One mistake with AI productivity is measuring the speed of generating an output while ignoring the time spent fixing it.
A better experiment is to measure the whole workflow.
For example, take a research task:
Without AI
Research → read documents → take notes → organize information → create summary
With AI assistance
Collect sources → analyze with AI → verify important points → organize → create summary
The second workflow isn't automatically better.
The real question is:
Did the AI reduce the total amount of useful work required?
I've been experimenting with Google's AI tools from this perspective.
NotebookLM is interesting for source-based research, Google Vids for turning information into video, and MusicFX for creative experimentation.
The tools are different, but the principle is the same: start with the problem, then choose the tool.
I'm also working through a structured Google AI learning resource covering these tools. I'm finding the hands-on approach more useful than simply collecting AI tool names.
Top comments (0)