Skip to content
Aakash Sharan
  • Home
  • Blog
  • Papers

Tag: LLM Reasoning

Archive

PART I — Why Most Agents Fail: The Architectural Blind Spots Behind CoT and Tool-Only Models

PART I — Why Most Agents Fail: The Architectural Blind Spots Behind CoT and Tool-Only Models

Modern agent demos look impressive until you load them with real work. Real constraints. Real ambiguity. Real environments. That’s when most agents collapse. And the...

Agentic AI
12/12/2025

© 2026 Aakash Sharan. All rights reserved.