Scott Gingrich

Apparently, it's Arts and Crafts day at the office today! My prototype of my "The Gravity Engine" framework in 3D.

Something so satisfying to see the plugins you've just made show up in WordPress Updates.

I added $50 to my Claude budget and within 30 minutes, it's gone. I've spent an extra $350 already this month with this kind of absurb/impossible token burn. What's going on @AnthropicAI ?

What is it about a results desk that gets me to stay up late, especially when I'm 99% sure it's a foregone conclusion? The Alberta Referendum vote is today. Results will start to come in at midnight my time. Maybe it's the peak patriotism happening in Canada these days? At any rate, I've got the news streaming on my PC whilst I burn AI credits on building cool stuff!

I'm envious... Over the last year, my wife (and business partner) has gotten into home renovation and woodworking. Me, I've been immersed in taking our agency AI-native and building really cool systems. Here's the rub... Jen gets great "show & tell" while I get really niche, geeky, hard-to-explain cool stuff. What Jen does is infinitely easier to brag to my Mom about, post on Insta, and show off! I just want the world to know: I'm building cool shit too! (LOL)

AI systems that rely on multiple connections to run can be glitchy. Sometimes, I feel like Scotty in the engine room needing to know just where to give the Enterprise a little nudge to behave! It's not so bad when it's just me because I know the quirks quite well. But to my team users of the system, it means I have a new role of tech support...that part's not so fun. I keep working on ways to make what I build more robust. Oh, the life of the Alchemist Builder/Founder LOL.

Claude Opus 5 is being dumb today.

Welcome to September, the unofficial start of the New Year. For me, summer has flown by... I've been in a season of (mostly) heads-down building. In business we have these seasons that call upon us Founders to double-down on an ambition goal. I've been re-founding my Digital Marketing Agency to be AI-native. In the process, I've discovered so much more than just "adopting AI". We're creating something that is both familiar and brand new to the world. We're able to do things, bring value to our Clients like never before. I've had to say "no" to some summer fun this season, but it's easier when it's in the pursuit of a worthy quest, when building something valuable. As Fall approaches in a few weeks, I'm getting ready to switch to a more outward posture, deploy what's been built, and engage the world for feedback. Harvest time.

This really captured the Canadian sentiment with our current trade war with the US/Trump. Catchy tune too!

I've been working on our "next gen" stack for WordPress sites using the Spectra One theme by @WeBrainstorm . With things like the new Spectra Block's mega menu block, I've quickly become a fan of FSE. So far, all of our design scenarios have been handled very well.

The lag between effort and evidence is where most people quit. After a month's long build (building the systems that build the systems) I'm seeing the fruits of that labour. Phew!

Took a few minutes to explore some new Jazz (I'm very much a newb learning). I gave Cal (my AI OS in Claude) some songs I liked and asked it to ask me questions to work out taste profiles for: Late Night Whisky, Focused Work, Backyard Chilling, and Jazz Discovery/Appreciation. Now, I can chat with Cal about what I like and don't like and listen & learn.

THIS. "maxing out your internal token window" I didn't have a good articulation for this until I read @jonnym1ller article at @every. Running multiple sessions at the same time is a new level of cognitive load. In the past month, I've been experimenting with better ways to work and there's real payoff. Worth the read. After automation: Regulating our nervous systems will become part of work every.to/thesis-statements/j…

In May, Perry Marshall's print newsletter mentioned Imre Lakatos, a philosopher of science I'd never heard of. Three weeks later it caught a problem in our agency that was flying under the radar. Jen tightened a call-duration floor on a client's Google Ads account so a five-second wrong number would stop booking as an inquiry. Small fix, logged, done. A week later, different client, different session, she found an account counting one caller as two leads when they rang back. Cleaned it up. Also logged, also done. Two good calls. Both quick fixes not requiring much thought. @PerryMarshall's piece was on how you tell a research programme that's alive from one that's dying. The dying one never gets proven wrong, because every time reality disagrees with it someone bolts on an exception (a patch) and the theory survives the encounter. It just gets heavier. Eventually it can't predict anything at all, but it's never once been wrong. Which is per-account patching exactly. Every fix works. Nothing improves. So I ran the test across our Ops Log, which tracks every fix we make with the client, the date, and the reason. Bingo. Same fault, twice. Different time. Different account. It was in onboarding. The setup routine had drifted, so the same configuration step was going in differently depending on when it was done. This is the part that matters to me. I'd far rather fix one structural thing with real downstream consequences than get good at whack-a-mole. That fix didn't save us two hours. It removed a class of problem from every account we'll ever open, before those accounts exist. Nobody triages it again. No client ever sees a number in a report that doesn't smell right, which is exactly why it's worth money. So we've embedded the Lakatos test into Cal, our AI operating system. First instance of a fault is an incident: patch it, move on. Second instance of the same class, we go upstream and fix the cause. Cheap to implement. Compounds every month you leave it running. Somewhere in your business there's a next instance coming. Whether you catch it depends entirely on whether anyone's looking across the jobs instead of inside them.

When Claude and I are working on something with asymmetrical risk (low cost to double check, high cost if I don't), these have become two of my go-to phrases: - "Measure twice, cut once", - "Let's look for the second and third order effects"

For 20 years my experience only scaled as far as my calendar. Now it doesn't need me there. I spent years working my way out of doing the doing. Now I spend my days doing what looks like coding, but it isn't. Not exactly. When I'm at the computer, I'm not creating output, the client deliverables. I'm training AI to do it the way Scott would. I'm encoding my judgment, frameworks, methodology. (Turns out creating a library of markdown files is real IP.) Instead of output, I'm creating capability. My team can run the systems without me because I'm in them. And as we learn, so do our systems.

"re-founding"...eureka! Finally, I have a word to describe what we're doing at @gravitystackm Thanks @ivanhzhao

There's a lot of general skills (e.g. the excellent marketing skills by @coreyhainesco) available. It's almost overwhelming how many skills have been made public. I think the real and strategic moat value actually comes in creating skills that are tuned to very specific niche purposes. Writing a headline for a local Family Counsellor is a very specific sub-set of general headline writing rules. Not only are there the sensibilities of local audiences, there's also the regulatory compliance considerations that need to temper headline writing for such industries.

I think people who just use AI to ask questions and get answers are risking weakening their critical thinking. Those who are building the AI systems that answer those questions are strengthening their critical thinking. My experience is that a lot of critical thinking and "knowing your stuff" is required to build the systems.