I've been feeling pretty worn out lately from the constant weight of AI being shoved down my throat every time I open VS code, log in to social media, view an ad on TV, or do any of the other countless regular digital activities that have suddenly become infested with the “magic AI icon” butting in to try and give me something I didn’t ask for. I was planning on composing a piece about why I'm so damn bothered by it all. But, while reading Ed Zitron's comprehensive guide to dealing with AI boosters (as usual from Zitron, it's a long read but a great one), I came across another great post by engineer Colton Voege, titled No, AI is not Making Engineers 10x as Productive: Curing Your AI 10x Engineer Imposter Syndrome. Honestly, Voege already lays out most of my issues pretty well, so I would encourage people to read his piece. So I therefore want to lay out a few of the points that have encouraged me to feel at least slightly better about my life as a software engineer in 2025.
Voege makes important arguments against the misconception that LLMs are making engineers 10x more productive: not only do agents still struggle with the context needed to contribute to large codebases, but the things that really take time in software (ideation, planning, reviewing, bugfixes, deployments) can't magically become 10x faster just because an LLM can churn out a bunch of code. He emphasizes that the promotion of this 10x productivity myth, at least in some small part, is coming from bosses wishing to make their employees feel precarious and replaceable. I am grateful that my boss has not done this explicitly, but I also have to admit that I have certainly been feeling more precarious lately. Do I have to go all-in on generative AI coding just to prevent myself from being laid off when someone higher up decides that my team can be replaced by thousands of AI agents? Would I be able to find another job if and when that happens, given the increase in tech layoffs since 2023? I appreciate Voege reminding me that these fears are, well, exactly the point for these people. As much as they are saying it because they want it to happen or think it will, they are saying it to make me afraid that it will happen. And while I certainly don't want to get laid off, I suspect that if I do, it would happen regardless of how many tokens I spend talking to an AI agent—someone who thinks my career can be equated to $270 worth of monthly subscriptions is already looking for a reason to get rid of me.
I especially love that Voege touches on how the LLM craze has affected morale. He describes his fear of becoming obsolete because he doesn’t particularly enjoy coding with generative AI, exactly how I have been feeling lately:
[AI coders] were using agentic AI. They were using thinking models that surfed the internet, ran tests, and corrected their own mistakes. Sure I popped into a chat window here and there and asked it to write some code, then promptly discarded most of the output once I got the idea that I needed. But these engineers were letting Claude fully take the wheel and had agents ripping 5 PRs for them while they made morning coffee. Was I becoming a dinosaur, an old man yelling at cloud?
Part of what made me feel so anxious was that it was entirely possible AI changed without me knowing it because I didn't use AI very much. Because I didn't like using AI that much. Reviewing code is a vastly less enjoyable process than writing it [my emphasis]. Had my stubborn desire to enjoy coding set me up to be left behind?
I have honestly wondered something very similar every time someone raves to me about how great the latest LLM from OpenAI or Google or Anthropic is. While code review is an essential part of my job, it’s not the part that most inspires me–and when it does provide inspiration, it’s because I’m growing from the back-and-forth process of teaching and learning with another human engineer. Can I grow in similar fashion from back-and-forth sessions with Claude, using the agent for a sort of enhanced rubber-duck session? Sort of, but it’s just not the same. When I’m speaking with a human colleague, we’re using our different experiences and shared contextual knowledge to solve problems and grow together, while making the products we’re building as great as possible. These products, which are intended to be used by humans, benefit from having humans who understand the larger context, need, and purpose behind the code. Conversely, when I’m typing to an agent, I’m feeding a beast that devours data indiscriminately and spits back whatever seems most plausible. While AI boosters defend this unreliability by commenting that human coders can also produce mediocre code, they fail to consider how the “traditional” approach—writing something that is mediocre, learning how to do it better through conversations with fellow engineers, refactoring based on those conversations, and then learning other new lessons during the refactor—may help engineers more.
Voege not only shows how LLM-assisted coding cannot make everything in software development 10x faster, but he also dares to voice something amazing that I’ve been too timid to say too loud: it’s OK to go a little slower if it helps you enjoy what you’re doing:
It's okay to sacrifice some productivity to make work enjoyable. More than okay, it's essential in our field. If you force yourself to work in a way you hate, you're just going to burn out. Only so much of coding is writing code, the rest is solving problems, doing system design, reasoning about abstractions, and interfacing with other humans. You are better at all those things when you feel good. It's okay to feel pride in your work and appreciate the craft. Over the long term your codebase will benefit from it.
I needed someone to voice this. I know we all have bosses and deadlines and stakeholders and end users demanding products to be shipped as quickly as possible. But it's also a valuable truism that sometimes to go fast, it's better to slow down—both for the quality of your product’s codebase and for the morale of those working on it. Viewing the posts of hustle culture cultists on LinkedIn will make you afraid to even think this, but there's more to good software than delivering things quickly.
LLMs aren't going anywhere, and I doubt the hype around them will either. I'm sure I will continue to feel frustration and inadequacy over the rhetoric, but I'm very grateful to know that I can find some solidarity with others who feel similarly. I don't know the future, maybe some day I'll have an LLM agent churning out PRs to do all my work, and I'll just be an AI babysitter (or out on the street). But I'm glad that day hasn't come yet (and, with luck, it never will!). I may not have become 10x more productive with LLMs, but I'm a better engineer today than I was six months ago. And I expect that if I can maintain my interest in software without getting burned out, I'll be able to say the same thing six months from now.