Angelina YangFounder and Host, TwoSetAI · AI builder for visibility and discoverability

Why They Killed RAG & MCP: Real AI Lessons with Gil Feig

10 Apr 2026 · 0 min

Watch on YouTube

0:24

In this segment, Gil Feig critiques the current state of Multi-Contextual Prompting (MCP) implementations. He argues that most offerings are unusable because vendors rushed to market, claiming MCP support without providing robust tools, reliability, or deep testing. Consequently, many users have abandoned these solutions after initial failures. Feig’s team avoids official MCP servers entirely, opting to rebuild them from scratch to ensure functionality. He also notes that APIs from large enterprises are often outdated, leading developers to attach MCP servers to problems they don’t actually address.

The vast majority are unusable. Everyone just wanted to be first to market and say, "We support MCP!" They didn't expose a lot of tools. They didn't make things super reliable. They didn't test really deeply. And so all these tools out there that people tested once and were like, "This doesn't work super well. I'm not going to use this." We don't use any official MCP servers. We rebuild all of them. A lot of the APIs at these big companies haven't been touched in years. And so you have people trying to tack on MCP servers with no real connection to the problem.