Adam Wenchel & John Dickerson, Arthur | AWS Startup Showcase S3 E1
summary
Arthur describes building an Arthur chatbot in January and rebuilding it weeks later because tooling shifted dramatically, demonstrating fast iteration on LLM stacks and measurable improvements. The company provides on‑prem and cloud monitoring for LLMs, tracks inference cost per token, watches input distribution drift and RLHF value‑alignment risks, and publishes performance metrics to calculate ROI — citing cases with over 100% first‑year ROI — and now offers the product via AWS Marketplace.
AI-written summary · about
