The shifting sands

Posted by: stak
Tags:
Posted on: 2026-08-23 10:26:19

Wow it's been a while since my last post. Since then AI capability and adoption has really started going vertical, at least in the world of software engineering. One of the things I've noticed with increasing AI adoption is a decline in collaboration, which is interesting. The main reason for this, I think, is people feel more empowered to do things, which is great in many contexts but not so great in others. Here's one specific example scenario.

Say you have a person Alice working on a product, and they need some new feature from the platform layer. Previously Alice would go talk to Bob (the owner of the platform component) about the problem she's trying to solve and they would collaborate to find a good solution. With AI tooling, Alice can just ask an agent to go and implement the feature in the platform directly. Alice unblocks her own work quickly, but the platform change still has to be maintained by Bob.

In the best case, Bob has enough context written down so Alice's agent builds a good solution. In the worst case Bob isn't even aware of the change and is still on the hook for maintaining it, regardless of the quality of the change. In the common case Bob is asked to review the change and can still collaborate with Alice to get it in line with the platform's standards - but the frame has shifted from "let's solve this problem together" to "my agent built this, please review what might be slop".

I have been on both ends of this so can appreciate both Alice and Bob's perspectives. It's not inherently better or worse, but really depends on the organizational structure and what is being optimized for. It's just one of many examples of how trade-offs are shifting with AI tooling, and highlights how organizations have to adjust and adapt in order to make the best of it. Org leaders who fail to account for shifts like these will find their orgs no longer run as smoothly as they used to.

[ Add a new comment ]