Raw coding speed isn’t the bottleneck. Alignment is the bottleneck.
That seems to be a zeitgeist-y theme lately. If you’re using AI to code, maybe you’re feeling it. You can code more and faster. And clearly a boatload of other developers are doing that too. But software doesn’t seem to be exploding in quantity or quality broadly. Maybe it’s a little? But if AI is 10✕ing our coding, we’re certainly not seeing software get 10✕ better.
Which is maybe why Andrew Murphy is saying: If you thought the speed of writing code was your problem – you have bigger problems.
Your developers are producing PRs faster than ever. Great. Wonderful. Gold star. Someone get the confetti cannon. Now those PRs hit the review queue, and your reviewers haven’t tripled. Nobody tripled the reviewers. Nobody even thought about the reviewers, because the reviewers weren’t in the vendor’s slide deck.
Or maybe you don’t even get to the “too many PRs” problem because nobody even knows quite what to build. Because you need team alignment to figure that out. You need research. You need stakeholder buy-in. You need a damn plan. And AI isn’t, for the most part, helping with those things. And those things are hard.
Or maybe you are just ripping PRs and your code is evolving rapidly. AI doesn’t help you know… is this the right thing to do? Is it working? Does anybody care? That probably should have been part of the plan, and again, that’s the hard part.
Maybe this is an industry-wide topic right now not just because it’s hitting the community feeling frequency just right, but because there is academic research supporting it. I can’t pretend to understand all that, but I appreciate it’s being looked at with mathematic rigor.
We’re also seeing tooling react to this situation. I think it’s fair to say that AI is increasing the productivity of individuals. But Maggie Appleton pulls out the classic saying: but 9 women can’t have a baby in 1 month. Fasters individuals don’t make a fast company, unless they are perfectly aligned. Maggie showing off new GitHub software that is designed to acknowledge and help with alignment issues. I tend to agree that software itself can evolve to help. Just the fact that AI, in “planning mode” isn’t sharing that plan with a team, is weird, and an easy target to make better.
I also think getting a bunch of humans in alignment is just a thing that takes time. It should be a bottleneck. I’ll forever think of Dave’s “Slow, like brisket.” Some things becomes good because they are done slowly, and it’s OK if software is one of them.