TLDR; AI productivity is not always real it just feels like it because of lower mental effort
I recently finished my project its a dashboard builder and it took me approx. 7 months to launch it in beta. This is despite using ai and not writing single line of code. It got me thinking about productivity with ai claims on social media.
If you spend any time on social media these days, you've seen it, post after post about how AI is making everyone "10x more productive." This isn't just casual talk, it's often tied to big predictions about job losses and the future of work.
Using AI really felt more productive. But then I wanted to put some concrete numbers to that feeling. I needed to know if I was actually saving time, or if something else was happening.
So, I decided to run a simple experiment.
Human vs. AI: The 30-Minute Bout
I picked one of the task from my current task list that required some real thinking to solve. My plan was simple, do it once on my own, and then do it again with AI, and compare the times.
Round 1: Just Me First up, I did it myself. No help, no AI. It took focus and some real mental effort to get it done. After a bit of work, I had a solid solution.
Time taken: About 30 minutes.
Round 2: Me + AI Then, I opened a new window and gave the exact same task to an AI. It didn't give me the perfect answer on the first try. The first attempt was a bit off. The second was closer, but still not quite there. It took several rounds of me refining the prompt and guiding the AI to fix its mistakes. This is despite solving the issue earlier so i had advantage of knowing the exact solution i wanted.
And the surprising part? When I added up all the time I spent prompting and waiting for the AI to finally get it right, the total time was…
Time taken: Also about 30 minutes.
The Sixth Sense: I See the Truth (About Cognitive Load)
The time was the same, so why did the AI version feel so much more productive?
That's when it hit me. It was about Cognitive Load.
Cognitive load is just a fancy term for how much mental effort you have to use to do something. Here’s how it played out in my experiment:
- Without AI: My cognitive load was high. My brain was doing all the heavy lifting, the problem-solving, the creating, and the editing. It was effective, but it was draining.
- With AI: My cognitive load was low. My role changed completely. I wasn't doing the hard thinking anymore. I was just guiding the AI and telling it what to fix. It was much easier and less of a mental struggle.
That feeling of ease is what I was mistaking for speed.
A Tale of Two Productivities
Now, I'm not saying the AI productivity boost is a myth. It's very real, but I think it comes in two different forms.
Type 1: The Real Time-Savers This is when you give AI repetitive or chore-like tasks. Things like summarizing a long document, writing simple boilerplate code, or searching for specific information. In these cases, AI is an automation tool. It does the job much faster than a human can, and the time savings are real and easy to measure. This is a true productivity boost.
Type 2: The "Felt" Productivity This is what happened in my experiment with a more complex, creative task. The time taken might be the same, or sometimes even longer. The "boost" here isn't about speed, it's about the lower cognitive load. The task feels easier, so we mistake that feeling for being more efficient. Both are valuable, but they are not the same thing.
Free Your Mind (and Measure Your Output)
The "feeling" of productivity with AI is powerful, but it's important to understand where it comes from. Is it making you faster, or is it just making the work feel easier?
I'd encourage you to become an experimenter, too. Next time you use AI for a task, don't just go by the feeling. Ask yourself:
- Did this task actually take less time than if I did it myself? (This is a real time-saving gain.)
- Or did it just feel less difficult and mentally tiring? (This is a cognitive load gain.)
Being aware of this difference helps you use AI more intentionally. You can then see if you're getting a real return on your time or just making a difficult task feel a little less painful. Both are great outcomes, but knowing which is which is the key.
I am now planning to write an article where I'm going to do a full post-mortem of my current project that was built almost entirely with AI. It was a medium-complexity project, and I didn't write a single line of code by hand. So why did it take me seven months to launch? Luckily i have both start/finish time of tasks in my org mode journal and also in histroy of AI tools (chatgpt, claude, gemini) i just need some free time to dig deeper.
Stay tuned for that.