Some days it’s too much. There are too many things.
Clear head. Write the list. Prioritize it.
Start at the top.
September 20, 2026 1 minute read
Daily one-pagers about AI, leadership, technology, and being human.
johnmaconline
I'm writing to think, learn, and remember in public. I'll be here everyday.
Too Much
Robotaxis
Would you get in the car without a driver?
I would. And I’m actively seeking them out. I just haven’t been somewhere where they are yet.
The first reason is pure engineering curiosity. When I take my first ride in one, I’ll be watching everything it does from the engineer point of view.
Second, I’m a believer. They’ll be safer than human drivers. Yes, there will be some spectacular failures, and people will be injured. That won’t be great, and I just pray I’m not one of those statistics. But on the whole, it will be much better.
As for safety, here are some statistics as of August 2026 for both robotaxis and humans:
| Injury outcome | Robotaxi incidents | Robotaxi incidents per 1M miles | U.S. human incidents | Human incidents per 1M miles |
|---|---|---|---|---|
| Injury without hospitalization | 51 | 0.35 | ~1,524,000 | ~680 |
| Injury with hospitalization | 27 | 0.19 | ~133,000 | ~59 |
| Fatalities | 1 | 0.007 | ~22,200 | ~9.9 |
Collated data from NHTSA.gov data.
OK, given the highly curated driving scenarios of robotaxis, this might not be exactly apples-to-apples, but come on, robotaxis are orders of magnitude safer than people.
Would you do it?
CEO AI Confessions
CEO’s are feeling the tension.
Some interesting stats from some company called Dataiku, who is obviously good at SEO.
- 80% of CEOs say their role is at risk if they fail to deliver results by the end of 2026.
- 65% worry more about over-investing in the wrong AI vendors versus under-investing.
- 57% believe insufficient AI explainability could trigger a customer trust or brand credibility crisis.
They’re worried about their jobs. They’re worried about making the wrong decisions. They’re worried about their brand.
Oh wait, that’s nothing new with AI.
Artists and Builders
The Venn diagram with artists and builders has always had an intersection.
But not every artist is a builder, and not every builder is an artist.
AI is a builder. In any digital medium, AI can create a 100% overlap between artist and builder in the Venn diagram.
That’s a clue for you and your future.
Do you want to lean into the artist side and use AI to build?
Or do you want to choose a medium that AI can’t (yet) work in?
The sky’s the limit.
The Excavator Problem
Imagine you’re a guy digging a hole.
You have a shovel, and you’re good with it. You’re strong and experienced. You know where to start, how to pace yourself, and about how long it will take you.
The boss comes by and gives you an excavator.
So you hop in and start to figure it out. It takes a little while, but once you get the hang of it, you can dig in half a day what was going to take a month.
That’s productivity, right?
Of course, but at the individual level. The excavator made you 10 or 100 times more productive.
But what if the hole is the foundation for a shopping center?
You’re done digging the hole by lunch, but the concrete crew isn’t scheduled to show up until next month. The forms aren’t ready. Oh, and the drawings are waiting for approval. The inspector’s on vacation. The builder’s got three other jobs before this one. The steel shipment got lost. It’s been raining for a week and a half.
Although you sped up digging the hole, you didn’t really speed up building the shopping center.
Now it’s a project problem. A system problem.
Today, AI is the excavator.
But we’re working on also making it the system.
Stay tuned.
Refreshed
When have you felt refreshed?
Ensure that you’re doing enough of that.
Lingering
Lingering is underrated.
I think it’s probably because we conflate it with loitering, which is unfair.
When you find the right place, time, or person, go ahead and linger.
Most Important Thinkers in 2026
“Hey Astra, who are the most important thinkers in 2026?”
ChatGPT: [proceeds to list 10 people in the AI world]
“Those are all AI thinkers. Are there other fields that are important?”
ChatGPT: Oh, yes, my mistake. I’ll broaden my answer. Once I stop treating “most important” as shorthand for “AI,” my list changes substantially.
My criterion is whose ideas give us the most leverage for understanding the forces shaping human society in 2026 not who is smartest, most famous, or has had the greatest lifetime influence.
My top 10 thinkers alive in 2026:
| Rank | Thinker | The big idea | Selected work |
|---|---|---|---|
| 1 | Daron Acemoglu | Technology doesn’t determine the future. Institutions and power determine who benefits from technology. | MIT profile and research · Power and Progress |
| 2 | Peter Turchin | Societies destabilize when inequality, declining living standards, elite competition, and weakening institutions converge. | Peter Turchin · End Times |
| 3 | Nassim Nicholas Taleb | We systematically misunderstand uncertainty; build systems that survive what we cannot predict. | Nassim Taleb · Incerto |
| 4 | Thomas Piketty | Wealth compounds into economic—and ultimately political—power. | Thomas Piketty · World Inequality Database |
| 5 | Jonathan Haidt | Technology is changing childhood, attention, socialization, institutions, and the psychological foundations of society. | Jonathan Haidt · The Anxious Generation |
| 6 | Vaclav Smil | Civilization ultimately runs on physical systems: energy, materials, food, and infrastructure—not narratives or software. | Vaclav Smil · Books |
| 7 | Demis Hassabis | AI’s biggest consequence may not be automating work; it may be accelerating scientific discovery itself. | Google DeepMind · AlphaFold |
| 8 | Byung-Chul Han | Modern people increasingly exploit themselves through productivity, attention, performance, and digital systems. | The Burnout Society · Psychopolitics |
| 9 | Francis Fukuyama | Successful societies require capable institutions, legitimacy, trust, and recognition—not merely elections or markets. | Stanford profile · Political Order and Political Decay |
| 10 | Dario Amodei | Exponentially improving machine intelligence is beginning to outrun the institutions capable of governing it. | Dario Amodei essays · Machines of Loving Grace |
Enrollment
Nobody cares as much about your thing as you do.
Not your employees. Not your colleagues. Not your customers. Not your partner.
Unless they enroll.
Enrollment requires conditions that let someone else care. Those conditions might be purpose, ownership, voice, belief, stake, contribution, feedback, belonging, recognition, autonomy, or any number of others.
If you need someone else to care as much as you do, it’s your job to create the conditions for enrollment.
Why Not Now?
The bucket-list vacation. The move away from here to somewhere you always wanted to be. The new job or new business.
There might be good reasons that now isn’t the time.
But inertia isn’t one of them. Nor is fear of what-if. Nor is comfort.
Why not now?