Reading the Naysayers

One thing I like to do with every tech wave is read what the skeptics are saying. IMHO it's a healthy way to counteract all the hype. David Gerard did a great job writing about crypto and like many of the crypto bros has "pivoted to AI". When headlines said MSFT code base is now 30% written by AI my reaction was bullshit. MSFT has hundreds of millions of lines of code and AI or not there is no way in two short years they have replaced or added new code to reach 30% of their total code base. My experiments with writing code with AI assist result in directionally correct but mostly wrong code and arguably more time spent debugging, plus there is an associated loss of learning because of the AI crutch.

I thought I'd share David's last piece which breaks my gut BS! reaction down. He also asks a great question, "If AI is so good at coding … where are the open source contributions?". To me that's an objective measure of the true effectiveness of AI coding immune to the hype. Even if PR's aren't the whole story, assuming widespread and effective uptake we should see the acceleration of open source projects as the AI assisted programmer should be more productive. There is no evidence yet.