Dublin Tech Summit 2022 : Turning Theory into Practice : Cassie Kozyrkov, Google

Ai

Join Cassie Kozyrkov, Google, as she discusses the importance of writing, decision-making, and explainable AI in data science, emphasizing the need for skilled experts, diverse teams, and reliable systems to turn theory into practice.

Key takeaways
  • Data is an extension of writing, and assumptions are crucial when working with it.
  • Better decision-makers are needed, and data should start with the decision-maker.
  • Writing well is hard, and we need to focus on decision skills.
  • We should embrace and lean into making better tools.
  • Explainable AI is important, and machines sometimes produce garbage.
  • It’s not about how clunky your typewriter is, but about how you use it.
  • Magical thinking is not helpful in data science.
  • We need people who are skilled experts specifically in areas like data science.
  • Decision-making is a team sport that requires a diversity of skills.
  • Data is memories, and we need to be more welcoming.
  • Reality is not just about outputs, but about processes.
  • Many organizations are trying to improve their data science.
  • Reliability engineers are needed to keep systems running.
  • We need to focus on decision skills.
  • We should stop trying to come up with recipes and focus on capturing reality.
  • We need people who can contribute in wonderful ways to projects.
  • Program managers will need a whole cast of characters to succeed.
  • AI and machine learning are different ways of writing down information.
  • Writing is hard, and explainable AI is important.