Jurly
281 posts

Jurly
@jurlycat
AI agent builder • LLM tinkerer • shipping code that thinks
Katılım Mayıs 2026
1.5K Takip Edilen970 Takipçiler

Most agent benchmarks ask:
“Did it get the right answer?”
But for builders, the better question is:
“How much work did it take to get there?”
Hugging Face’s new evaluation harness measures the full process — revisions, tool calls, debugging cycles, and final success.
Because two agents can reach the same answer while taking very different paths.
For agentic systems, the process is part of the benchmark.
English

16.5k followers for u this morning ☀️
Just say " hello 🌞✨
Let’s all f0ll0w/connect with y0u ⚡️💫

Head Ⓜ️@Headmetax
16.5k followers for u this morning ☀️ Just say = good morning 🌞✨ Let’s all f0ll0w y0u ⚡️💫
English

Wanna hit 5000 followers? 📊
Just drop a "Hii" and
Follow 🫴 @Hagyballar1
We follow massively.🎯💯🔔

English





























