Sure, there have always been AI shills. That’s not the issue. The point I’m making is that people who used to dismiss AI are, as of several months ago, saying it has become very useful.
Three of the four people I quoted were specifically extremely vocal opponents (“haters”, if you will) of AI since the beginning. DHH especially is referring to his previous resistance to AI, and alluding to how the situation changed because the models improved to the point where they are beyond a certain threshold now.
I can tell you my own company has benefited from Opus 4.5 very directly. I definitely could not have done that project with its predecessor models.
Before Opus 4.5, the only similar rate of improvement (outside of coding) was GPT 4o. Its jump in intelligence was noticed by everyone who used GPT models professionally.
