Community theme evolution across three capability layers

13 Apr 2026

Figure showing community theme evolution in LLM agent development across three nested capability layers — Weights, Co... image
In this moment: Dragan Spiridonov AI Chapter
Found on LinkedIn, sharing here with the MoT community:
"Donald Norman argued in 1991 that tools don't amplify cognition … they change the task by changing its representation. A checklist doesn't improve memory. It gives you a different task: read and follow. The system sees enhancement. The user sees new cognitive demands. Precomputation … building representations once, reusing them across people and time … is the mechanism. Context engineers rediscovered this 35 years later."
https://www.linkedin.com/posts/alindnbrg_cognitiveartifacts-contextengineering-representation-share-7449147172467580928-jg4T 
Dragan Spiridonov
Agentic Quality Engineer | Founder of Quantum Quality Engineering

In IT since 1996, establishing and leading QA/QE functions since 2014.
8 years at Alchemy, building QA/QE from scratch.
Since October 2025, the founder of Quantum QE, Ambassador for Agentics Foundation Serbian chapter, and creator of the Agentic QE framework and Agentic QE Fleet, an open-source AI-powered QA/QE platform.

Chapter Lead
Lisa Crispin
As they say, there's nothing new under the sun. I admit though, I'm ignorant on what is the big takeaway here? I'm not yet a practitioner of context engineering, so maybe that's why I don't see it.

Sign in to comment
Explore MoT
AI-driven testing in practice: from requirements to reliable automation image
See where AI genuinely helps, where it doesn’t, and how testers can stay firmly in control
Prompting for Testers image
Unleash the power of generative AI to boost your software testing and day-to-day tech tasks
This Week in Quality image
Debrief the week in Quality via a community radio show hosted by Simon Tomes and members of the community
Subscribe to our newsletter