Hacker News: MehdiBelkacemhttps://news.ycombinator.com/user?id=MehdiBelkacemHacker News RSShttps://hnrss.org/hnrss v2.1.1Mon, 18 May 2026 08:45:04 +0000<![CDATA[New comment by MehdiBelkacem in "Ask HN: Does anyone know what a monad is?"]]>chimichangas

]]>Thu, 14 May 2026 10:47:45 +0000https://news.ycombinator.com/item?id=48133581MehdiBelkacemhttps://news.ycombinator.com/item?id=48133581https://news.ycombinator.com/item?id=48133581<![CDATA[New comment by MehdiBelkacem in "Launch HN: Ardent (YC P26) – Postgres sandboxes in seconds with zero migration"]]>[flagged]

]]>Thu, 14 May 2026 10:45:56 +0000https://news.ycombinator.com/item?id=48133568MehdiBelkacemhttps://news.ycombinator.com/item?id=48133568https://news.ycombinator.com/item?id=48133568<![CDATA[New comment by MehdiBelkacem in "Ask HN: What will happen as AI costs increase?"]]>Token anxiety is real. What worked for me: prompt caching on fixed system prompts cut my Anthropic bill by ~60% overnight. Most devs don't realize cache writes are 25x cheaper than input tokens on Claude.

Local models for classification/routing + frontier only for generation is the other move — but the latency tradeoff is real if you're in a user-facing flow.

]]>Fri, 08 May 2026 03:12:40 +0000https://news.ycombinator.com/item?id=48058066MehdiBelkacemhttps://news.ycombinator.com/item?id=48058066https://news.ycombinator.com/item?id=48058066