I have one of those days today where I feel pretty restless. Yet again, it feels like I’m not really doing much for myself, or getting the kind of rest I would really love to have.
What’s different, though, is that I don’t have this feeling every weekend anymore. I don’t even have it every other weekend anymore. It’s been a while since I felt like this the last time.
And that realization made me realize something else: AI has already done a shift in my life. A subjective one, sure, but still another real shift.
The shift is simple. I finally have more time to be present for my little daughter.
I’m trying to be very intentional about that. Well, no kidding if this is kind of a blog about intentional AI, really.
People say AI is in early stages, I don’t think that’s really true anymore. AI is in early stages of consumer products.
And while we are blazing through development, it feels like a peak that’s never really been seen before. Right now, I feel like I’m living through a technological revolution.
At the same time, it still feels like early days. Products are still kind of raw and unstable.
I’m a product guy, so you don’t need to kid me on what stability means. There’s metrics for that. There’s uptime for that.
I’m obsessed with measuring ROI, and I still don’t really have it with AI.
There’s only a handful of repeatable tasks where AI is actually helping me on its own. And even there, at least once every three months, something kind of falls apart.
That’s really the rawness and the early-days aspect of all this. We still keep seeing massive pricing changes, massive shifts in user bases between competitors, and constant movement overall. At this point, I don’t even think the models themselves are the real bottleneck anymore. I mean, the models are already pretty amazing. Whether you’re an Anthropic fan, an OpenAI fanboy, or trying to run things locally, I think there’s valid use cases for everything nowadays.
And yeah, it’s an amazing place to be.
But back to ROI.
The first thing is that you’re probably not even able to reliably measure ROI.
From my own practice, I do have one specific repeatable task. Every Monday, I used to spend roughly four to six hours planning and setting myself up for success for the coming week.
And it worked. I wasn’t complaining.
But I tried to delegate parts of that process to AI, or at least use AI as a really good sparring partner.
At some point, I realized I was actually saving time and money. I brought the process down to roughly two hours, and it felt great.
I even started tracking it against my hourly rate, so I finally saw some monetary savings.
Nothing massive. Just some nice spare cash. Maybe three figures, occasionally four.
But then I had this realization: what would happen if I stopped using it?
Actually, I didn’t even choose to stop. OpenAI changed some rules, my agent suddenly stopped working, and it happened on a Monday morning when I still had to do the planning.
So I just did it manually.
It took me 90 minutes.
That’s when I realized AI technically didn’t do the work for me.
It helped me refine the process.
And at some point, those savings stopped feeling like they belonged to AI anymore.
Not until the moment I realized:
“Okay. Now I need to bring this down to 30 minutes.”
So I kept wondering at some point: is the added value of AI really just the babysitting part?
You give it a task, or it picks a task, but you still have to give it roughly the same amount of attention. I’d compare it to running a sprint.
Now, I suppose if you define the tasks properly in one go beforehand, that will generally take you less time. But still, AI just keeps trying to execute tasks on its own. And if it’s not sure, it asks for your opinion or your decision, and then it continues.
So I keep wondering whether the real usage and the real benefit is actually something different.
Maybe the real benefit is that AI becomes part of your process of figuring out the exact workflow that works for you as a human being.
And over time, it helps you smooth out the edges.
Then the workflow works better, and suddenly you have another opportunity to improve it again in the next iteration.
So the process kind of goes on and on.
And yeah, right now, while I’m writing this, I’m pretty sure this won’t even be the main benefit forever. This is probably going to change again with a couple of new and more modern approaches.
Once everything settles (and I actually expect that to happen within something like six months) I think we’re going to see an era of amazing ROI.
People already keep claiming they have it, but I’m not really sure they fully understand what that means.
Take people who are obsessed with productivity.
Does AI help with productivity?
Yeah, probably.
But do people understand that productivity isn’t just measured by reducing the time it takes to get tasks done, or by getting more tasks done in the same timeframe?
I don’t really think they account for things like energy levels.
If you’re babysitting an AI for four hours and, technically, you got eight hours of work done... yeah, sure, maybe you did. If you know your stuff, you probably did.
But if after those four hours you feel exhausted and irritated with your life, is that really productivity?
If you are losing relationships because the main relationship you really have is with your AI agent, is that actually the boost you were hoping for?
Is that the ROI you were really looking for?
I mean, you do you, but I doubt it.
Let me just park that statement right here: I really doubt that productivity simply means making more things happen in the same timeframe.
Because if doing that costs you more energy, and if it costs you relationships, then I honestly don’t think that’s more productive at all.
And so, after a few minutes of reading my rambles, now we’re finally getting to the point.
Do I feel like I’m getting more tasks done over one week's sprint?
No. Actually, the data told me otherwise.
I’m getting roughly the same amount of tasks done, roughly the same amount of work done.
Now, there’s a completely different discussion around measuring work by actual impact and different levels of leverage, but I don’t really want to go into that today. Although, side note to myself: I should probably write about that too, because measuring work in absolute numbers is sometimes not really the best thing you can do for yourself.
But anyway, let me get straight to the point.
I’m actually not doing more work.
But I do feel like I have (and here’s the punchline) more energy to just... live. To be in the moment, as cliché as that sounds.
Most importantly, I actually have much more capacity to be present for my daughter.
And that’s one of the KPIs I intentionally set out for myself to observe, uphold, and honestly even enjoy.
What does that actually look like?
My day is really simple.
I get up at 8, prepare breakfast, brew coffee for my partner, feed my little one, get her dressed, change her diaper, and then leave for work.
I work for, I don’t know, something like eight hours.
Then I come back, pick up my little one, play with her until around 8 p.m.-ish, and then go to sleep at 10.
Usually, there wasn’t much more energy left after that. Actually, I used to have basically no energy left after that. Right now, I get to watch some of those YouTube videos I always wanted to watch. I actually have the opportunity to try out new things. And most importantly, I don’t hate myself for this life anymore.
Because it is intense.
I really didn’t enjoy weekends where I would wake up already scared of the fact that I had to figure out what I was going to do with my daughter.
Or just sitting there thinking:
“Oh, there’s another weekend where I’m not going to get the level of rest I really need.”
That was very slowly eating my mind. Now it’s not.
The point where this started changing was when AI started taking over parts of my planned week.
And now I’m looking for ways to make that even smoother. How to make the work less transactional. How to make the work have more impact with the lowest possible amount of human intervention.
I’m by far not there yet.
Right now, this whole thing is raising more questions than answers, which I think is perfectly fine if you’re an explorer like me.
But yeah. That change is big. That change is great.
I enjoy being in this position right now. Would you enjoy this, too?