When AI becomes your co-pilot, UI alone isn’t enough—you’re also designing the conversation, the context, and the underlying behavior of the model itself. Whether you’re shipping an AI product or adding AI to an existing one, you need to bridge the gap between human needs and model behavior.

In this talk, Vitaly will explore how core levers like temperature, scoping, and prompt presets shape user experience just as much as layout or flow. You’ll learn how concepts like token nudging, daemons, and the Assistant Pattern can be used intentionally to build AI features that feel intuitive, responsive, and aligned with user intent.

AI is fun to say, but is it BS? Join Scott Hanselman as he explores "LMs both VL and S" (very large and small) and learns if they are useful or not? What about NPUs on your phone or laptop? What can a small model do that can save power and money? How many TOPS (tensor operations) are needed for a model to add value in your life and be worth your time?

In this talk, Scott will dive into how he uses AI in his day to day to be more productive, without turning into an AI Grifter!

Let's take a little time to discuss all that we have heard, together. Bring your questions, and let's get the discussion going!