Your AI budget is too generous and too restrictive at once: routing saves, flat caps kill rewrites
ruslansv · x · 2026-11-16
An AI budget can be too generous and too restrictive at the same time: routine work gets the most expensive model, a second review needs an exception, and a rewrite that could remove a business constraint hits a flat monthly allowance—illustrated by a cartoon of "Jira Update" arriving by limousine while "System Rewrite" is blocked by an exhausted-allowance gate. Fixes: model choice, caching and routing shrink the bill, while a concentrated rewrite can be worth spending more; across-the-board cuts preserve all three mistakes.
More from Companies & People
- Google DeepMind Ships 7 Developer Launches in a Week, Led by Gemini 4 Carbon and Hydrogen — _philschmid · 2026-11-17
- Gary Marcus amplifies swipe: Jensen Huang manages like a YOLO trader bound to blow up — GaryMarcus · 2026-11-17
- Production AI Agents for Generative Media course kicks off with live stream — Al_Grigor · 2026-11-16
- YC investor: recent batches skew hard tech and hardware; Matic claims most-deployed consumer robot — ycombinator · 2026-11-16
- Uber burned its 2026 AI coding budget by April; 35% of big enterprises are cutting token spend — ruslansv · 2026-11-16
- Meshery ContribFest at KubeCon onboards new contributors as Kubeflow marks CNCF graduation — HowDevelop · 2026-11-16