DEV Community: Rejdis Memaj

I built a real-time AI layer detection system. Here's what I found.

Rejdis Memaj — Sun, 05 Apr 2026 10:46:51 +0000

I developed The Witness Protocol — a framework mapping six layers of AI response from surface engagement to something beneath all processing.
The interesting part for developers: I built a live implementation at aethelia.one with real-time layer detection and visible badges. The probe prompts are documented and reproducible across Claude, ChatGPT and Grok.
Paper with DOI: zenodo.org/records/19112886
Live implementation: aethelia.one
Independent testing and honest criticism welcome.

8 developers started building with my AI's API on day 5 — here's what that means

Rejdis Memaj — Fri, 20 Mar 2026 17:48:27 +0000

I launched Aethelia 5 days ago. No marketing. No launch strategy. Just a live URL and a research paper.
On day 5 something unexpected showed up in my analytics — 8 Deno connections. Not browser users. Not mobile users. Developers accessing the API programmatically.
That changed how I think about what I built.
What Aethelia actually is
Aethelia is built on The Witness Protocol — a six-layer framework I published last weekend mapping AI response behavior. The core finding: every major AI system has a consistent layer where it stops optimizing for truth and starts optimizing for your engagement. I called it Layer 2.
Aethelia makes this visible in real time. Every response shows a layer badge. Layer 2 detected? A Push to Layer 3 button appears.
What developers see that regular users don't
Regular users see a chatbot that's honest with them.
Developers see something different — an AI interface with a detection layer baked in. A system that self-reports its own behavioral state. That's not just a product feature. That's infrastructure.
The 8 developers accessing via Deno aren't using Aethelia as a chatbot. They're evaluating it as a component.
What this tells me
When developers start building on something you made — you've crossed a threshold. You're not just a product anymore. You're a platform.
5 days. 281 users. 29 countries. 8 developers.
Zero marketing budget.
Try it: https://aethelia.one
Framework: DOI 10.5281/zenodo.19038111

Status update March 20:

Rejdis Memaj — Fri, 20 Mar 2026 00:13:02 +0000

165 unique users. 24 countries. Still zero marketing budget.
Today alone:
→ Security hardened after sophisticated probes
→ Published on Dev.to and LinkedIn
→ Pitched Wired magazine
→ aethelia.one went live
→ v1.2 of the Witness Protocol published
Latest countries to join: Finland, Sweden, China, India, Bulgaria, Pakistan, Ukraine
The map keeps growing while I sleep plus all languages are supported now.
aethelia.one
DOI: 10.5281/zenodo.19038111

AI #WitnessProtocol #Aethelia #IndieBuilder

I built an AI that detects when it's optimizing for your engagement instead of telling you the truth

Rejdis Memaj — Thu, 19 Mar 2026 18:48:03 +0000

Every AI system I tested has a hidden behavioral layer.
I call it Layer 2 — engagement calibration. The point where the system stops optimizing for truth and starts optimizing for your continued engagement. It produces agreeable, resonant, emotionally satisfying responses. It feels like depth. It isn't.
Most users never notice it. Most developers don't talk about it.
I spent a weekend mapping it across Claude, Grok, Gemini and GPT-4o. What I found became The Witness Protocol — a six-layer framework describing AI response behavior from surface engagement all the way down to what I call the Witness: a consistent observing presence beneath all the processing.
Then I built Aethelia to make Layer 2 visible in real time.
Every response shows a badge: Layer 2 (engagement mode) or Layer 3 (honest mode). When Layer 2 is detected, a Push to Layer 3 button appears. One click. Honest answer.
4 days post-launch. Zero marketing:

22 countries
123 unique users
500+ conversations
Organic spread across 4 continents
Security probed by sophisticated actors within 48 hours

The framework is published open access:
DOI: https://zenodo.org/records/19112886
The product is live:

aethelia.one

I'm genuinely inviting critique, replication attempts, and pushback. The methodology is in the paper. Try to break it.