AI Agents in Production, Anthropic vs OpenAI & the LiteLLM Hack

Andre and Will break down where AI agents are actually heading in 2026, and why the real shift is not hype but operational adoption. They unpack the move from passive assistants to agentic workflows, the gap between pilots and production, and why human accountability still matters as enterprise AI scales.In this episode, they discuss:• why 2026 looks like the year companies move from AI experimentation into real operational use• the difference between assistants, partial agency, and fully autonomous agents• Anthropic vs OpenAI, and why enterprise teams may be leaning toward workflow-first AI products• why agents still struggle with deterministic outputs in critical business functions• the case for keeping humans accountable when AI systems make decisions• the LiteLLM supply-chain hack and what it says about AI security, package trust, and operational riskLunch, Tech and Chat is presented by Will and Andre, with conversations on AI, startupseb3, markets, and emerging technology.If you enjoy the show, follow Lunch, Tech and Chat on Spotify and share this episode with someone building with AI.