Dario Amodei: belief

Transcript context

…Even if generalization is weak and you can only do verifiable domains, it’s not clear to me you could automate software engineering in such a world. You are “a software engineer” in some sense, but part of being a software engineer for you involves writing long memos about your grand vision. I don’t think that’s part of the job of SWE. That’s part of the job of the company, not SWE specifically. But SWE does involve design documents and other things like that. The models are already pretty good at writing comments. Again, I’m making much weaker claims here than I believe, to distinguish between two things. We’re already almost there for software engineering. By what metric? There’s one metric which is how many lines of code are written by AI. If you consider other productivity improvements in the history of software engineering, compilers write all the lines of software. There’s a difference between how many lines are written and how big the productivity improvement is. “We’re almost there” meaning… How big is the productivity improvement, not just how many lines are written by AI?…

Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.