How I Use AI Every Day

(Tip: I’m not in marketing)

I don’t know about you, but I am settling into this feeling that AI, while powerful, is also kind of looping over the same territory. I find myself turning to AI to answer questions instead of Google search.

I think most people who use AI either use it as a search replacement or they ask it to write a document for them that they themselves have low confidence in being able to generate themselves.

It seems to me the primary use of AI for the vast majority is:

1. Replaced Google with better, more relevant answers.
2. Writing well is hard, so do it for me.
3. Make a pretty/cool/interesting picture for me.

Okay, that’s not fair, is it? There are a few more.

1. Note takers: Transcribing meetings and capturing what everyone is saying.
2. Video creation tools are powerful.

I have seen some cutting-edge applications, like Eleven Labs and Go High Level voice agents. And that brings me to the next giant category of AI use: marketing. Since marketing seems to be about persuasion with words and pictures, it makes a ton of sense for AI to thrive in that arena. Makes sense, right?

But what if… you are not in marketing? What if you are running a company? What if you want to use AI to be more personally productive? What are some ways you can use AI to increase your own effectiveness?

That is what I want to write about. I want to share, a little at a time, how I have been implementing AI for my own personal workflow. And yes, this will definitely include some content about how I use my digital second brain.

At present, I am thinking about covering:

1. What I use and why (my three tools: Claude, ChatGPT, and Gemini)
2. The environments (mobile, desktop, command line)
3. The types of problems I solve and how I solve them.
4. Tips and tricks for being more productive.
5. Model Context Protocol techniques.
6. Vibe Coding
7. Power Tools:
   1. GitHub - managing personal projects, backups, and revisions.
   2. Visual Studio Code
   3. Marta (or other dual-pane file managers)
   4. Markdown
   5. Obsidian
   6. Calibre
   7. Anna’s Archive
   8. Automator
   9. AppleScript
   10. Python
   11. Node.js

Kinds of Problems

What kinds of problems am I solving with this “stack” or workflow?

Here are a few examples:

• Take a rambling description of a topic and turn it into a well-organized presentation, with thematic images I can use for a keynote.
• Convert 99 screenshots of recipes into 73 linked text recipes formatted for my Obsidian database and published so I can share with my mom. (Including stitching together recipes that span more than one screenshot.)
• Performing a financial analysis on my operating business locations, as well as helping me build a forecast model for a cash-based business (non-trivial problem).
• Working with complex concepts and ideas across multiple knowledge domains (psychology and neuroscience) to find interesting insights and connections which have a practical application.
• Story coaching. How to take my highly conceptual presentations and bring them alive with real stories for greater stickiness and impact.
• Using multiple models to cross-check work and minimize sycophancy.
• Using multiple models to fractionate large projects into more manageable pieces so quality can be checked before running off the rails.
• Techniques for both 1. Preserving context across large project runs and multiple sessions and 2. Isolating projects so as not to cross-contaminate work.

If any of these sound interesting to you, then. You might enjoy this series. If this is not interesting to you, then just have your favorite LLM reduce this to a few trivial bullet points, scan that, and then move on with your life, or ignore it altogether (which amounts to pretty much the same thing).

Note: You can also find this article on medium.com: scottnovis.medium.com/how-i-use…