Don’t make the model pay 200 lines to find a functionCoding agents still read like find | cat: dump the tree, grep wide, page 200-line windows until context is landfill. Codeloom’s engine makes reading a cost hierarchy, then puts TypeSafe System One (je12ABSep 21 · 6 min readblog.iresharma.com
Trimming the Fat, Not the SignalMost writing about coding agents focuses on the loop: call the model, run its tools, feed results back, repeat. codeloom.engine's loop is deliberately unremarkable — an OpenAI-style tool-calling loop 1Sep 10 · 9 min readblog.iresharma.com
Why a Coding Agent Needs Three Different Ways to "Read" CodeIf you're building something like Devin, an agent that goes end-to-end from a task description to a working code change. The very first problem you hit isn't code generation. It's comprehension. Befor69JESAug 18 · 8 min readblog.iresharma.com
I never finished Venture Deals, so I built a pipeline to read it for meI've started Venture Deals three times. I've never finished it. That's not really a discipline problem, it's a format problem. I care about maybe 15% of a book like that: the definitions, the mechanic21JAug 7 · 17 min readblog.iresharma.com
Building an AI-Powered YouTube Shorts Generator: A Complete Technical Deep DiveIn the era of short-form content dominance, creating engaging YouTube Shorts consistently can be a time-consuming challenge. Today, I'm excited to share a comprehensive technical breakdown of an open-source YouTube Shorts generator that automates the...1TJun 10, 2025 · 5 min readblog.iresharma.com
