Hacker News: haukebri

New comment by haukebri in "Deterministic Simulation Testing in Celld"

haukebri — Sun, 11 Oct 2026 06:14:30 +0000

[flagged]

New comment by haukebri in "Show HN: Quake ported to safe Rust, playable in browser"

haukebri — Fri, 09 Oct 2026 06:08:42 +0000

Adding to what you both said. I run pixel checks against live websites in a real browser, and I stopped doing full-page screenshots early on: scrollbars, lazy-loaded images and font timing shift pixels between runs like clockwork.

What works for me is diffing small stable regions, one component or one flow at a time, with a small per-pixel tolerance. I also keep a DOM-level assertion in front of the visual check, so when something fails I already know what changed structurally before I compare two images.

For the game port it's genuinely harder, the renderer decides the pixels, not the test. But the region idea transfers: lock the viewport and only diff what's stable.

Show HN: Built a little helper for answering questions while gaming

haukebri — Fri, 02 Oct 2026 12:38:18 +0000

I've built wizbert because I was constantly switching screens from Valheim to a wiki page, and got annoyed with it. So I build a little wizard that can stay in my game screen, while still helping with questions!

Comments URL: https://news.ycombinator.com/item?id=49932815

Points: 1

# Comments: 0

New comment by haukebri in "Pi 1.0"

haukebri — Fri, 02 Oct 2026 12:34:31 +0000

I did a performance test between pi (agent, not dev) and langgraph a few weeks ago. Since then I am absolutely sold on pi. Clear winner! The "old" tools are not automatically better <3

New comment by haukebri in "AI in Test Automation: Real Limitations vs. User Error"

haukebri — Wed, 09 Sep 2026 07:42:26 +0000

I love AI, but understanding errors and "bad" things is so tough.

I have built systems from Skills to full SaaS solutions for testing with AI. Sometimes it's great, sometimes it's just full on weirdness.

New comment by haukebri in "The Problem with Your AI Tests"

haukebri — Wed, 09 Sep 2026 07:38:40 +0000

Reading AI written tests feels like reading AI written stories from a few years ago. The text is there, things are happening but they are contradicting themself and make no sense. Sometimes you have a gem that is great.

I've been building agent skills against exactly this, cutting worthless tests and pushing the agent to write fewer but sharper ones. It's kind of the same approach as yours, just as a skill. --> github.com/haukebri/mira-test-engineer

New comment by haukebri in "Ask HN: Why don't LLM APIs have a first-class test mode?"

haukebri — Sun, 06 Sep 2026 08:07:06 +0000

It happens and I hate it. What happens: the tests are green and all looks falsely good. I have a bunch of mechanisms that try to prevent this.

I have a skill that acts as a guide to test writing. I have a CLI command that can test end to end in a browser, and I have a prod monitoring / testing SaaS.

New comment by haukebri in "What happens when your AI agent edits its own tests to pass?"

haukebri — Sun, 06 Sep 2026 08:03:13 +0000

It happens and I hate it. What happens: the tests are green and all looks falsly good. I have a bunch of mechanisms that try to prevent this.

I have a skill that acts as a guide to test writing.

I have a cli CMD that can test end to end in a browser and I have a prod monitoring / testing SaaS

Coding-agent skills that cut your worthless tests and write fewer, better ones

haukebri — Sat, 05 Sep 2026 17:17:51 +0000

Article URL: https://github.com/haukebri/mira-test-engineer

Comments URL: https://news.ycombinator.com/item?id=49578579

Points: 3

# Comments: 0