Entrepreneur
Key Takeaways
- The real bottleneck for founders isn’t speed — it’s that their thinking gets spread across thousands of tiny, half-formed decisions. AI will gradually start taking that load off their plates.
- Instead of just summarizing or generating content after a human decides, AI will act as a filter — removing noise, structuring vague problems and eliminating weak options before founders ever see them.
- AI will likely become a “system layer” for decision support (not a replacement for thinking), freeing founders to focus on strategy, creativity and real conversations.
When people talk about AI in business, it almost always comes down to the same promise: write faster, generate more, calculate in seconds. More emails, more tests, more code, more decks. But if you look at real founder life, the bottleneck is almost never speed. The bottleneck is the constant stream of daily decisions you have to carry in your head.
I’ve been running tech startups for 11 years. We started small, and today we work across dozens of countries. Over that time, I’ve seen the same pattern again and again: The problem isn’t that founders “don’t think enough.” The problem is that their thinking gets spread across thousands of tiny, half-formed decisions. That’s the kind of load AI will gradually start taking off your plate. And it’s worth noticing that shift now.
From “thoughts” to cognitive states
The popular line about “tens of thousands of thoughts a day” sounds dramatic, but it doesn’t help you run a company. Cognitive science offers a more useful lens: Instead of counting thoughts, it looks at cognitive states, more stable modes of attention your mind moves through during the day. Research by Tseng and Poppenk (2020) points to roughly 6,000-8,000 of these states per day.
For business, this isn’t an academic detail. It’s a new unit for understanding performance. What matters isn’t how much you think, but which states you spend your time in and how often you bounce between them.
The 3 layers of a founder’s day
A founder’s day tends to fall into three layers.
The first is active anchor themes, usually five to 15 per day: money, team, product, market, long-term horizons. These set the emotional tone and color everything that follows.
The second is conscious problems and ideas, usually 20 to 80 per day: work tasks, everyday decisions, plans, doubts, internal conversations.
The third is micro-ideas and their variations, roughly 100 to 300 per day. Around 80-90% of these repeat and don’t add new meaning. This background creates the feeling of being busy, but rarely produces real progress.
What drains you isn’t strategy, and it isn’t your task list. It’s the repeating noise of micro-ideas that pushes out time for your anchor themes. Until now, technology barely saw this internal structure. It just sat on top of it.
Yesterday’s technology helped after the decision
For most of tech history, machines kicked in after a human had already made a choice. A calculator sped up arithmetic but didn’t tell you what to calculate. A spreadsheet helped process data but didn’t form the question.
The heavy part — framing the problem, choosing what to focus on, deciding which fork in the road to take — stayed in the brain.
Most AI products today still live in that same logic. They generate text after you’ve decided to send an email. They summarize a call after you’ve decided it matters. That’s useful, but it barely touches the core of entrepreneurial load: the swarm of decisions that exists before tasks, and never becomes clean tickets.
The next wave starts exactly there.
AI as a partner: Not instead of you, but next to you
There’s an obvious dead end: letting AI make the founder’s important decisions. Which markets to enter. Where the ethical lines are. If you outsource strategy and meaning to a machine, the quality of thinking will degrade quickly. That territory should stay human.
The interesting space is somewhere else.
Look at your day without filters. What makes fatigue build fastest? Not a strategy discussion, but dozens of small choices: “How do I answer this?” “Which version do I send?” “What goes first?” Not one big fork in the road, but fuzzy questions like “Something feels off in marketing, but what exactly?”
AI can already work with that layer better than most people expect. Not by “thinking for you,” but by removing friction around your decisions.
It can eliminate obviously weak response options before you ever see them, so you’re choosing from two or three drafts, not 20. It can act as a first filter on incoming signals, so what reaches your attention isn’t everything, but what actually has a chance to change the situation. It can take a vague, intuitive worry and turn it into a structured question: not “everything is breaking,” but “we have three hypotheses, here’s where we lack data.”
It can hold and update “what if” scenarios outside your head instead of forcing you to re-run the entire simulation from scratch every time. It can spot stable patterns in how you make decisions: what you constantly delay, what you automatically agree to, where you repeatedly take on too much.
This isn’t about replacing thinking. It’s about creating conditions where the same thinking gets spent on different things: meaning, direction, real conversations with people, instead of endless sorting of noise.
From a set of tools to a system layer
If the main constraint for a founder isn’t typing speed but the architecture of daily decisions, the role of AI starts to change. The question stops being “which AI services should I add?” and becomes “what if AI sits not on top, but underneath, as a layer that works on the load before it turns into tasks?
A layer like that could filter noise before it hits your attention. It could shape decisions before they land on your desk. It could hold context between meetings so important threads don’t dissolve. Not by replacing a strategic yes or no, but by reducing the number of unnecessary “maybe” moments.
We’re still early. The market is still dominated by point solutions that solve fragmented tasks. But the direction is visible. Models are getting better with context, learning to hold memory over time and seeing patterns inside your own decisions. The ingredients for a system layer are starting to come together.
The value of that shift won’t show up as one “magic button.” It will look more like weeks that feel calmer: fewer jolts, fewer repeating thoughts, less of that feeling that your day shattered into hundreds of micro-decisions. And more space for strategy, creativity and real conversations.
Why it’s worth thinking about this now
It’s easy to treat AI as a tactical tool: save an hour, speed up a task, remove a couple of routine steps. But if AI moves even partially closer to the decision process itself, it will start changing what it means to be an entrepreneur.
It changes how much of your day goes to real thinking versus servicing noise. It changes the standard for what you personally work on and what you hand to a system for pre-processing.
The AI conversation stops being only about “the right prompts.” It becomes a conversation about the role you’re willing to give a system inside your decision-making process. Not to hand off responsibility, but to stop spending a rare and expensive resource — your own thinking — on the forks in the road that should have left your head a long time ago.
That’s the most interesting near-future value of AI for entrepreneurs. Not doing more and more, but learning to manage the thing that ultimately defines the business: the quality of attention and decisions.
Key Takeaways
- The real bottleneck for founders isn’t speed — it’s that their thinking gets spread across thousands of tiny, half-formed decisions. AI will gradually start taking that load off their plates.
- Instead of just summarizing or generating content after a human decides, AI will act as a filter — removing noise, structuring vague problems and eliminating weak options before founders ever see them.
- AI will likely become a “system layer” for decision support (not a replacement for thinking), freeing founders to focus on strategy, creativity and real conversations.
When people talk about AI in business, it almost always comes down to the same promise: write faster, generate more, calculate in seconds. More emails, more tests, more code, more decks. But if you look at real founder life, the bottleneck is almost never speed. The bottleneck is the constant stream of daily decisions you have to carry in your head.
I’ve been running tech startups for 11 years. We started small, and today we work across dozens of countries. Over that time, I’ve seen the same pattern again and again: The problem isn’t that founders “don’t think enough.” The problem is that their thinking gets spread across thousands of tiny, half-formed decisions. That’s the kind of load AI will gradually start taking off your plate. And it’s worth noticing that shift now.
From “thoughts” to cognitive states
The popular line about “tens of thousands of thoughts a day” sounds dramatic, but it doesn’t help you run a company. Cognitive science offers a more useful lens: Instead of counting thoughts, it looks at cognitive states, more stable modes of attention your mind moves through during the day. Research by Tseng and Poppenk (2020) points to roughly 6,000-8,000 of these states per day.
Read the full article here
