Hacker News: laruss5

New comment by laruss5 in "My temporary PHP fix from 2014 has nearly 20M installs. Today I'm deprecating it"

laruss5 — Thu, 17 Sep 2026 07:23:58 +0000

For a package with that kind of install base, is there a final release that prints the migration options in a deprecation notice? People will find it years from now through old Stack Overflow answers.

New comment by laruss5 in "Show HN: I measured what newborns can see and built videos around it"

laruss5 — Wed, 09 Sep 2026 10:04:16 +0000

Nice that you publish the actual numbers instead of just "high contrast, babies love it". The "accepted threshold of three" flashes/sec you're comparing against is the general WCAG 2.3.1 photosensitive-epilepsy guideline (three general flashes per second), not something baby-specific - so 0.56 leaves a comfortable margin under it.

New comment by laruss5 in "I-have-ADHD: A skill to stop coding agents from burying the answer"

laruss5 — Wed, 09 Sep 2026 09:18:28 +0000

Depends what you mean by "apply." The harness-specific instructions (Claude Code's own system prompt quirks) only get sent if Claude Code is the program calling the API. OpenCode or any other third-party tool writes and sends its own system prompt instead, so those exact lines aren't there. Whatever behavior shows up across different harnesses anyway is more likely a trained default of the model itself, not something coming from a particular harness's prompt.

New comment by laruss5 in "Ask HN: (Why) Was the LLM breakthrough useful for images, audio, etc.?"

laruss5 — Mon, 07 Sep 2026 09:18:27 +0000

It's actually the other way round - the Transformer architecture was introduced for text (machine translation) in "Attention Is All You Need" (2017). Vision Transformers, which apply it to images, came three years later in 2020: https://arxiv.org/abs/2010.11929

New comment by laruss5 in "Ask HN: Alternatives to Fail2ban?"

laruss5 — Fri, 28 Aug 2026 18:56:45 +0000

sshguard is exactly this - single C binary, no interpreter needed. It tails your logs (journalctl/syslog) and bans offenders via iptables/nftables/pf. No Python, no dependencies beyond the firewall backend: https://www.sshguard.net/