---
title: "Rost Glukhov (@rgl) | Hashnode"
description: "Rost Glukhov (@rgl) writes on Hashnode. Read their latest articles and follow their work."
url: "https://hashnode.com/@rgl"
---

@rgl

# Rost Glukhov

@rgl · Joined December 2025

[Website](https://www.glukhov.org/)

### [I Tested 12 Self-Hosted Deep Research Tools — Most Are Just Search With Extra Steps](https://rosluk.hashnode.dev/i-tested-12-self-hosted-deep-research-tools-most-are-just-search-with-extra-steps?utm_source=hashnode&utm_medium=feed)

I ran the same research question through twelve different self-hosted Deep Research systems. Eight of them produced output that, on inspection, was no different from hitting a search engine five times

10 min read · [rosluk.hashnode.dev](https://rosluk.hashnode.dev/i-tested-12-self-hosted-deep-research-tools-most-are-just-search-with-extra-steps?utm_source=hashnode&utm_medium=feed)

[Oct 25](https://hashnode.com/posts/i-tested-12-self-hosted-deep-research-tools-most-are-just-search-with-extra-steps/6add6ffbdd2f4668d2269e29)

### [I Tested ROCm vs Vulkan for AMD Local LLMs — One Backend Silently Tanked by 6x](https://rosluk.hashnode.dev/i-tested-rocm-vs-vulkan-for-amd-local-llms-one-backend-silently-tanked-by-6x?utm_source=hashnode&utm_medium=feed)

ROCm vs Vulkan for AMD Local LLM Hosting

8 min read · [rosluk.hashnode.dev](https://rosluk.hashnode.dev/i-tested-rocm-vs-vulkan-for-amd-local-llms-one-backend-silently-tanked-by-6x?utm_source=hashnode&utm_medium=feed)

[Sep 15](https://hashnode.com/posts/i-tested-rocm-vs-vulkan-for-amd-local-llms-one-backend-silently-tanked-by-6x/6aa92eb317e1d2cddab1b503)

### [Your LLM Guardrails Don't Protect Your Agents — Here's What Does](https://rosluk.hashnode.dev/your-llm-guardrails-don-t-protect-your-agents-here-s-what-does?utm_source=hashnode&utm_medium=feed)

You've set up input validation. You're filtering outputs. Your model is sandboxed behind guardrails. Now your agents are calling tools, delegating to other agents, and acting on user data — and none o

10 min read · [rosluk.hashnode.dev](https://rosluk.hashnode.dev/your-llm-guardrails-don-t-protect-your-agents-here-s-what-does?utm_source=hashnode&utm_medium=feed)

[Aug 7](https://hashnode.com/posts/your-llm-guardrails-don-t-protect-your-agents-here-s-what-does/6a75a278103f169a06393a00)

### [Google's A2A Protocol in 2026: What Actually Works](https://rosluk.hashnode.dev/google-s-a2a-protocol-in-2026-what-actually-works?utm_source=hashnode&utm_medium=feed)

What Is the A2A Protocol? Agent Cards and Tasks Explained

10 min read · [rosluk.hashnode.dev](https://rosluk.hashnode.dev/google-s-a2a-protocol-in-2026-what-actually-works?utm_source=hashnode&utm_medium=feed)

[Jul 2](https://hashnode.com/posts/google-s-a2a-protocol-in-2026-what-actually-works/6a464aef2a3896c100e3166a)

### [GPU and RAM Prices Surge in Australia: RTX 5090 Up 15%, RAM Up 38% - January 2026](https://rosluk.hashnode.dev/gpu-and-ram-prices-surge-in-australia-rtx-5090-up-15-ram-up-38-january-2026?utm_source=hashnode&utm_medium=feed)

Today we are looking at the top-level consumer GPUs, and RAM modules. Specifically I’m looking at RTX-5080 and RTX-5090 prices, and 32GB (2x16GB) DDR5 6000. I’ve taken screenshots of cheapest prices for in-stock graphic cards on the three prominent ...

3 min read · [rosluk.hashnode.dev](https://rosluk.hashnode.dev/gpu-and-ram-prices-surge-in-australia-rtx-5090-up-15-ram-up-38-january-2026?utm_source=hashnode&utm_medium=feed)

[Jan 24](https://hashnode.com/posts/gpu-and-ram-prices-surge-in-australia-rtx-5090-up-15-ram-up-38-january-2026/69746eeed4df342f82ffb983)
