Scott Hanselman: Semantic Kernel - Episode 280

Scott Hanselman is a teacher at heart. He speaks all over to whoever will listen. He's written code that you, dear listener, has used. Scott has been blogging, coding, and podcasting a LONG time. He codes, writes, speaks, empowers, promotes, braids, learns and listens - usually not in that order. And he's a Vice President at Microsoft in his day job. You can find him on YouTube, TikTok, Instagram, and Mastodon! His Hanselminutes podcast has surpassed 900 episodes, and his Azure Friday show, over 750 episodes.

Topics of Discussion:

[4:33] What should developers be focusing on?

[625] Understanding the complexity of AI development.

[8:09] Without understanding context, we can't make good prompts.

[8:42] What are the levels of the pyramid that every developer should understand?

[9:16] Developers should start by learning the basics of AI.

[9:22] The question of who is responsible when a website or model goes down.

[14:15] Grounding your AI in reality.

[15:19] Edge deployed AI model.

[17:00] A foundational model is a machine learning model that has been pre-trained on a data set.

[20:40] The limitations of large language models.

[21:00] AI transformer models and their growth in size and complexity.