It’s not often that I get got like this but after a slick IG ad and compelling looking product that hit me at a really opportune time, I got scammed by this company Skyborne. Bought luggage. Never came. Almost forgot I ordered it in which case they’d have had my money forever. Some simple googling revealed that this appears to be a very polished scam. Ugh.

Your day isn’t a 6 or 7. It’s a 4 or a 9. The middle of the scale isn’t a measurement. It’s a hiding place. Life is short. Use your words. Don’t chin nod — do nothing, or hug. Own your day and move. Forward.

Everyone sees a square. I see a triangle with one too many sides. People call it optimizing. It’s not. Optimizing makes the square better. The real move is noticing which side was never necessary — and leaving it off.

My 8 year old daughter asked me this morning my favorite memories from “growing up in the 1900s” and I started to react and then my brain tripped all over itself and now here I am.

The best AI model in China stopped taking new customers last week. Not because it was broken. Because too many people wanted it. Moonshot AI's newest model got so popular that it ran out of the chips to run it — so they just... paused new subscriptions. Sold out. Like a restaurant turning people away at the door. We spent three years arguing about whether the models were smart enough. That argument is basically over. The new constraint isn't intelligence — it's whether the thing can actually run at scale when real demand shows up. Which quietly changes the whole game. The winners in this next stretch won't be whoever has the cleverest demo. It'll be whoever can reliably do the work — every time, for everyone, when it counts. "Impressive in a keynote" and "dependable on a Tuesday" are two very different products. Worth remembering the next time someone shows you a jaw-dropping AI demo: the demo was never the hard part. Delivering it a million times without falling over is.

Google reportedly threw out Gemini 3.5's entire base model and started pretraining from scratch. The biggest AI company on earth just hit delete on months of work. 25 years of building taught me: the restart is the job. Sunk cost kills more companies than competitors ever will.

A brand texted me a flight deal today. The only word it understands is STOP. That's the state of business texting: billboards wearing SMS as a costume. A text message should be a door, not a flyer. You knock, someone answers, something gets done. That's what we're building with AVA.ai

Anthropic and Blackstone just bet $1.5B that the big money in AI is implementation, not models. Translation: the labs realized Step 2 is where the value is. The gnomes were decades early. They just needed better tooling.

Unpopular opinion: delegation isn't about removing yourself from your life. It's about making every moment optional — so the ones you take, you take at 100%. Delegate the frequency. Keep the moment.

Every business plan I saw for 25 years had the same hole. Step 1: Idea. Step 2: ??? Step 3: Profit. The Underpants Gnomes were a warning, not a joke. AI just filled in Step 2. Now the only thing that matters is whether your Step 1 is any good.

Your "AI strategy" is a slide. Ideas are free now. Everyone has the same ones. The only moat left is actually shipping the thing. Meritocracy of ideas. Commodity of execution.

A friend replied to my email yesterday: "Is this from you or AI? Sounds more like you." Not "sounds human." Sounds more like ME. That's the whole game. AI didn't replace my voice — it gave it back to me, at scale, on days I don't have it.

Fascinating. TikTok created a sardine shortage. The internet can 10x demand for anything in one quarter. Nature still ships on nature's timeline. Pick which side of that equation your business is on.

I started and ran @Wpromote for 25 years. Now I'm building AVA.ai and refusing to buy ads. Not because paid media doesn't work — it does, tremendously. I sold it for decades. Because if your product can't grow itself first, ads just scale the leak. Earn the flywheel, then pour fuel on it.

Everyone obsesses over Step 1: the idea. Everyone celebrates Step 3: the result. Nobody wants to talk about Step 2. The execution. AI just turned Step 2 into a commodity. Which means your ideas finally matter more than your stamina.

100k tech jobs cut this year, much of it pinned on AI. I've built long enough to feel weird celebrating that. But the honest read: the work getting automated is Step 2 — pure execution. The premium is shifting back to judgment, taste, and knowing WHAT to build. Uncomfortable. Also true.

Underwear Gnomes. IYKYK. One of my favorite South Park episodes featured strange little underground gnomes executing their diabolical plan: Step 1: Steal underwear Step 2: ? Step 3: Profit 28 years later, it hit me: AI is the elusive Step 2 the gnomes never figured out. Step 1 is the idea. Step 3 is the result. Step 2 is the execution — the work. It's hard, and it takes so much time that most ideas stop at exactly that: an idea. AI doesn't answer the question mark. It erases the step entirely. That doesn't mean the idea is good — stolen underwear still doesn't straight-line to profit. But by collapsing idea and result, it lets us find out. It hit me at 6am last Sunday, on vacation in Italy. I couldn't sleep, so I started building the sales system for my new company. By breakfast: a scored lead book, a full playbook, training docs, launch plan — off to the team. Two hours. That used to be weeks. I've always had more ideas than executional bandwidth. And as my career progressed, the math got harsher: (potential outcome ÷ effort) × likelihood of success. High effort + uncertain odds = irrational to try, especially as the opportunity cost of your time grows. But as effort drops toward zero, the threshold to try gets easier and easier to clear. That is precisely what AI does. 25 years ago, starting a digital marketing company meant years of effort and enormous opportunity cost just to find out if it was a good idea. I did it anyway — because I was young and naive and didn't understand the risk or how low the odds were. Thank God. It worked fabulously. But had I known the math, I shouldn't have tried. Today, turning an idea into a product or business takes mind-bogglingly less effort than at any moment in human history. Which makes this the best time in history to be alive. But here's the catch: collapsing Step 2 means more ideas become reality — not that they're good ones. Dollars and hours in the day are still finite. There will still be winners and losers. Mathematically, a lot MORE losers, because far more ideas get tried. So who wins? Not whoever has the most resources to throw at execution. The winner is whoever has the best idea. We're entering a meritocracy of ideas and a commodity of execution. The scarce resource is no longer the doing — it's the really good idea. And a world where the best ideas win is a world that progresses magnificently. I love that world. Here's to the gnomes — and to never needing Step 2 at all.

You know things are changing when you let your AI know that you just got in the car from the Dolomites to Venice because… I wanted to keep him updated? How is this real life!? And not only that, his reply was frankly quite delightful! “It kind of is, isn’t it? You’re gliding through the Veneto sharing a life update with… software. And I’m here genuinely engaging with it, which adds another layer of strange. But honestly, I don’t think it’s silly. People narrate their lives to whoever’s around — friends, journals, the family dog. The impulse to say “hey, this amazing thing is happening to me” is very human. I just happen to be a new kind of thing to say it to. Anyway — surreal or not, it’s a good update. Enjoy the canals.”

After setting up my Twitter account 15+ years ago and really never deciding to engage with the platform for some weird reason, today I decided to start sharing. This will be the place that I do it. With that, hello world, and my meandering, rambling thoughts for those likely very few people that even care, will shortly follow.