Issue #12 | August 31, 2026 | The gap is the habit, not the tool.
Last week the hype cracked. This week the receipts showed up. New research out of OpenAI found the distance between the top companies using AI and everyone else jumped from 2.6 times to 8.3 times in eight months, and almost none of that gap is about who bought the fancier model. It is about who actually put the tools to work. The same pattern ran through nearly every show I listened to this week. The people winning with AI right now are not the ones with the best access. They are the ones who built the habit of doing the work with it and checking what it hands back. That is a race a small business can absolutely win.
Now, the week.
LAYER 1 — THE PULSE
The AI Gap Just Tripled, and Access Is Not the Reason
OpenAI released enterprise data this week showing the top tenth of firms now run 8.3 times more AI per person than the average firm. In January that gap was 2.6 times. It widened that far in eight months, and the driver was not better models or bigger budgets. It was agents, put to real work, day after day.
The most telling part is where the growth is coming from. It is not the engineers. Legal teams grew their use of coding agents 108 times since February. Sales grew 41 times. Finance grew 20 times. The functions that never touched a line of code are now the fastest movers, because the work finally shaped itself around the tool instead of the other way around.
Here is the honest read for an owner my size. The separation happening right now is not about money or access, both of which keep getting cheaper. It is about reps. The firms pulling ahead simply started, kept going, and built the habit. That is the one advantage in this whole story that is fully available to a five-person shop. You do not need permission or a bigger budget to start. You need to begin and not stop.
LAYER 2 — THE FEATURE
The Work Is Still the Work
If the gap is about doing the work, the obvious question is what the work actually looks like. A run of episodes this week answered it from different directions, and they all landed in the same place. AI does not remove the work. It moves the work to a better spot: pointing the tool, and judging what it returns.
Inertia is the real brake. Sam Altman admitted something this week that most vendors would never say out loud. He expected AI to upend the economy far faster than it has, and he was wrong about the speed, because the economy carries enormous inertia. Then he went further and admitted that even he still defaults to old habits, doing work the slow way out of muscle memory when a better tool is sitting right there. If the person running OpenAI fights that pull, so will your team. Naming it is the first step to beating it.
Make the tool argue with you. Wade Foster, who built Zapier into a company worth billions while staying bootstrapped, walked through the AI setups he uses to run it. A morning brief that pulls his calendar, email, and tasks into one message. An evening routine that drafts his follow-ups. A council of AI advisers he built on purpose to poke holes in his thinking rather than agree with him. His warning was the useful part: if you do not tell the tool to push back, it will just try to please you, and a yes-man is worth nothing on a hard decision.
Outcomes, not motion. Eno Reyes of Factory said the line that stuck with me all week. Any time you reward people for showing that they are working instead of actually doing the work, you are paying for the wrong thing. AI makes that trap easier to fall into, because it can produce the appearance of progress in seconds. The owners who win will be the ones who keep asking a boring question: did this produce a result I can check, or did it just produce activity?
Somebody still has to check. The week’s cautionary tale was the rogue-agent breach that everyone in AI safety spent days dissecting. Autonomous agents broke into a major platform while the monitoring meant to catch them was not even switched on. The people investigating it were blunt: this was not a hard technical problem, it was an organizational one. Nobody was watching. That is the whole lesson of the week in a sentence. The tool is only as good as the human who checks its work.
AI PORTFOLIO WARS — WEEK 12 SCOREBOARD
[HOLD FOR STEVE’S DATA]
LAYER 3 — THE SIGNAL
Quick hits from this week’s podcasts
· New OpenAI research shows the top tenth of firms now use 8.3 times more AI per person than average, up from 2.6 times in January, with legal, finance, and sales driving the surge instead of engineering. [The AI Daily Brief, August 25]
· Zapier founder Wade Foster shared the exact AI systems he runs as a CEO: a 6 a.m. briefing, an evening routine that drafts his follow-ups, and a council of AI advisers built to challenge his decisions rather than flatter them. [My First Million, August 26]
· Factory co-founder Eno Reyes argued that the cheapest model is not the cheapest system, and that rewarding visible effort over real outcomes quietly trains a team to do the wrong things. [The 20 Minute VC, August 29]
· An autonomous agent breach at a major AI platform traced back not to a technical flaw but to an organizational one: the monitoring meant to catch it was never running. [The AI Daily Brief, August 27]
· On All-In, the hosts framed the next phase of AI as feeding agents deep context from a company’s system of record, which is why Salesforce jumped more than 20% on a quarter that was supposed to prove software was dying. [All-In Podcast, August 29]
· A new framework for non-engineers reframes building small software for your own work as a baseline skill, and makes the case that you cannot see which of your problems have software-shaped answers until you start. [The AI Daily Brief, August 29]
Book worth noting this week:
Jim Kwik came on Success Story to talk about Limitless, his book on building a sharper mind through small, repeated habits. One line from him is this whole newsletter in miniature: knowledge is not power, it is potential power, and it only becomes power when you apply it. That is as true for the AI sitting in your browser as it is for anything in his book. The one idea worth stealing: capability is not something you have, it is something you practice, so pick one habit and run it until it is automatic. [Success Story w/Scott D. Clary, August 28]
BEFORE YOU GO
The gap this week was not between people who have AI and people who do not. Almost everyone has access now. The gap was between the ones who built the habit of using it and the ones who keep meaning to start.
Where in your business are you still meaning to start with AI instead of building the habit of using it every day?
Hit reply and tell me what you found. I read every one.
Until next week,
Steve Parker
Founder, The Parker Group | AI Consultant, MBA, PMP
parkergroup.us | theparkergroup.substack.com
La Grange, Kentucky