Megan Scott

We had an AI agent refactor our database submission tables to fix load times. The pull request looked super clean and performance metrics improved instantly so everything seemed fine at first.

A few days later we realized it stripped out essential audit columns because nothing in the database constraints explicitly flagged them as required. It optimized strictly for raw execution speed over domain context.

We've been fully remote for over two years now.

While daily coding output felt high we kept hitting a wall where features got shipped with small logic gaps that no one caught during planning.

We recently started getting the team together in person just once a month not for long presentations but just to sit and work in the same room.

Having that physical proximity dropped our rework rate significantly.

It s so annoying when you feed an error to an agent it fixes one thing, breaks another, and then reverts right back to the first issue two turns later.

Once the chat gets flooded with raw logs it feels like it completely forgets the big picture and just starts guessing.

Do you just wipe the chat completely at that point or is there a prompt trick that actually gets it to stop and rethink?