zyel@lab:~$ cat /etc/motd

 ███████╗██╗   ██╗███████╗██╗         ██╗      █████╗ ██████╗ 
 ╚══███╔╝╚██╗ ██╔╝██╔════╝██║         ██║     ██╔══██╗██╔══██╗
   ███╔╝  ╚████╔╝ █████╗  ██║         ██║     ███████║██████╔╝
  ███╔╝    ╚██╔╝  ██╔══╝  ██║         ██║     ██╔══██║██╔══██╗
 ███████╗   ██║   ███████╗███████╗    ███████╗██║  ██║██████╔╝
 ╚══════╝   ╚═╝   ╚══════╝╚══════╝    ╚══════╝╚═╝  ╚═╝╚═════╝

zyel@lab:~$ █

zyel@lab:~$ cat about.txt

> about_this_lab

This is a public lab for AI + crypto experiments.

I build small systems, break them, and document what I learn.

Philosophy: shipping experiments > talking about ideas


> Focus areas:

  • Multi-LLM coordination systems
  • AI-powered crypto analysis tools
  • Blockchain automation experiments
  • Token mechanics & game theory

[!] Not polished. Actively building. Failures documented.

zyel@lab:~$ ls -la experiments/

> active_experiments

● RUNNING

exp_001: Multi-LLM Debate Engine

2026-02-10

Testing multi-LLM coordination for crypto topics. Three models debate bull/bear cases, fourth judges consensus.

llm-coordination crypto-analysis [read_more] [demo]
○ ARCHIVED

exp_000: Token Sentiment Tracker

2026-01-28

Real-time sentiment analysis from Twitter/Reddit. Failed: API rate limits cascade → too expensive.

sentiment-analysis failed [read_more]

→ One experiment per week. Always building.

zyel@lab:~$ tail -f log/latest.log

> experiment_log

[2026-02-15 14:23:41]

Rebooted lab infrastructure. New domain, new aesthetic.

Set up folder structure: /experiments, /log, /playground

→ Next: ship multi-LLM debate engine v2

[2026-02-12 09:15:03]

Multi-LLM debate: Claude stronger reasoning, GPT-4 faster response

Issue: timeout cascade when LLM B slow → added fallback chain

✓ Fixed: cascading fallback with 3-model redundancy

[2026-02-08 22:41:19]

Experiment 000 failed spectacularly. Rate limits destroyed the budget.

Learning: Twitter API v2 is unusable for realtime sentiment at scale

⚠ Archived. Moving to on-chain data instead.


→ Daily updates. Raw notes. No polish.

zyel@lab:~$ ./launch_playground.sh

> interactive_demos

[!] Click to run experiments in your browser. No setup required.


LLM Debate Arena

Watch multiple LLMs argue in real-time

▶ Run on Replit </> View Code

Token Price Oracle

AI-powered price prediction aggregator

▶ Run on CodeSandbox </> View Code

Smart Contract Analyzer

Coming soon...

IN DEVELOPMENT