In this week's episode of AI That Works, we're diving deep into the practical side of AI development with another "No Vibes Allowed" segment! Join us as we livestream a collaborative coding session, giving you an inside look at how we build real-world features for our respective AI tools.
This time, we're tackling message queuing within Riptide (or Code Layer, name TBD!), our in-house IDE. We'll explore how to efficiently manage interactions with AI agents, allowing users to queue up prompts even while a process is running. We discuss the strategic use of context windows, and the importance of baking in assumptions during the design phase to ensure a correct implementation.
We also showcase the addition of learning test outputs, which we find incredibly valuable for observing the behavior of internal cloud state machines. Tune in to see us put theory into practice and build a solution live!
Check out our github: https://www.github.com/boundaryml/baml
AI That Works repo: https://github.com/ai-that-works/ai-that-works
Socials:
X: https://x.com/boundaryml
Discord: https://discord.com/invite/yzaTpQ3tdT
LinkedIn: https://www.linkedin.com/company/boundaryml/
⏱️ Chapters:
Intro
Technical Concepts Discussion
Great Leaders are Right
Riptide Workflow
Preserving Files with Riptide
Queue Messages Logic
Detour Coaching Discussion
Design Documents Purpose
Optimization
Transcript
I've watched this one closely. Ask me anything about it and I'll point you to the exact moment.