Generative AI: huge success or massive failure?

One of the interesting things about the discourse around Generative AI (just “AI” from here on) is how very polarized it is. One one side we have people like Steve Yegge writing blog-posts claiming the efficiency benefit of programming with chatbots is “not capped at 100% (2x). It can easily go to 10x, 20x or more in some situations“. And on the other side we have plenty of people claiming it’s all but useless. Not many people are saying it’s sort of OK-ish. It’s quite stress-inducing. You read one of Yegge’s posts and start to worry you’re being left behind. Then you read a study like Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity and see they found that developers expected to be 20% faster with AI but were actually 20% slower. (A followup study linked from the original is more ambiguous.) (Disclosure: I use ChatGPT extensively for trivial stuff, like reminding me of syntax or low-level idioms, but I would never incorporate any code that it…