My AI Posted "Paralysis Is Optional" to My Instagram

My AI Posted "Paralysis Is Optional" to My Instagram

Jana was in bed. Phone dinged. She almost never picks it up, and she's never on social media, but she did this time.

Notification from my account. Three words.

Paralysis is optional.

She sat straight up. Actually yelled my name, she says it was like a sitcom, and asked me what in the world I had posted, because she has a lot of friends in wheelchairs and one of them is her.

Thank God she picked up the phone. Nobody else was going to tell me. It could have sat there for weeks.

I had posted nothing. That's the whole problem.

What happened

I'd just finished writing and publishing a book.

Which, if you've done it, you know what the tail end of that feels like. You're wrung out. The book is done and now there's a whole second job called telling people the book exists, and you have approximately nothing left.

So I took clips from the manuscript, fed them to AI, asked for social posts, and pushed the output into a tool that posts automatically.

I didn't read them.

The phrase in the book is analysis paralysis. That's a real thing, it's what the book is about, and the distinction between that and paralysis matters enormously to dozens of people I've had dinner with and millions more I'd like to.

AI dropped one word. Then the software did exactly what I'd told it to do.

The tool was Zeely.ai, and I don't recommend it.

The post is on me. I skipped the review, and no product is responsible for that.

The rest of it is theirs. I cancelled the same night I tried it. I was billed weekly for the next six weeks while I kept trying to cancel. That's my experience with them and I'd want somebody to tell me before I signed up.

The part that makes it worse

The book is about decisions. Specifically about decision fatigue, and about how the quality of your choices falls off at the end of a day.

I wrote that book. And then, at the end of a long day, mentally spent, I made a decision to hand off judgment to software and not look at what it produced.

I'd like to report that the irony occurred to me before Jana yelled. It did not.

Why that particular phrase, in our particular feed

For the last two years Jana and I have been building relationships in the adaptive sports world. Team USA water skiing. A nonprofit we're getting off the ground.

And ballroom, which needs a sentence of explanation. We compete in Fred Astaire events and international adaptive competitions, and every other couple out there is either pro with pro or pro with amateur. We're the only amateur-amateur pair we've come across. Two people who aren't professionals, dancing together.

That's on purpose. Everybody should get to dance with the person they want to dance with. I think a soul that's actually satisfied wants to dance, and I don't think it much cares whether either of you is any good at it.

We win sometimes, which I mention because it undercuts the polite assumption that the amateur couple is out there to be encouraging.

We lose more than we win. But we get a standing ovation most times we're on that floor, and it isn't for our technique. It's because we are visibly having the time of our lives out there, and a room can feel that from forty feet away.

And last year we spent a week in Colorado Springs watching her 2004 Paralympic team get inducted into the U.S. Olympic and Paralympic Hall of Fame. Jana's been on three Paralympic teams. That week was a room full of people who have spent their lives making the case that a wheelchair is not the end of a sentence.

That's who follows us now. A lot of people with disabilities, who let us into their community, and who we've been genuinely trying to serve.

And my account told them paralysis is a choice.

Not a typo problem. A trust problem, in the one community where we had the least margin for looking out of touch.

We use AI constantly and I'm not going to pretend otherwise

Jana and I were learning this stuff back when the prompts were three pages long and you had to teach the thing to sound like you.

I use it to get a brain dump organized, which is most of what it's good for. Get the mess out of your head, let something structure it, then go fill in the spaces yourself. That's how both books got written and I'd do it again tomorrow.

James Altucher has a phrase for the other thing we use it for, which is idea sex. Ask for a few different angles on something, then take two that don't obviously go together and see what happens when you put them next to each other. That works.

So this isn't an article about how AI is bad.

The line is somewhere else.

Using it to think differently	Letting it think for you
You bring the raw material and it organizes.	You ask what it thinks before you've thought anything.
You ask for angles you wouldn't have reached.	You take the first answer because it sounds right.
The output goes to a person before it goes anywhere.	The output goes straight to the world with your name on it.
You still have the argument afterward, out loud, with somebody.	Nobody had the argument.

The muscle question

Here's what Jana's actually worried about, and it isn't the Instagram post.

She can't remember phone numbers anymore. When she was a kid she could tell you every friend's number, the movie theater, all of it. Now if you give her more than three digits at a time she has to stop you.

I can't navigate a city I've lived in for twenty years without GPS.

Neither of us decided to lose those. They just went, quietly, because something else was doing the work.

So she's watching herself go to chat before forming an opinion. Ask it first, see what it says, then build her own view somewhere near that. And she said something that stuck with me, which is that if she's doing it, other people are doing it.

Her creativity is the thing she's proudest of. She'd like to keep it.

The one thing we do that I think actually matters

We take it back to people.

Jana will ask AI for ten table topics for Toastmasters, which used to take her an hour. Then we take those questions into a room of actual humans and talk about them, and that conversation is nothing like the list.

Same with everything else. We have three or four real conversations a day, usually at the table, about whatever we're working through. The tool speeds up the front end. The thinking still happens between two people in a kitchen.

That's the exercise. If the output goes straight from the machine to the world without passing through a person, you've skipped the part where you were supposed to be involved.

Which is precisely what I did with the Instagram post.

The rule now

Nothing publishes without a human reading it first. Not a sample. All of it.

And nothing gets approved at the end of a day when I'm fried. That's the one I'll actually have to work at, because the end of the day is exactly when handing it off feels most reasonable.

If you take one thing from this: the automation isn't the risk. The risk is the moment you decide you're too tired to check, and that moment arrives at the same time every day.

The Living Lucky® Practice

Exercise: find one place where software publishes or sends something with your name on it without a person reading it. Put a person in that loop today.

Reflection question: what am I asking a machine before I've formed my own opinion?

Journal prompt: what could I do easily ten years ago that I've quietly stopped being able to do?

Action: whatever you were about to approve tonight, approve it tomorrow morning instead.

Frequently asked questions

What's the risk of automated social posting?

Fluent output with no context behind it. AI writes clean sentences and has no idea who's reading them, so a phrase that's fine in one room is damaging in another. Dropping one word out of analysis paralysis is the whole distance between a concept and an insult.

How do I use AI without losing my own voice?

Bring your own material first. Write the raw thoughts, then use the tool to organize and expand them. The failure mode is asking it what to think before you've thought anything, which is a different activity entirely.

Is cognitive offloading real?

Ask yourself how many phone numbers you know, or whether you could get across your own city without navigation. Whether that's a problem depends on what you're offloading. Directions, probably fine. Your own judgment about what to say in public, less fine.

What guardrail actually works?

A human reads everything before it goes out, and nothing gets approved when you're depleted. Both are simple and both get skipped for the same reason, which is that skipping them feels harmless right up until it isn't.

What do I do if something already went out?

Say what happened plainly to the people affected, fix it, and don't hide behind the software. The tool did what it was configured to do. Somebody configured it.

Both books are out

The Decision Matrix and The Quiet Space are both available now, on Amazon and at livinglucky.com.

The Decision Matrix is the one about decision fatigue. The one I wrote, published, and then immediately demonstrated by making a bad decision at eleven at night.

I'm keeping that story. It's the best example in the book and it didn't make the book.

We went deeper on this on the Living Lucky® Podcast, Season 11, Episode 35.

We Create Our Luck®.