From the head of Claude Code, Boris Cherny
https://www.lennysnewsletter.com/p/head-of-claude-code-what-...
My comment is somewhat tongue in cheek: I know that Claude Code (and AI agents generally) do not mean the end of software engineering altogether, because I have been creating code with AI for a year, written a coding harness, and have non-coder colleagues who use said harness to create software.
But "where people get that from" is precisely from the same guy who is now saying "oh well, if you're going to commit your passwords then well!! you should know better than that!!"
"I think by the end of the year, everyone is going to be a product manager, and everyone codes. The title software engineer is going to start to go away"
Or this:
"Someone has to prompt the Claudes, talk to customers, coordinate with other teams, decide what to build next"
https://x.com/bcherny/status/2022762422302576970
He doesn't mention "configure the git repository so we don't accidentally publish our passwords". That would sound WAY less impressive.
Dario has gone out and said that AI will cause mass unemployment. He said that AI could do 50% of jobs within 1 - 5 years.
OpenAI are much less flagrant with this shit but Anthropic have been banging the drum that their coding agents can fully replace workers for ages, at every available opportunity, and then something as fucking fundamental as publishing passwords shows up and he's all like "Oh well, you shouldn't do that should you that's obvious!!".
It's massive bullshit artistry.
But something like fully configuring a git repo isn't really the concern of a software engineer, per se; their concern is to create code that works, and git is just something to help with managing and passing on their creation. It's the one with the devops role who's really responsible for publication aspects (of course the same person may be doing both creation and publication) and should be ensuring repos don't have things committed which shouldn't be published (or even security if that's a separate role).
> Dario has gone out and said [...]
Different person from Cherny, but also I don't see anyone giving reasonable arguments as to why his prediction may be incorrect, only a lot of outrage. And I'm inclined to agree that AI will be able to do a lot of jobs; it's already on the way with quite a few, for example the software engineer role. I for one have been successfully using Claude Code for said role, while I handle the other aspects, and as it becomes good enough at other things I'll keep handing off more of the software production process to it.