The SDLC just got bigger
Engineering hasn't gone away. It's just different: more complex than it's ever been, and simpler too.
Anthropic’s AI-SDLC playbook is, in my opinion, one of the best-documented simple approaches to getting that AI-native feeling. Have a read.
Moving to it is still hard for most teams, though, and it takes real effort. Why?
It stopped being an engineering problem
Honestly, I think a lot of people are struggling with one thing. Two or three years ago the development life cycle was treated as an engineering problem, with product sometimes invited along. Now it’s much bigger than that, and it’s hard to make peace with.
The guardrails come first
What stands out to me is that guardrails and governance are built in from the start. Take a look at the playbook and you’ll see managed settings and control through hooks.
Normally that sort of thing gets picked up by an IT or security group after the fact. Engineers don’t give it much attention, and sometimes even the security people skip past it.
But there it is, part of the SDLC.
Same job, new shape
The hardest part of AI, I’m finding, is the transformation. Engineering hasn’t gone away; it’s just different. It’s more complex than it’s ever been, and at the same time simpler than it’s ever been.