Uploaded August 2025 | Updated November 2026, 17 minutes ago
A model can crush benchmarks and still flop in the real world.
UC Berkeley Professor and RunLLM Co-founder Joseph Gonzalez explains why leaderboard rankings aren’t the whole story — and why he pays closer attention to what actually works for his students.
What you’ll learn in Ep 21:
✅ Why Joey uses grad student feedback as his “source of truth”
âś… How tuning has led to overfitting for benchmark performance
âś… Why the best model depends on your task, not your leaderboard rank
As Joey puts it: benchmarks help, but real-world results matter more.
📺 Watch more LLMs on the Run:
Playlist → tinyurl.com/msxmhph7
✨ Subscribe for more AI insights:
tinyurl.com/445y4sd3
🤖 Explore RunLLM — The AI Support Engineer for advanced technical support: tinyurl.com/kcmy67kh
#LLMsOnTheRun #AIModelEvaluation #AIUX #Benchmarking #FoundationModels #RunLLM
A model can crush benchmarks and still flop in the real world.
UC Berkeley Professor and RunLLM Co-founder Joseph Gonzalez explains why leaderboard rankings aren’t the whole story — and why he pays closer attention to what actually works for his students.
What you’ll learn in Ep 21:
✅ Why Joey uses grad student feedback as his “source of truth”
âś… How tuning has led to overfitting for benchmark performance
âś… Why the best model depends on your task, not your leaderboard rank
As Joey puts it: benchmarks help, but real-world results matter more.
📺 Watch more LLMs on the Run:
Playlist → tinyurl.com/msxmhph7
✨ Subscribe for more AI insights:
tinyurl.com/445y4sd3
🤖 Explore RunLLM — The AI Support Engineer for advanced technical support: tinyurl.com/kcmy67kh
#LLMsOnTheRun #AIModelEvaluation #AIUX #Benchmarking #FoundationModels #RunLLM










![RunLLM Introduction [2025-01]
Note: This is an outdated demo of RunLLM. For the latest demo, please go to runllm.com/demo-video
RunLLM is an AI-powered technical support engineer that saves your team time, accelerates customer adoption, and generates valuable insights from customers interactions. It does that by building expertise in your product, which comes from reading and understanding documentation, guides, and APIs.
RunLLM is currently in private beta. To learn more about RunLLM, please visit runllm.com and request access! RunLLM Introduction [2025-01]](https://i.ytimg.com/vi/jxf2GmCB5eU/mqdefault.jpg)