The Productivity Paradox of AI Coding Assistants

I work for a massive corporation that employees tens of thousands of engineers. Actually sitting down and typing code is literally never the bottle neck. I'm lucky if I get to sit down uninterrupted for a couple hours to actually work on something technical at all. The majority of every engineers time is spent in meetings, on documentation, or fighting with either the development, staging, or production environments. Shit I spend more time on trying to prevent management from making a stupid decision then I do on writing code.

When I have gotten time to do something deeply technical, even that time isn't spent on writing code. I've spent hundreds of hours in labs with colleagues debugging complicated systems trying to track down where a single code line change needs to be made. The ratio there is probably 1 line of code written per two dozen hours spent testing and debugging.

AI has little to no value in any of that. If companies really want to save money, their leaving billions on the table in operational inefficiencies, compared to a few thousand dollars in incremental improvements to developer productivity.

The security side of this is going to really hurt a lot of organizations. Increase productivity today because security issues are a risk that might hurt tomorrow.