Are we wasting that much time?
Hey!
Welcome back to another week of musings.
We're almost halfway through July, and now with half of the year behind us. Have you done any introspection on what happened? What do you want for the next 6 months?
Was this forwarded to you? You can subscribe here!
Things I enjoyed in the past week
- A podcast from The Pragmatic Engineer: What is a Principal Engineer at Amazon? With Steve Huynh, I've heard most of these comments in some form or another, but it's always good to hear from experience rather than general direction.
- TBM 367: The Wicked "Big Picture" Loop, a good post that talks about why talking only in "Big Picture" terms loses a lot of fidelity, which ends up burning teams.
With the recent push for the adoption of AI tooling for development, I feel that leadership believes they can either squeeze another engineer out of everyone or reduce the workforce even further by having "AI do the work."
Where is all the developer's time going? If you've been in the industry for a while, you'll know (maybe anecdotally?) that coding is a small portion of the time. Other activities that occur include code reviews, story grooming, sprint retrospectives, and quality assurance (QA).
This, excluding meetings outside of "agile ceremonies", if your company holds those. Meeting with Product Managers, other Engineer Managers, and other staff peers.
Productivity Boost?
If you've been using AI-related tools for development, you'll know that, well, they might be getting better, but not really at replacing us at all. Additionally, they may not work well with all cases; they can create code but struggle to review it. Even with agentic workflows, they may not be able to replace a QA engineer.
As I've been interested in providing a point of view on AI tooling, I've been following developments such as the DX AI Measurement Framework, which offers dimensions and maturity to measure the impact of AI tools.
Lately, to better understand how the tooling affects me, I've been using all the available tools at my disposal, such as Cursor, Claude Code, and GH Copilot.
Being honest, I'm not saving time, like I don't open PRs faster, if that's the measurement of speed used. But the tools have helped me get in the zone faster, since I don't have large blocks of time to code, having an idea, writing it down as text and getting some code back is great, I generally correct or tweak it in some way to our coding standards, and I get to do that task of the day maybe.
The real time sinks
But what about other things that take up my day? System design, product meetings, and providing subject matter expertise are not being replaced at the moment. I see people using AI for summarizing meetings or generating notes from recordings, but those efforts often fail to yield any new insights, sadly.
If anything, we have more text to review, because people refuse to have a detailed meeting agenda, and action items as a conclusion. We're automating the wrong thing, as the meetings themselves are a side-effect of a culture that needs them to move things forward, so they're just events for people talking to each other.
Your turn!
Where is your organization in its AI adoption journey? Is the leadership team believing they can replace all engineers? Are your PMs creating branches with proof of concepts? Let me know your thoughts by replying to this email!
Happy coding!
website | twitter | github | linkedin | mastodon | among others