<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>amenophis.dev — Writing</title><description>Notes on voice AI, agents, and developer infrastructure by Amen.</description><link>https://amenophis.dev/</link><item><title>The anatomy of time-to-first-audio in a voice agent</title><link>https://amenophis.dev/writing/anatomy-of-first-audio/</link><guid isPermaLink="true">https://amenophis.dev/writing/anatomy-of-first-audio/</guid><description>From turn-commit to first audio: LLM ~59%, TTS ~28%, my orchestration ~14% — and one first token took 2.9 s. I was tuning the wrong layer.</description><pubDate>Sat, 08 Aug 2026 00:00:00 GMT</pubDate></item><item><title>Giving AI agents a real phone line</title><link>https://amenophis.dev/writing/voice-agents-real-phone-line/</link><guid isPermaLink="true">https://amenophis.dev/writing/voice-agents-real-phone-line/</guid><description>50 concurrent phone calls in userspace Go: p50 held at 20.00 ms at every load level — measured, not claimed. The tail is where it gets interesting.</description><pubDate>Sat, 08 Aug 2026 00:00:00 GMT</pubDate></item></channel></rss>