What is Serverless, and can it run agents & other genAI apps?
summary
Explains what serverless actually does: containers plus autoscaling to zero, Cloud Run as Cloud Run with GPUs that runs containers (any language, any binary) and now hosts inference and agents; Cloud Functions is folded under Cloud Run; AppScript provides tight Workspace automation but not autoscaling. Shows a working GenAI web app under 200 lines using the Gemini API code in Python & JavaScript and deploys to App Engine or Cloud Run with no code changes. Notes historical context like 2008 App Engine launch and a 10k requests per second scaling anecdote.
AI-written summary · about
chapters (11)
- 0:00Serverless and AI intro
- 5:59What is serverless
- 12:40Google Cloud serverless
- 16:11App Engine basics
- 19:24Cloud Functions events
- 21:51Cloud Run containers
- 25:05AppScript Workspace
- 27:10GenAI on serverless
- 30:39Choosing serverless
- 34:28Cost tiers credits
- 35:32Q&A
related talks
07mServerless: Accelerate your AI
38mGCP Serverless Updates
50mGo Serverless with Google Cloud
43mGoogle Cloud's new serverless features
43mWhat's new in Google Cloud Serverless
35mServerless Experiences -- On Your Own Servers?
34mServerless Computing with Python
38mServerless tools for machine learning, big data and app backends