Blog | Srijan Shukla

Blog

AI

The Shortcut That Backfired: One-Shot Prompting and the Step-Skipping Problem

How I learned that bundling multiple tasks into a single prompt is a false optimization, and why sequential gated prompting is the better model for agentic workflows.
What Changed When Agent Design Stopped Being Mostly Prompt Engineering

Once the frontier models got good enough at long-horizon work, the bottleneck moved outward. The real leverage started showing up in the harness, the context, the memory system, and the control loop around the model.
What Happened When I Tried to Coordinate Two AI Agents Over NFS

A shared mount looked like the cleanest way to make two agent instances talk across machines. A 30 minute directory cache made the whole thing unreliable, and Git ended up being the simpler bus.
Vibe Coded a `watch` Command for Autonomous LLMs, and It Became My Debugging Loop

Once I had LLMs editing code autonomously, I wanted a better way to watch a directory breathe in real time. Existing tools covered file events or diffs or a TUI, but not all three together.

Infra

Others