Nothing in your business moves faster than the decisions you are still waiting to make.

Good Tuesday. Two solo sessions this week. One was about a decision you have been rehearsing in your head for months. The other was about a company that quietly stops the week you go on vacation. They are the same problem wearing different clothes. In both cases you are the thing everyone is waiting on, and in both cases you are waiting on a feeling you have decided to call readiness. The obvious response is to prepare more, and preparation is real work right up until the moment it becomes avoidance with a respectable name. After that it produces nothing. Waiting creates exactly zero confidence. It also creates exactly zero systems.

This week, pick one thing that routes back through you and move it a single step. Make the call you keep drafting. Or take one process that currently lives only in your head and write it down in a form someone else could follow without asking you a question. Not twenty steps. One. Action creates feedback, feedback lets you adjust, and clarity almost always shows up after you move rather than before.

George

Key takeaways

From this week's solo sessions with George.

Confidence is downstream of action. The sequence is courage, action, experience, evidence, confidence. The step gives you experience, the experience gives you evidence that you can handle what is in front of you, and stacked evidence is what belief is actually made of. You cannot think your way into the confidence that only comes from having done the difficult thing.

Diagnose the fear before obeying it. Ask one question: is this fear warning me about real danger, or is it telling me I am outside my comfort zone? Those are two completely different signals, and confusing them costs people years. Growth requires unfamiliar territory, so fear showing up is a condition of the work, not a verdict on it.

Ask what stops when you stop. Step away for a week, actually or on paper, and list everything that halts. Marketing, content, sales follow up, client fulfillment. That list is not a guilt trip, it is a map of every place you are still the bottleneck, and it tells you what to fix first.

Systematize one process, not the business. The ladder is do it, document it, delegate it, automate it, optimize it, and it works one repetitive thing at a time. Pick the activity you touch every week (content, follow up, onboarding, reporting) and move it up a single rung. Trying to fix the whole company at once is how founders end up a year later with the entire operation still living in their head.

The Courage to Act · Build an Authority Engine

Guest spotlight

Adrian Rosebrock

PhD computer scientist and founder of PyImageSearch, acquired in 2021, now CEO of Chief of AI, on why the AI everyone can buy is the AI that gives you no advantage.

Most businesses think they have adopted AI. They have the accounts, the prompts, maybe one power user doing impressive things. Adrian's point is that none of that is an edge, because your competitor bought the identical thing for twenty dollars a month. The advantage is sitting in the data, processes and relationships you already own and have never captured. This is the same bottleneck conversation from the solo sessions, just pointed at your technology stack.

“There's no moat around a $20 a month account. You have to be looking internally: what's your competitive advantage, what are you doing that no one else can, and start engineering your AI efforts surrounding that.”

Three things worth stealing from the conversation:

1. Know which wave you are in. Adrian splits the market into three waves. Access, roughly 2022 to 2026, measured success by how many people on your team were using AI. Economics, roughly 2025 to 2028, asks what it costs, what the return is, and what your governance policy says. Differentiation, roughly 2028 to 2032, is decided by proprietary data and custom models. If you are still counting seats, you are competing in a wave that is already closing.
2. Your own data beats better prompts. Bridgewater had recorded internal conversations and documented the reasoning behind investments for years, so they owned a dataset nobody could copy. Off the shelf prompting got them around 50% accuracy, a coin flip. Expert written prompts pushed it to 60 to 70%. Fine tuning a model on their own data took it past 80%, and self hosting that model cut token costs by 14x. Start capturing decisions and reasoning now so you have something worth tuning on in three years.
3. Write the policy before the leak. The biggest mistake Adrian sees is governance, precisely because it is boring. A Midwest bank had to self report a violation after an employee, doing nothing nefarious, uploaded personally identifiable information to ChatGPT, and studies he cites suggest 50 to 60% of CEOs believe their people have already done something similar. Write real policies with your lawyer, educate your team on the consequences, and stop file downloads to local disk so work stays in permissioned cloud folders.

Listen to the full conversation →

This week on the show

Resources & links

Where this leads

When you’re ready to go deeper than a weekly report, there are two ways in.