Improve episode 1

[gcroci2]

• Cut on the theory (especially the redundant parts)
  • For example, key points for being more aware of how AI generates content are 1) the training factor (bad data comes in, bad output goes out) and 2) the probabilistic nature of AI response. I think these were there in the introduction, but they could have been delivered simpler, shorter, and more straightforwardly. You may consider trimming a bit the introduction
• In the episode, add a mention to temperature settings for some tools, such as Copilot (but to Codeium!)
• In the episode, look up if benchmarks about the tools exist, and in case link them (callout)
• Add note on code churn and degrading code quality due to AI coding (maybe as callout in the episode, and evaluate whether it needs to be added in the slides)
• Consider adding Cursor AI to the list of coding assistants, it is one of the most popular ones at the moment.
• Slide with table: show only logos, no info, just mention major differences (very high level) you can find and then refer to the lesson materials