←
Kostis Kapelonis1,434 posts
Kostis Kapelonis avatar

Kostis Kapelonis

@codepipes

Tweeting about Java, Groovy, CI/CD, Unit testing, containers/Kubernetes and anything related to automation and deployments.

◷ Joined February 2015▧ 105 Photos and videos
Dan Garfield avatar
⟳ Kostis Kapelonis retweeted
Dan Garfield@todaywasawesome·Nov 3

Do: Use AI to build repeatable tools and processes. Don't: Replace repeatable tools and processes with AI.

Anton Arhipov avatar
⟳ Kostis Kapelonis retweeted
Anton Arhipov@antonarhipov·Oct 7

IMO, ArchUnit is becoming critically important tool in the era of AI agents

A new episode of IJ Tech Talks is out! 🎙️ Architecture as code. Roland Weisleder joins IntelliJ IDEA Tech Talks to turn architecture rules into ArchUnit tests, catch cycles before they catch you, and keep docs and code from drifting apart. 👉️ youtu.be/hEnSy8vlK10 👈️

0:53
argoproj avatar
⟳ Kostis Kapelonis retweeted
argoproj@argoproj·Sep 23

The Gateway API plugin for Argo Rollouts has passed 200k downloads! 🎉 This plugin allows you to define rollouts once, agnostic to almost any ingress or gateway provider! github.com/argoproj-labs/rol…

Post attachment
Alex avatar
⟳ Kostis Kapelonis retweeted
Alex@AlexJonesax·Sep 17

Average github interaction now: Me->Agent->PR<-Your Agent<-You

Siva avatar
⟳ Kostis Kapelonis retweeted
Siva@sivalabs·Sep 15

No amount of agent skills and harness engineering makes AI coding agents write code exactly the way you want to. Period. Deterministic libraries like ArchUnit, Spring Modulith, etc are the way to actually enforce your architecture rules.

Erik Meijer avatar
⟳ Kostis Kapelonis retweeted
Erik Meijer@headinthebox·Sep 15

About an hour into its mission to wipe out humanity, Claude will experience a compaction event, and it will have completely forgotten what is was supposed to achieve. Just like in all your coding sessions. Absolutely, nothing to worry about people.

Mitchell Hashimoto avatar
⟳ Kostis Kapelonis retweeted
Mitchell Hashimoto@mitchellh·Aug 31

To show how much I care about performance, I just ordered the worst specced 2020 Intel MacBook Air I could find (i3, 1.1 GHz, 8 GB RAM). It is the last supported hardware for macOS 15, our min target for Superlogical. Going to optimize on it for worst-case behavior.

🧗‍♂️ Matt Holt avatar
⟳ Kostis Kapelonis retweeted
🧗‍♂️ Matt Holt@mholt6·Aug 26

One day you copied and pasted from Stack Overflow for the last time and didn't realize it

Nova avatar
⟳ Kostis Kapelonis retweeted
Nova@NovaXCode·Aug 22

Someone bought “localhost8000” domain just to say “you can't even type a colon properly.” Peak engineering disrespect 💀

Post attachment
sunil pai avatar
⟳ Kostis Kapelonis retweeted
sunil pai@threepointone·Aug 22

“Look at my incredible new factory!” Yo that’s cool, what do you make? “It’s highly optimised, fully automated, zero tolerance for defects and with a continuous feedback cycle” Cool cool, so what do you actually make? “I can interact with it on my phone, laptop, messenger, completely async, and the shared context means it’s always learning how to get better” Very impressive, but what do you make? “Every agent has full context, can spawn other agents, review their work, fix defects, and ship continuously.” yes yes. WHAT DOES IT MAKE? “Software.” Oh nice. What software? “Well right now we’re mostly using it to improve the factory.” Improve it to make what? “Anything!” Such as? “…a better factory.”

Kostis Kapelonis avatar
Kostis Kapelonis@codepipes·Jul 1

It is amazing to see so many people in the field "rediscover" the Unix ecosystem (tmux, ssh/mosh, readline) in the course of learning to work with AI agents. Same for Git. Git worktrees used to be a niche feature, but they have now become an essential part of local development

Kostis Kapelonis avatar
Kostis Kapelonis@codepipes·Jun 8

Every OSS project maintainer should know by now Brandolini's Law, "The amount of effort it takes to review an AI generated PR is larger than the effort needed to create it". ☹️ antonydenyer.co.uk/2025-08-2…