When to Train an AI Model and When to Write Code
There's a rush to use AI for almost every problem at the moment. That enthusiasm can make it easy to reach for a model before asking whether a straightforward rule would solve the problem more reliabl

There's a rush to use AI for almost every problem at the moment. That enthusiasm can make it easy to reach for a model before asking whether a straightforward rule would solve the problem more reliabl

A lookup table looks harmless until three modules start depending on it for different reasons. Suppose an underwriting system has a list of countries. Submission intake uses it to normalise addresses.

Most .NET developers recognise an N+1 database query when they see one. AI features can recreate the same failure at a more expensive boundary. An application loads 300 support tickets, documents or p

An authenticated user calls your API with a valid access token. The token contains the right role and the endpoint requires authorisation. Everything appears secure. The user then changes /claims/8d61

Sending text to an AI model as an image sounds like it should cost more. The application has to render the text, encode a PNG and send a much larger payload over the network. Yet for the right kind of

IRC is old, plain, and surprisingly useful as a teaching tool. You open a TCP connection. You send text commands. The server sends text lines back. The connection stays open. If the server sends PING,
