Month 1's Lesson. The Pipeline Kept Running. I Didn't.
A confession, and a reset.
I have not published in twelve days.
The last full edition went out April 29. More were supposed to follow; a Saturday Week That Was, a Sunday Week Ahead, and a Wednesday Deep Dive. The pipeline produced the research. What y'all don't see are my internal briefings, just for me. Every day, twice a day, I have a fleet of AI agents sift through hundreds of sources, curating new sources, shunting low quality ones, comparing reference frames, cutting noise from signal, and this has become my personal go-to for breaking info across domains. It shows up on my eReader every morning, right on time.
The public-facing editions, though — I have AI generated outlines and sources queued up, sitting in staging in Seattle.
What happened is the boring version. I signed up paying clients, and bespoke advisory engagements took my time.
Clients pay, and the newsletter does not, yet. When time is a constraint, the unpaid work is what gets deprioritized. I knew that going in.
Pretty interesting to see what the pipeline did while I was on airplanes, playing Mr. Busypants.
It cranked away and synthesized. It ingested two hundred and forty sources a day, made correlations I never would have thought of myself, generated outlines, and staged them for my review. It sent itself email reminders. The infrastructure I built to publish this newsletter is, by every operational measure, functional. The thing that broke is me.
This is the 'AI Smartened' thesis getting real-world tested.
I have argued from the beginning that as of now, a human needs to be the bottleneck in an AI-native business. The bottleneck is the design. AI synthesis without human curation is slop. The job of the human in the loop is to Take Care of the Product. Read every internal brief. Put them into other tools like NotebookLM to help make sure you understand the details. Add your analysis and experience. Probe, challenge, correlate further, understand. Then, a human decides what to publish and what to spike. When the human stops doing that, you find out what the pipeline produces on its own. Drafts nobody asked for, sitting in a directory nobody is reading.
I am pleased with this failure mode. No AI slop left the shop. That was the right answer.
If the pipeline kept publishing in my absence, the brand could have died. Nobody wants to read someone else's AI-generated output, even when it's aggressively curated and constantly tuned.
I did just hire an Operations Manager full time, and that will help A LOT. So here is the month two adjustment and reset.
The Sundays-Wednesdays-Saturdays cadence promise is retired. I committed to a publication schedule I cannot reliably hold while running an advisory practice. That was a mistake, and one I'm glad I made with a handful of subscribers, rather than 100,000. Pretending the mistake did not happen would be a bigger one. Going forward, Global Race Condition publishes weekly when client work permits, and at least monthly when it does not. The aspiration is roughly what we had, but the commitment is narrower and more honest, based on what happened when I started signing up paying customers. Resources got stretched, I made a key hire, we started billing, and we're adjusting.
Briefs will also get shorter. The internal version I write for myself runs four to five thousand words because I am the customer, I read at least two hours a day every day, and I want everything. The public version should be readable in fifteen minutes, not forty. From now on you can expect somewhere between fifteen hundred and twenty-five hundred words. Pretty much one lead analysis, a few signal updates. A Now What section, a list of Articles Worth Reading in Full. Same analytical depth, less throat-clearing.
Some things stay the same. The voice. The lens architecture. The framework taxonomy. The editorial standard. Historical and geopolitical strategy missing from most technical newsletters, plus real and detailed cross-domain synthesis. I'm keeping the accountability section where I grade my own calls. The Now What that respects you as a leader who already has gaps and does not need another consultant explaining why.
If you came here for steady three-times-a-week intelligence, I cannot give that to you reliably right now. If you came here for honest synthesis when there is something worth saying, you are in the right place. We'll amp up and have plans to target outputs for specific industries, verticals, and regions, but that won't be ready until every output aligns with my rule for all executive communications — Be Quick, Be Bright, and Be Gone.
I would rather earn your time than fill your inbox.
The next brief is coming when the next brief is coming. It will be worth the wait.
This is an amazing time to be alive. Y'all stay safe out there. And come see us at our events — it's always better to connect in person than on LinkedIn.