writing 8 posts 8 topics
Updated May 12, 2026
Field notes on comment spam, moderation, anti-spam patterns, and the operational shape of text classification at scale. Written for the engineers and site operators we'd want to read this.
Why LLM-generated comment spam slips past keyword rules, the structural patterns that still expose it, and how a calibrated classifier scores it in production.
Server-side classification, honeypots, time-trap fields, and trust scoring — how blog owners can stop comment spam without making real readers solve a puzzle.
What to use instead of Akismet when you've outgrown WordPress-first plugin spam filtering: API-first options, plugin-style replacements, and how to migrate without losing leads.
How honeypot fields catch basic bots, where they fail, the accessibility traps to avoid, and how to layer them with content classification.
How on-page anti-spam plugins add JavaScript, database work, third-party calls, and Core Web Vitals cost, plus the server-side alternative.
What trackbacks and pingbacks are, why modern platforms disabled them, and how old WordPress settings can still leak SEO value.
A practical moderation playbook: triage queues, trusted-author allowlists, edge cases, and throughput traps that wear teams down.
How spam comments affect crawl budget, link equity, user-generated spam policy risk, and rel=ugc handling on public discussion pages.