Archives on rkar.org

The Insatiable Urge to Change Everything on Day One

Sun, 28 Jun 2026 00:00:00 +0000

You worked at a place for six years. Six years is a long time to be anywhere. You absorbed the culture, hoarded a small mountain of tribal knowledge, and earned the kind of respect that comes from simply having been there. Somewhere in all that you quietly decided you can solve any problem, because you solved the last ones. Your brain is happy, proud and a little too courageous.

Then you switched jobs because you wanted something different, or because the problems stopped being hard enough to be interesting. But you do not leave empty handed. Nobody tells you that the most dangerous thing you bring to a new job is not your bad habits, it is your good ones.

The Art of Communication in Tech

Sun, 21 Jun 2026 00:00:00 +0000

More often than not, I get messages like this:

Mark: Hey! …4 hours later… Me: ??????? Mark: I have a few questions. Up for a quick call? Me: Hmm… what's the question? Mark: So, my client’s kid forgot their lunchbox. I was wondering if we have any protocols for that. Me: Yep, that’s actually a common issue. There’s a Confluence page with all the details. I wrote it, in fact, because you've asked me this same question multiple times over the last five years. Mark: Whoa, easy there. You know I decide your yearly salary hike, right? Me: Haha, just kidding! I know you’ve got a lot going on.

Solving Bongard Problems using VLMs

Thu, 28 Aug 2025 00:00:00 +0000

In 2018, this post made an interesting attempt at solving Bongard problems using deep learning. That work may have laid the foundation for further attempts at creating various Bongard-like datasets such as Bongard-LOGO¹, Bongard-HOI², and Bongard-OpenWorld³. This post is an attempt to replicate the same methodology with state-of-the-art vision-language models (VLMs).

A Quick Intro

Bongard Problems are puzzles formulated by Russian scientist and intelligence theorist Mikhail Moiseevich Bongard in the 1960s in his book The Problem of Recognition (Problema Uznavaniya). A typical Bongard problem consists of two sets of six images. All images on one side illustrate a shared concept that is absent from the other side.

Agentic Programming is No Fun

Wed, 25 Jun 2025 00:00:00 +0000

This post is based on my personal experience after dabbling with AI-based agentic programming for a few days. Not that anyone actively reads my blog, but I want to document how I am feeling at this point in time, during what feels like the early rise of AI automation.

How I Use AI Today

I have been consistently using AI chatbots over the past couple of years for programming-related tasks. I barely Google error messages anymore; I just paste them into ChatGPT or Claude and move on. I am on the free tiers now, but I did subscribe for a few months after realizing Claude was better at generating code.

Elegance is in the Eye of the Beholder

Sun, 08 Dec 2024 00:00:00 +0000

Introduction

Go recently added support for a syntax that a fellow gopher had requested 12 years ago on the forum. The OP proposed a range syntax for Go that was "simpler" than the traditional C-like iteration.

The discussion that followed highlighted an interesting divide - while a few found the proposed syntax cleaner, others argued that the traditional C-style loop was more explicit and natural, especially for developers with C background. This debate about what makes loop syntax "elegant" has historical roots.