Roman Kolesnev

@ffloyd
3 Followers
10 Following
21 Posts

I'm passionate software engineer with humanistic mindset.

#elixir #nix #vim

Personal Websitehttps://ffloyd.space

Today I found my answer to the question: "Why I didn't join Fediverse earlier?"

https://ffloyd.space/thoughts/2026-03-18-dopamine-platforms-fediverse

#mastodon

Dopamine, Platforms & Fediverse ⋄ ffloyd.space

I created my first Mastodon account years ago, but abandoned it almost instantly. Today I realized why.

My favorite Elixir operator is `&%{&1`.

It's a good time to start using all that LLMs power to stop enshitification!

#ai #programming #enshitification #ffloydspace

https://ffloyd.space/thoughts/2026-02-28-it-s-just-docs

It's Just Docs ⋄ ffloyd.space

Agents, skills, memories... aren't we overcomplicating? What if I show, that you can replace them with investments into high quality?

I see the following trend:

> Why to buy software if we can vibecode our own solution?

Does it mean that classic schema of "making startup, selling startup" does not work anymore?

Why to pay if you can "vibesteal"?

Therefore, if you want prospering business in the new reality - it cannot be vibecoded. Anything that can be vibecoded can be "vibestolen".

#ai #software #startups

❓Have you noticed that digital products and services are getting worse? So have we!

➡️We have published a report about enshittification, on how and why digital products and services keep getting worse - and how we can turn the trend (hint: open tech, enforcement, public policy++)

Obviously @pluralistic is a big inspiration and help in this work.

More than 80 groups in Europe and the US has joined in a call to action.

More here: https://www.forbrukerradet.no/breakingfree

Enjoy this short film!

Oh... I spent several hours writing this. But I really needed it. To offload all that thoughts about how LLMs will affect software industry.

Will return to this in a year or two. But now, I will use my freed mental capacity to make something awesome! =)

#ai #llm #programming #future #ffloydspace

https://ffloyd.space/thoughts/2026-02-25-q1-2026-how-i-see-ai-and-its-future

Q1 2026: How I See AI and its Future ⋄ ffloyd.space

I want to return to this article in 1-2 years and check how good I am at predicting the future and grasping essence of emerging technologies.

Option A: 10k lines of code done in one day.
Option B: 1k lines of code done in one week.

The same functionality is implemented.

What is better for a product that aims to survive in the long run? 🤔

#programming #ai

When you use AI agents from the Internet - it's like copy-pasting someone's documentation into your code. Agent - is no more than a doc with suggestions and conventions.

If AI works much better with agents on your project - it can be a sign that you have a huge problems with documentation, not that agent is "smart".

#ai #agents #programming

i built an entire x86 CPU emulator in CSS (no javascript)

you can write programs in C, compile them to x86 machine code with GCC, and run them inside CSS

https://lyra.horse/x86css/

We've been on X/Twitter for many years, but it's time to reduce our activity there and instead promote Mastodon as our main social media channel now. So we've done just that: https://x.com/LibreOffice/status/2026204949760131158 – Welcome to all our new followers here 😊
LibreOffice (@LibreOffice) on X

Hello, world! 👋 From now on, Mastodon is our preferred social media channel. It's an open source, decentralised platform – not controlled by tech giants. Follow us here: https://t.co/KZpwR61V5R

X (formerly Twitter)