What using AI all day is doing to my life

The good and the ugly of life as a heavy AI user. I have never been more capable. I have also never felt so scattered.

I spend almost every working hour with AI. I use it to research, write, build systems, coordinate other AI agents, make products, create content, and do technical work I couldn’t have imagined doing a few years ago.

I love it. I wake up excited to work almost every day.

Every new model, feature, and release gives me another rabbit hole to disappear into.

As someone who has spent her whole life running away from boredom, I seem to have found the one field where boredom is almost impossible.

And somewhere inside all that possibility, my relationship with time, work, focus, and even my own brain has started to change.

I move faster and know more, yet I always feel there is more to catch up on. I delegate more, yet I often feel mentally drained. I can do things I never could have done alone, yet I’ve started struggling to sit with a book.

This is the part of using AI that I don’t write about enough.

You usually see the systems I build, the work they save, the experiments, and the thousand small problems I solve so you don’t have to solve them yourself. All of that is real. So is the cost of maintaining those systems, supervising their work, and living with the knowledge that there is always more I could be doing.

So I want to give you a broader picture:

• What happens when your capabilities expand faster than your time

• Why building and managing AI systems can be as draining as the work they replace

• How AI is changing my focus, memory, patience, and articulation

• Which new skills it is helping me develop, and which abilities I am trying to protect

• How I decide what to delegate, what stays mine, and why I still wouldn’t go back

This isn’t a warning against AI. It is a personal account of some of the changes I’ve noticed in myself from spending this much time working with AI, every day. The good and the ugly.

More is possible, so less no longer feels sufficient

Before AI, every idea came with a fairly obvious cost.

Could I build this? Probably not without learning to code or hiring someone. Could I research that market, turn the findings into a product, write the launch, and create content for five platforms? Maybe, but it would take weeks or require a team.

Those limits made many decisions for me.

Now I can open my laptop, explain what I want in plain language, and begin. I can learn the technical part while I build. I can coordinate several pieces of a project in parallel. I can turn ideas that once belonged on a distant wish list into active work today.

That freedom is incredible. It also means almost every idea feels actionable, so my backlog never reaches an end.

Every task I complete gives me new ideas while I am working on it. I cross off one item and add three more. By the end of the day, the list may be longer than it was that morning.

And when everything feels possible, everything you don’t do feels like something you chose not to make happen.

Of course, many of those are simply choices between competing priorities. But I also know that some of the tasks I leave untouched deserve more attention than the ones I complete.

The growing list changes how I see my own progress. At times, I can finish more in a day than I ever could before and still end it feeling less accomplished, because my definition of “enough” keeps expanding alongside what AI makes possible.

This is one of the reasons I started building systems to carry more of the work. Without them, far fewer ideas would ever leave my backlog.

I delegated the work and became its manager

We all want the AI system that does the work while we live our lives.

I build those systems for myself and write about them here, spending a ridiculous amount of time making the setup easier for you, because I want to help you skip the trial and error, the broken connections, and the thousand questions between a good idea and a system that works.

Building one requires a huge amount of systems thinking. The deeper you go with AI, the more it forces you to develop that skill.

You have to map how work moves between agents, where human judgment returns, and what happens when information is missing, instructions conflict, or an edge case breaks the logic.

Once the system works, the setup is only the beginning.

Every system needs maintenance. Every agent needs context, feedback, approval, correction, and occasional rescue. Someone has to provide all of that.

That someone is still me.

I feel this most in the system I built for my social channels. I have delegated content creation almost completely to AI agents, using my voice skills, my carousel skill, Amplifiers that research for me, Amplifiers that create my infographics, and everything in between.

Together, they have become one of the best systems I’ve built for my business. I still catch myself staring at the work, amazed that it can autonomously create something this jaw-dropping. It lets me maintain channels I never had time for while keeping the content close to how I think and speak.

But now that my system can create 30 posts for me in a single day, I’m left with 30 posts to review and approve.

I built the system for this scale. I’m not complaining that it works. I’m simply noticing that reading 30 AI-generated posts in a row, while trying to give each one the same level of attention, is far more draining than creating one post at a time with AI beside me.

When I make the post myself, I follow my own thought, shape it as it develops, and notice problems as they appear. When I review a batch, I am searching for something that may be wrong inside language that already looks finished.

Execution has given way to workflow design, orchestration, evaluation, and judgment. I can operate at a scale I could never reach alone, but my brain feels more strained in this role.

