Building a ChatBot – live blog Oracle Code New York

Title: DevOps at Scale; Greek Tragedy in Three Acts
Speakers: Maria Kaval and Shaun Smith

See my live blog table of contents from Oracle Cloud

Started by talking about 6 trends

Serverless functions

  • Spins up when funtion calls
  • Goes away after
  • Like Cinderella’s carriage – but with a server. Only there for a short time

DevOps -> NoOps

  • Taking work of ops away from you
  • As developer, just want to write your code
  • Less emphasis on memory management and such

Open Source

  • Oracle cloud based on open source
  • Not focused on profit

Chatbots

  • Teens like testing and emojis
  • Adults like to text too; good interface
  • With chatbots, don’t know if talking to human or bot

Blockchain

  • More than just bitcoin
  • Ledgers build trust

Machine Learning

  • Now have proessing/compute power to enable machine learning

Use case – selling and buying a car

  • The chatbot asks basic info and calls serverless functions to look up value
  • Showed Oracle bot builder service – set up intents (phrases that represent what you want to do), train bot (ex: linguistics, machine learning)
  • Test bot by trying a chat in the config screen
  • Artificial intelligence integration trains bot

Serverless Functions

  • Serverless is a category
  • There are still servers; you just don’t have them. Instead the cloud provider does
  • Economics – only pay when service is being used – so only when called vs standing service
  • Agility – small amounts of code. So easier to write/debug/etc
  • Reliability – in cloud
  • Innovation – easier to try things out since easy to deploy

Fn

  • http://fnproject.io
  • Sample commands – init, run, test, deploy, call
  • Flow UI lets you watch as code runs. Looks like a sequence diagram except live; shows how long took, etc

Leave a Reply

Your email address will not be published. Required fields are marked *