Jab Petersen / AI Won’t Save You

Chapter 2 · The Cost of Keeping Every Door Open

AI keeps every door open. Progress needs you to close some.

Every new comparison feels like progress. Usually it is motion sickness: constant intellectual movement with very little distance covered.

At 11:47 p.m., Elias has six tabs open. ChatGPT. Claude. A course about starting an agency. A video called "The SaaS Opportunity Nobody Is Talking About." A spreadsheet comparing five business models. And a blank document called "FINAL PLAN" that has been open for forty-two minutes.

Elias is not lazy. That matters. He is doing what intelligent people do when choosing feels dangerous: he keeps comparing.

When options become anesthesia

Options are valuable. Savings, skills and several suppliers all buy freedom. But options have a side effect. An unchosen option cannot disappoint you. The business you have not started can still become huge. The book you have not finished can still be brilliant.

Possibility is clean. Reality is messy, because reality requires exclusion. If you spend Tuesday on the software idea, you are not spending it on the agency.

The word decide comes from Latin roots connected to cutting off. A real decision does not merely pick an option. It cuts off the others, at least for a while. Most of us want the decision without the cutting. AI makes that almost effortless: whenever anxiety shows up, you can ask for one more comparison.

The option tax

Every open option costs more than attention. It costs evaluation. If you are weighing four businesses, every new piece of information has to be interpreted four times. Does this trend help the agency? Hurt the software? Should the newsletter move up?

That is the option tax. AI raises it because it keeps every branch intellectually alive. A new model release is a reason to reconsider your stack. Someone else's success is a reason to reconsider your strategy. None of these facts are fake. The problem is that you have no threshold for reopening a decision.

The One-Bet Rule

The fix is not a permanent commitment. Jeff Bezos wrote in his 2015 shareholder letter that most decisions are two-way doors: if you do not like what is on the other side, you can walk back through. So use a defined window. For many projects, thirty days is enough to create evidence.

Write one sentence:

For the next thirty days, my one bet is ________.

Then start a second list called NOT NOW. Every competing idea goes there. Not because the ideas are bad, but because they do not get to interrupt the experiment.

Then change what you ask your AI. Not What else could I do? but:

Given that this is my bet for thirty days, what is the highest-value action I can take today? Do not suggest changing the bet unless new evidence directly contradicts the assumptions behind it.

The model is no longer expanding the map. It is helping you move.

Set reopening rules in advance

A commitment window does not mean ignoring evidence. It means deciding beforehand what evidence is strong enough to reopen the choice.

  • Business: "I will reconsider if, after twenty real sales conversations, fewer than two people describe a painful version of the problem."
  • Learning: "I will not switch programming languages because another one is trending. I will reconsider after shipping the project."
  • Content: "I will publish twice a week for eight weeks, then review retention and replies."

This protects you from mood-based strategy, and it gives AI a clear job: watch the evidence you defined in advance.

The 24-hour version

If thirty days feels too long, start smaller. Whenever you are stuck between options, ask:

What can I do within twenty-four hours that would produce information unavailable from comparison?

Torn between two freelance niches? Message three prospects in each. Two content formats? Publish one of each. Design or coding? Spend two real hours on each, actually doing it. Not sure there is demand? Ask for money.

Give your AI a constitution

As assistants gain memory, give yours a rule that protects decisions you have already made:

My current priority is X until [date]. When I ask for ideas, optimize inside that priority. If I start exploring unrelated projects, remind me of my NOT NOW list. Only recommend changing the priority if I bring new evidence that invalidates the original assumptions.

Most digital systems are designed to feed you novelty. Yours can be designed to defend depth.

Optionality without commitment becomes paralysis.