The Balanced Engineer • Issue #30
Breaking down AI reasoning, mastering async comm and evaluating AI hype with videos and articles from Annie Sexton, Lizzie Matusov, and Korny Sietsma.
The Balanced Engineer Newsletter
Week of July 28, 2025 • Issue # 30
🔧 Technical Excellence
Building skills that last beyond any framework
Why it feels like AI can reason by Annie Sexton (Youtube Video)
Summary: This philosophical video from Annie explained what reasoning is and why LLMs don't necessarily do it, even though it feels like they can. The missing key is the lack of emotions. Humans can't separate their emotions from their thinking. On the other hand, LLMs don't experience emotions at all, so emotions can't impact how they think. She recommends thinking of AI as "very smart parrots". They typically get the things that they repeat correct, but they can also just repeat absolute nonsense!
Why this resonates: I really enjoyed this video and probably wouldn't have watched it if I didn't get to sit next to her while she was editing it. I don't often watch videos in general, but it has convinced me to check out more of Annie's videos. I love this very easy to understand explainer and the philosophical twist on how LLMs work. Since AI has blown up recently it does feel like many of the conversations I have about it border on philosophical (after all, what even is thinking??) so I really enjoyed seeing Annie's take and recommend her videos!
Tags: AI
💬 Communication & Collaboration
Because the best code means nothing if you can't work with humans
What async communication behaviors lead to better outcomes for software engineers by Lizzie Matusov (Research-Driven Engineering Leadership)
Summary: This article described a study of async patterns of communication on team performance. Teams that had short but intense bursts of conversation had better outcomes than those that had evenly spread activity. Communication patterns were better predictors of behavior than skill levels of the teams. The best teams were the ones that were the most synchronized.
Why this resonates: This was one of the first RDEL articles that I read and I love the entire concept behind this newsletter. I am personally pretty data and research-driven, so I loved seeing the approachable summary of a study like this. I'm also on an async and remote-first team that is spread across many time zones, so finding ways for us to better communicate and stay on the same page is incredibly important to me! In particular, reading that the diversity of topics discussed was more important than the overall level of communication was very interesting. Being on the west coast, I only have 1-2 hours of overlap with my European colleagues, which makes synchronous communication pretty tough. But understanding how to best spend that overlapping time to improve team performance seems like a huge win!
Tags: Async communication
🧠 Mental Models & Problem Solving
How we think about complex problems
Clowns to the left of me... by Korny Sietsma (Korny's Blog)
Summary: I actually don't recall how I came across this post, but I did and it was one of my favorite things I read in the last week. Korny describes being "stuck in the middle" between the AI hype squad and AI skeptics, which feels very real right now. One thing he mentioned in the article was the METR study that has been heavily shared recently, reminding us that while it is definitely a good idea to be cautious around how much AI is actually improving dev cycles, the study isn't necessarily an "emperor has no clothes!" moment, due to many limitations of it.
Why this resonates: This article really resonated with me because I feel like I'm in a similar boat. While I have written about having some AI optimism, I do feel like my own viewpoint is also very realistic. There are some wonderful ways that AI has sped up my development process. It is totally one of the coolest developments in the world of software engineering in the recent decades. But also I think it's important to be realistic about its limitations and to constantly stress how incredibly important it is to verify what comes out of AI tools.
Tags: AI
🎯 Try This
One small thing to practice this week
Evaluate your async communication patterns
If you have a lot of asynchronous communication at work, put some extra effort into it this week. As we learned from Lizzie this week, how we communicate matters much more than how often we communicate. Try to increase the overall diversity of what is discussed with your team asynchronously. Here are some ideas of what that might look like:
- Ask questions about what other folks are working on
- Share something interesting that you came across this week (like this newsletter ;) )
- Try designating a specific window for bursty asynchronous communication, when everyone is online and able to unblock each other
What I've Been Building
A quick look at what I've been working on this week
- Overcommitted: Ep. 17 | Empowering Women in Tech with Jennifer Harris - The crew met with Jennifer Harris, a founder and CEO of a multi-million dollar software consultancy. She has been in the tech industry for decades at this point and had wonderful insights on how to actually empower women in tech and how to embrace the changes that come with AI.
- Overcommitted Book Club: Check out the insights from participant's read of the third chapter of Looks Good To Me!
Have comments or questions about this newsletter? Or just want to be internet friends? Send me an email at brittany@balancedengineer.com or reach out on Bluesky or LinkedIn.