Senior Engineer: I'm tired of talking about AI unless you actually do something valuable and cool with it.

A engineer who uses AI daily admits: AI has indeed boosted productivity, but endless tool worship has left him exhausted. The community has shifted from sharing cool projects to everyone competing over whose AI workflow is trendier. This article is based on Jake Saunders’ piece “Is anybody else bored of talking about AI?”, edited, organized, and translated by Dongqu.
(Previous summary: Claude Code ultimate quick reference: hotkeys, slash commands, skills, agents, MCP full operation tips)
(Additional background: NVIDIA’s Jensen Huang calls DLLS 5 “the ChatGPT revolution for graphics,” only to be mocked by artists: just applying beauty filters)

Although it might sound rebellious, I really am getting a bit tired of talking about AI. I understand, AI is truly incredible. I use it every day, and it has completely changed my workflow.

Recently, I took on a new role handling the highly challenging Web Scale (hey, do you remember the term Web Scale?) field. Thanks to AI, I was able to achieve a leap from 0 to 1 in productivity within weeks.

That said, all this is starting to feel a bit… monotonous. I don’t want to deny how astonishing this transformation is, but in daily life, I seem to have run out of fresh things to say about it. Worse, AI has almost completely monopolized the attention of the online communities I follow.

The most frequented Hacker News used to be full of interesting projects and unresolved problems, but now it seems to have devolved into three similar Claude coding workflows or yet another article about how you can let OpenClaw pet cats while gaming, just to free up more time for… configuring AI tools. It all feels like going in circles.

Kagi Small Web Is Another Victim

Here’s a challenge: open it up and click “Next Page” 20 times. How many of those articles are related to AI?

Before you see me as just “an old man shouting at clouds,” please understand my perspective. Back in the good old days (2023), before anyone who could run Claude terminal was called an “AI engineer,” the most popular title was “Product Engineer.”

The core idea was that engineers should stop obsessing over code and focus on the value of the products they deliver. I loved this idea; it made a lot of sense to me. But now, it seems we’ve regressed. We no longer focus on code but on this “overgrown auto-completion tool”—which is just making the simplest parts of an engineer’s daily work easier.

It’s like I walk into a woodworking forum, and instead of showcasing their tables, everyone is posting about the hammers they use. And most of these hammers are basically the same, used in similar ways, yet people are still shouting the same nonsense at each other.

Even worse, now the bosses are convinced too. My managers used to care little about database tech, IDEs, or JavaScript frameworks—they just wanted features shipped so they could sell. But now, management seems to recklessly dive into implementation details.

I bet most people’s KPIs this year include some kind of “more AI usage” company policy. Management involvement in the software development lifecycle (SDLC) isn’t new; DORA metrics have been around for years. But historically, it’s always been about “output”—faster deployments, shorter response times. Now, we’re measuring how many tokens each developer uses, which is as pointless as measuring “lines of code” back in the day.

What I want to say (besides just venting) is: talk to me more about the cool stuff you’re building, not the tools you’re using to build it. Don’t forget, the essence of programming, like any craft, is about creating something valuable for someone—even if that someone is just yourself.

…Oh, and I’m fully aware of how ironic it is to write a rant about AI. Sorry.