My work has become one long series of interrupted thoughts

I have always done my best work by disappearing into one thing. One task, no phone, no one talking beside me. That kind of uninterrupted focus could keep me in a flow state for hours.

AI-heavy work rarely happens that way.

I give one agent a task, move to another session, and answer a question there. A third agent finishes and needs approval. I return to the first output, remember a related issue in another project, leave a note, and come back to logic I now have to reconstruct.

Every open session holds a small piece of my attention and every waiting agent creates another unfinished loop. I can coordinate more work while being less present inside any individual part.

I don’t know whether my memory itself has changed, but it feels as if I can hold far more projects, questions, and unfinished threads in my mind at once.

That capacity helps me coordinate more work, and it may also be why I feel so scattered. There is more in my head at any given moment, and every extra thing adds to the weight.

What upsets me is that the scatteredness follows me outside those sessions too. I have less patience to read for hours without interruption. I can open a long article that interests me and still want to skim.

When I notice it, I force my attention back. If I need to start again ten times before I make it all the way through, I start again ten times.

I don’t know how much of this comes from AI and how much comes from the fragmented digital environment we all live in. I only know that as my work became more AI-assisted, multitasking became a larger part of it, and staying focused on one thing became harder.

The small things I stop doing still count

After hours of designing workflows, answering agents, and reviewing their work, I still have dozens of smaller things to do.

Write a message. Answer emails. Finish a prompt. Each of these looks tiny next to the things taking up most of my brain, so I rush through them or ask AI to make my thoughts coherent.

I notice this a lot in the way I type.

Before, typing was part of my thinking. I would write a sentence, read it as I went, notice that the idea was unclear, and slow down long enough to understand what I was trying to say.

Now I type so much that the process looks more like free-journaling. I type, type, type, put everything down as quickly as it comes, and leave myself with a mess that has to be cleaned up later.

Knowing the tool can do that cleaning makes me less patient with the thought itself. I begin rushing before I have fully constructed it.

That leaves me pulled in two directions. I don’t have enough left in the tank to do everything as carefully as I would like or to think deeply about every task that looks small compared with the rest of my work. Yet those small tasks are also where I practice abilities I want to keep.

And when the shortcut is always open beside me, preserving focus, articulation, patience, and the ability to think something through on my own becomes a choice I have to make again and again, even in seemingly low-stakes situations.

AI is forcing me to explain what I know instinctively

At the same time, AI is developing another kind of articulation in me.

The more I delegate, the less I can rely on “I know it when I see it”. A system cannot follow a judgment that has never left my head. If I want work that feels like mine, I have to make my standards explicit.

This can be maddening. I know what I want, yet I cannot find the words that will make the system understand. By the time I do, I may have spent longer explaining the task than doing it myself.

Deep expertise can make this even harder. The longer you do something, the more of your judgment becomes automatic. You have built years of knowledge into a feeling that arrives in seconds, and now you have to unpack that feeling into words another system can follow.

And every time I have to explain why an output is wrong, I learn something about what I consider right. I am forced to examine my taste, define the principles behind my work, and turn years of instinct into something I can see, reuse, teach, and improve.

So I can feel less articulate when I rush through a sentence and spill the mess in my head into the chat, and more articulate when I explain a standard I used to apply without words. Both are happening at once.

And once I can explain a judgment, another question appears. Do I turn it into instructions for a system or is this one of the judgments I still want to make myself?

The decisions I outsource reveal the ones I care about

It is incredibly tempting to tell AI, “You pick”. And I do it.

My systems make hundreds of small decisions for me. I don’t care deeply about every one. Giving them away preserves energy for the choices that matter.

But when I bring AI into a decision that matters to me, my experience changes. It gives me a clean answer with sensible reasons, and almost every time, something in me pushes back.

Sometimes it agrees too much or challenges the wrong part. Sometimes its suggestion is perfectly reasonable and still feels completely wrong to me.

So I explain what it misunderstood. I add the context I left out. I write three paragraphs about what I believe while trying to prove that its answer missed the point.

By the end, the AI has not made the decision. It has given me a surface to push against until my own decision becomes visible.

Here’s another small, innocent example.

I have more than 100 drafts in Substack, more ideas in notes, and others still in my head. Sometimes I simply don’t feel like writing any of them, so I ask AI what I should write next.

