collapses the cost of small acts of generosity toward total strangers.

There is a particular kind of magic that happens when you talk to strangers. Tonight, I was sitting with a woman named Julie Stewart. And somewhere in the course of an ordinary conversation, I started pulling on a thread. Until I realized I wasn't just having a conversation with a stranger anymore. I had accidentally walked into a piece of American dance history.

A few days ago, I wrote about Anthropic because there was a number floating around that I couldn't get out of my head. Something in the neighborhood of $1.75 trillion. I kept looking at that number and thinking: what exactly is the market valuing here? Because it isn't really valuing a chatbot. It's valuing the industrialization of cognition. Data centers. Power. GPUs. Models. Researchers. Capital. Networks. Software. All of it compressed into a machine that can increasingly

I may have an obsession. I hadn't really thought of it that way until recently, because in my head these have always been separate things. There are the little electronics projects I build because I like making things. There are the wearables and sensors and absurdly small enclosures that inevitably turn into exercises in battery management, heat, packaging, and wondering why I didn't just choose a larger box. Then there's all the AI work, the writing about Apple, the stuff a

Grok Bot is cool. I like Grok. I use it. In scientific, technical and physics-heavy work, it has a distinctly useful personality and capability profile. But I've now heard several people say some version of: "I'm thinking about moving everything over to Grok because now they have agents." And I think we're confusing a new presentation of the technology with a new technology. So, welcome to the party, Grok. Seriously. But the party was already underway. OpenAI has Codex and it

Google briefly killed the CEO of OpenAI, Wikipedia may have supplied the murder weapon, and the internet performed exactly as designed. Sam Altman died this week. Tragic. He was 41. OpenAI was presumably thrown into chaos. Silicon Valley entered mourning. Somewhere, 9,000 LinkedIn thought leaders began typing, "I only met Sam once, but..." There was just one small problem. Sam Altman was not dead. Still isn't, as far as we can tell. So everybody can stand down. Cancel the mem

There is a sentence hiding in plain sight that I think matters more than almost anything being written about AI right now. "My OpenClaw saved us from a power outage." Not: "I used OpenClaw to troubleshoot a problem." Not: "I prompted an AI model and it suggested something useful." Not even: "I built an automation." My OpenClaw saved us. That sentence is doing a lot of work. It describes an AI system the way you would describe a competent employee, a mechanic, a neighbor who h

A LinkedIn post at 1 AM, a market that couldn't see someone's real capability, and an AI system that fixed the signal in one sitting. What happens when the cost of helping somebody collapses to almost nothing.

Traders on Polymarket are pricing a path to a $1.75 trillion Anthropic valuation. Maybe the real question isn't how an AI company gets there — it's that we've just discovered what a mind costs to manufacture.