Ai on Hrushikesh Dokala

context switching is all you need

Sat, 30 May 2026 00:00:00 +0000

i have 4+ terminals open right now. each one is running claude code (in bypass permissions). each one is working on a different project.

2 of them are running experiments in independent worktrees, and the other 2 which im working on closely to debug and fix a few prod issues.

i’m not doing any of this. i’m just… orchestrating.

the old advice #

there’s a popular take that calls context switching “the silent killer of productivity.” the advice: batch your tasks, use do-not-disturb, single-task your way to clarity.

is ai making us dumber?

Wed, 26 Nov 2025 00:00:00 +0000

something which i’m observing in myself is:

when i try to read, chat with ai to understand a concept at a deeper level, why does it fade away much faster than what i’ve just written down and put some conscious effort into understanding?
it’s easy to understand things with ai, but is it easy to forget as well?
do you think faster learning without cognitive effort will result in losing the learning from memory/understanding?

are people jumping into answers and thinking less?

moe - mixture of experts

Mon, 17 Nov 2025 00:00:00 +0000

hey, i was trying out the composer 1 in cursor, and was trying to understand how is inference super fast and even the code is very accurate at the same time, and how much does it cost (interms of gpus usage) and how did they scale to all cursor users.

so i started digging into a few articles and papers, got my hands on a few open source models like ibm-granite/granite-3.1 and mixtral 8x7b which are sparse mixture of experts models. but then, i didnt understand how does it work, was wondering what is 8x7b? is it 8 times better than 7b? or 8x the size of 7b? well, the model has 8 experts, each with 7 billion parameters mistral style model. what the hell are these experts?

how to almost reduce llm costs by 80%

Sat, 01 Jun 2024 00:00:00 +0000

If you are an AI developer or building a SAAS wrapped around AI foundational models, its really necessary to think about spending a lot of money on LLMs without any prior knowledge on how to reduce these costs..

change the model #

By replacing the LLM like GPT-4 with a small language models like phi-3 or Mistral for specific tasks that doesn’t need more precise and optimized responses. this way you can have major cost cuttings.