> Rendering of an HTML page vs MD in an AI IDE or even tokens MD is usually converted to HTML and then rendered by a webview, which means MD is more expensive.

> Rendering of an HTML page vs MD in an AI IDE or even tokens MD is usually converted to HTML and then rendered by a webview, which means MD is more expensive. > compounding costs for a communication doc may be billions in a year Lol absolutely not. It would be surprising if the "compounding costs for a communication doc" reached $1 a year. I don't think you understand how efficient computers are and how optimized HTML rendering has become. > As tokens needs to conserved Not really. No large org is really optimizing for fewer tokens yet.