Chain AI models from different vendors to catch blind spots
Don’t run AI models in parallel. Chain them in series, across vendors. The output of one becomes the input to a critique from another. Different training data means different blind spots.
Data modelling and DAX
Using VS Code and GitHub Copilot as a personal productivity environment.
Don’t run AI models in parallel. Chain them in series, across vendors. The output of one becomes the input to a critique from another. Different training data means different blind spots.
Running multiple topic-based chat sessions in parallel beats one monolithic chat. Tighter focus, faster context, easier to review weeks later.
I use VS Code as a personal productivity environment, not a coding tool. Customer notes, briefings, emails, research, diaries, presentations, planning. Almost no actual code. This is the first post in a series writing down the setup.