
© Piotr Sikora 2026 | All Rights Reserved
A recap of a prompting workshop for young people at IdeaLab in Kielce: talking to an LLM is not magic, just simple, concrete communication. A few rules, some practice — and AI answers exactly the way you want.
A recap of the second Eskadra Bielika mission at Hotel Binkowski in Kielce: a hands-on workshop on deploying a RAG system with the Polish Bielik model on Google Cloud Platform. Event photos in the gallery.
Learn the six-part RTCROS prompting framework to structure your AI prompts and get consistent, high-quality responses from ChatGPT, Claude, and other LLMs.
Understand how the Simple Memory node works in n8n AI agents, how window size controls context, and why your chatbot suddenly forgets what you told it.
How I raised my Next.js website's Ahrefs Health Score from 28 to 99 in six weeks using Cursor, Vercel, and Google Search Console to fix SEO issues.
Analysis of 14 real-world scenarios: TOON cuts LLM token costs 75% and TRON 40% vs JSON, while staying compatible
Analysis of 14 scenarios: TOON cuts LLM token costs 75% vs JSON while maintaining full feature parity. Comprehensive comparison.
Since the emergence of LLMs on the market, I've been quite skeptical of them. Now I can say that they are a great tool for IT professionals.
Does it matter which JavaScript framework you use? A short reflection on labelling yourself an Angular, React or Vue developer and switching between them.