Σημειώσεις επεισοδίου
AI is easy to use — but hard to scale.
In this episode of Domesticating AI, we’re joined by Daniel Dowler (Red Hat) to break down what actually happens when you move from calling APIs to running AI systems yourself.
Recorded on April 21st
Most developers interact with AI through APIs — fast, simple, and pay-per-token. But behind the scenes, those systems rely on GPU scheduling, batching, and infrastructure that doesn’t behave like traditional software.
We cover:
• Why GPU scaling is fundamentally different from CPU scaling
• Why tools like vLLM are becoming the default for high-performance…
Ιστότοπος: podcasters.spotify.com

