"The teams that pull ahead in the next five years will not pull ahead by being smarter at writing boilerplate. They’ll pull ahead by choosing the right agent foundation and spending their engineering hours on the problems only they can solve."
"Ideally, each organisation should design and run their own capability to observe, test, interpret, and improve agentic systems as they operate, as an ongoing function embedded into how systems are built and run."
A photo booth entrepreneur runs 20+ AI agents handling quoting, booking, content editing, marketing, bookkeeping. Sounds modest. But that's a labor-heavy local service now scaling across regions with one person.
After a year of experimenting with LLMs, my data looked like it had been distributed by a hand grenade across file systems and cloud services. I used Claude, NotebookLM, and Perplexity like a five-year-old uses crayons. Time to grow up.
Most people think tyranny requires dictators and military parades.
I've seen it in schoolyards, corporate hierarchies, structureless startups, and algorithms that won't explain themselves. We've grown disturbingly comfortable with its presence.
Excellent data can CREATE business opportunities you weren't even seeking.
If you're not actively listening for signals in the noise, you'll miss the idea that could lift you from the swamp next year.
Spent 30 minutes simplifying a Make scenario. Past Me used three filters where two seemed enough. So I 'fixed' it. Everything broke. Past Me had reasons. Present Me just couldn't remember them.
Everyone's published a 'how I built my second brain' article. Full of Obsidian workflows and Claude prompts. I don't care how to build one. I want to know how to use one. Sixteen techniques. No recipes. Just a map.
Business models look the same. Thinking processes look the same. Pricing tactics look the same. Every startup copies Airbnb's pitch deck. Nobody does anything stupid.
"The only way to secure the promised returns is a hard and verifiable workforce redeployment clause. Leave the social contract blank, and workers will fill it with sabotage."
"Usage data from the major labs shows the growth is not coming from enterprise seats. It’s coming from heavy individual users and small teams inside companies."
With agility, we aim for acceleration.
With viability, we aim for calibration.
In 2026, nobody wants to accelerate themselves to death.
"Expectations about productivity are rising, even though how AI actually helps here remains something of a mystery. One thing it does do is increase surveillance, which means employees are under even more scrutiny and pressure."
Nobody mourns Rembrandt's pigment-mixing recipes. Most painters buy materials in shops now. So why do we panic when writers offload spelling to machines?
A client wanted 100 copies of my book but couldn't invoice for 'books.' Asked me to inflate my keynote fee, bring the books 'for free.' Corporate budgeting is absurd. But adapting to absurdity is the job.
Solo work bundles freedom with risk, cash-flow swings, and that charming moment when a client pays three months late while your mortgage remains weirdly punctual.
You don't have a "network."
You have a distributed museum of people-you-once-met, curated by SaaS vendors, funded by your subscriptions, held together by hope and the fantasy that you'll "clean this up someday."
Most of us systematically undervalue ourselves.
We anchor to historical salaries. We adjust to what we assume orgs can afford. We let imposter syndrome negotiate against us.
You are probably worth more than you think.
Tensions between autonomy and coordination, spontaneity and discipline? Those aren't problems to solve. They're features to manage. Organizations thriving with AI agents surf these tensions instead of fighting them.
"That is the only question worth asking. Not: are these people good enough? But: is this system designed well enough to let good people perform?"
"People are overthinking how to use AI. I say, very simply, that they should use AI to do the things they know they should do and aren’t doing. Don’t overcomplicate it."
The physiotherapist wrote a book about psychosomatic injuries. So naturally, he diagnosed my hamstring pain as being in my head. When your favorite tool is a psychological hammer, every aching body part looks like a repressed emotion.
E-books about writing e-books. Podcasts about making podcasts. YouTube channels about growing your YouTube channel. Monetized recursion engines are the business model equivalent of a snake eating its own tail.
AI can summarize a 200-page board pack in minutes. Whether that's efficiency or an admission those pages shouldn't have existed in the first place remains an open question.
Everyone stares at downsizing and misses widesizing. They see layoffs at the cathedral and ignore the bazaar springing up outside. Millions of solo businesses launching won't make headlines. Still happening.
AI evangelists breathlessly declaring AGI by Tuesday get my eye-rolls. Doomsday prophets shrieking about mass unemployment get my pushback. A recovering optimist with trust issues sees both the promise and the bullshit.
"Reality eventually makes a fool of every model. Every model is eventually proven wrong. But some models are still useful. And the intellectuals who cling to their un-useful models in the face of reality are the ones who become the biggest idiots."
"When AI becomes the coordination technology, agents need to understand the commitment structure, dependency map, and capability boundaries. Without explicit organizational semantics, AI becomes an entropy multiplier."
