The InfoQ Podcast

Generally AI: Time to Travel

Dec 23, 2024
Time zones can be baffling, and the discussion reveals their quirks, including oddities like Troll station and Lord Howe Island. Delving into predictive modeling, the speakers share the challenges of machine learning with temporal data, including humorous takes on daylight saving time. The transformative role of the Transcontinental Railroad is highlighted, showing how it revolutionized travel and commerce in America. The conversation also draws parallels between historic projects and current software industry challenges, particularly in adapting to new technologies.
Ask episode
AI Snips
Chapters
Books
Transcript
Episode notes
ANECDOTE

Google Flu Trends Failure

  • Google Flu Trends aimed to predict flu outbreaks by analyzing search queries.
  • The project failed due to correlation not equaling causation and noise in the data.
ANECDOTE

Dynamic Pricing Mishap

  • A bus rental platform expanded from Germany to other countries using the same pricing model.
  • They failed to account for different national holidays and events, leading to underpricing.
ANECDOTE

E-Scooter Problem

  • A self-driving car company's model didn't recognize e-scooters after they became legal.
  • Roland had to collect and annotate new data to adapt the model to this new object class.
Get the Snipd Podcast app to discover more snips from this episode
Get the app