

IBM partners with Anthropic, plus OpenAI drops AgentKit
Oct 10, 2025
Join Olivia Buzek, a savvy AI developer advocate, Chris Hay, a distinguished engineer, and Mihai Criveti, an expert in agentic AI, as they dive into the latest in AI partnerships and products. They discuss OpenAI's AgentKit, exploring its accessibility and trade-offs between code and no-code solutions. The trio also unpacks IBM's collaboration with Anthropic, highlighting the importance of secure enterprise AI. Finally, they ponder whether AI could replace radiologists, navigating the balance between automation and human intuition in healthcare.
AI Snips
Chapters
Books
Transcript
Episode notes
AgentKit Lowers Access Barriers
- OpenAI's AgentKit makes agent creation broadly accessible via a visual builder and evaluation tools.
- Visual builders lower the barrier for non-technical users but don't eliminate the need for engineering practices.
No-Code Complements, Not Replaces Code
- Visual no-code agent builders complement rather than replace code-centric development for scalable systems.
- Probabilistic LLM behavior requires deterministic programmatic scaffolding to ensure consistent outcomes.
Prototype Visually, Ship As Code
- Start with visual builders for quick automations, then export to code for version control and scaling.
- Use SDK-generated TypeScript/Python from the builder to integrate with enterprise workflows and source control.