They never excite me. They don’t inspire another thought. They just sound like ideas that exist because a content calendar needs an empty box filled.

Sometimes I close the window. Other times I begin explaining what I was considering instead, then disagree with nearly everything it says about those ideas too.

Still, my thoughts are now outside my head. The weak suggestions help me locate the standard I couldn’t articulate five minutes earlier, because I discover what I think more easily inside a debate than during a monologue in my head.

So the less I care about a decision, the more easily I delegate it. The ones that matter rarely end with AI’s answer. They end with me hearing my own more clearly.

Choosing what stays mine means accepting the trade-offs

Once almost every part of the work can be delegated, I can no longer decide what to hand over based on efficiency alone. I also have to consider what I want to protect and what I am comfortable releasing.

These decisions involve values, quality, joy, and capacity. Keeping too much gives me more work than I can carry. Delegating too much risks giving away parts of the process that make me care about it.

That same social media system is a good example. The content would be better if I created every post myself, no question about that. I would notice more and make more decisions in the moment.

I have still decided that automating it is okay.

The system creates remarkably good work on channels I never had time for. Creating every post myself would take attention away from work I care about more. Without the system, that content would never exist.

Articles are a different choice.

AI can produce a draft in seconds. It can organize messy notes, find missing connections, and help me express a thought that arrived in Romanian but needs to reach you in English. I use all of that.

And I still spend at least eight hours on a single article. Most take several days once I include testing, research, restructuring, and making sure I am giving you something I have earned the right to say.

I don’t want to narrate the news or repeat what someone else discovered. Whether I’m exploring a tool or building a workflow, I want to test it from every angle, understand what works and what fails, and make sure you can get great results from it.

That depth is my contribution. AI can help me express it, but it cannot care about the contribution on my behalf.

So articles are one of the battles I choose. I stay deeply involved because the process brings me joy, and the quality depends on what only I can bring to the work.

But I cannot stay that deeply involved in everything AI helps me create. The volume it makes possible depends on trade-offs, and I have to be able to stand behind the ones I make.

I found work that fits my brain, and it never lets my brain rest

I am thriving here.

I am a generalist by nature. I want to understand writing, products, design, research, marketing, software, psychology, and whatever else catches my attention that week.

Before AI, being interested in everything felt like a lack of direction. Now I can move between fields, bring pieces from one into another, and expand what I can do every time a system breaks.

I love opening a tool I don’t understand and staying with it until I do. I love building systems and turning what I learn into something useful for you. A random question can become a two-day rabbit hole and then a product, an article, or a new way of working.

I wouldn’t return to my pre-AI work life, but I also work more than I ever have.

I can pick up my phone and delegate a task, then put it down. I can approve an agent while I wait in line. I can send one more instruction before bed so something will be ready in the morning.

Each action takes a minute or so. Together, they erase the line around the working day.

The price of never being bored is that I am almost always switched on.

There is always another experiment available, another task that could run in the background, another improvement that feels close enough to start now.

Rest has to compete with possibility. And possibility usually wins.

AI did not make that decision for me. I don’t want to blame a tool for the fact that I protect those boundaries less than I used to. But it did make them much easier to cross.

Before, starting many kinds of work required time, energy, skills I didn’t have, or help from someone else. Now I can set even complex work in motion from anywhere with a single sentence.

I don’t want a clean verdict

Conversations about AI keep getting pulled toward extremes. You are excited about it, so every effect must be positive. You are worried about it, so you must want to reject it.

Neither position describes my life.

I would not give up what AI has made possible for me. But working with it this much has changed some things for the better and others in ways I am still trying to understand.

That tension is where these observations come from. They are incomplete and deeply personal, drawn from my experience as one very heavy user.

For now, I am paying attention. To the decisions I hand over and the ones AI helps me understand more clearly. To the abilities I practice less and the new ones I am developing. I notice when AI makes me rush, and when it forces me to slow down and articulate judgments that used to live only in my head.

That attention shapes what I delegate, what I protect, and where I allow myself to take the shortcut.

And I suspect most of us will spend the next years learning how to hold both. The good and the ugly.

So tell me what this feels like for you. Maybe your experience can continue this article where mine ends.

---

This post is free. Paid subscribers get access to all premium prompts and tools inside Amplifiers (the AI blew my mind MCP), weekly premium articles, all premium resources inside the AI blew my mind Lab, and exclusive partner discounts. Upgrade here.