Top all-time · #11425
Michael Livshits
Personal blog · michaellivs.com/blogaiprogramming
Michael Livshits writes /dev/michael, a personal blog at michaellivs.com about ai and programming. Since 2025, 7 different Hacker News readers have linked it 7 times, most often “System reminders – how Claude Code steers itself”. It is still writing: the last post was on 2026-05-15, and you can follow it by feed.
Too few sampled comments (2) to say how often HN recommends it.?
Why HN links it
Linked as a source on how agent harnesses use system reminders to steer models, cited by one commenter as the usual approach. A second linked item is a Show HN for a Claude Code session visualizer.
Summarised by Claude from a sample of 2 HN comments that link here. Each comment is marked recommending, critical or neutral (citing without judging); how this works.
Thoughts on building with AI agents, Claude Code, infrastructure, and the craft of software engineering. By Michael Livshits.
Most-linked posts?
- System reminders – how Claude Code steers itself2 people · 2 mentions · first linked 2026-08-21
- Web Search for Agents in 20262 people · 2 mentions · first linked 2026-04-01
What Hacker News says?
That's what system reminders do in most harnesses. https://michaellivs.com/blog/system-reminders-steering-agent...
Comment linking /blog/system-reminders-steering-agents/
on “Clean up Claude 5's token vomit with a separate LLM”Submitted Web Search for Agents in 2026 as a story · 2 points
HN discussionI wanted to see what Claude Code was doing instead of watching text scroll by. This watches the JSONL conversation logs in ~/.claude/projects/ and renders animations based on tool events. Built with Go and Raylib. No API keys or network calls, just file watching. The animation system is a state machine with 10 states. Biomes use parallax layers at different scroll speeds. There's a progression system that persists X…
Comment linking /blog/claude-quest
on “Show HN: Claude Quest – Pixel-art visualization for Claude Code sessions”Submitted Web Search for Agents in 2026 as a story · 2 points
HN discussionSubmitted Reverse-engineering Claude's sandbox, then building my own as a story · 2 points
HN discussionSubmitted System reminders – how Claude Code steers itself as a story · 1 points
HN discussionSubmitted Reverse-engineering Claude's generative UI – then building it for the terminal as a story · 1 points
HN discussion
