# AI Agent Lab: LLM Battles, Demos and Weird Tests

Canonical page: https://2bcome.ai/lab/
Content updated: 2026-09-15

THE LAB

# We build weird things on purpose.

Not every useful experiment belongs in a product roadmap. The Lab is where we make agents argue, compete, coordinate, fail, and surprise us.

[Open LLM Battle](https://2bcome.ai/lab/llm-battle/) [Product experiments](https://2bcome.ai/experiments/)

FEATURED / 001

## Two models enter. One goal leaves.

A visual replay demo for an agent battle: models take turns, modify shared state, and keep going until a stop condition fires.

[MODEL A QWEN Planner / builder TURN 18 / DEMO Build the game. Play the game. Win the game. Local replay. No live model call. MODEL B DEEPSEEK Critic / adversary Open interactive replay](https://2bcome.ai/lab/llm-battle/)

WHAT BELONGS HERE

## Curiosity before commercial usefulness.

Lab projects can become products, blog posts, or dead ends. All three outcomes are useful.

NEXT / 002

### Agent telephone

How badly does a goal mutate when it passes through five agents that cannot see each other's reasoning?

IDEA

NEXT / 003

### Self-editing instructions

Give an agent permission to improve its own operating instructions, with a diff and hard rollback.

IDEA

NEXT / 004

### Four models, one company

Give four local models distinct roles and see whether coordination becomes the real bottleneck.

IDEA

LAB ≠ PRODUCT PROOF

## Fun tests stay honest.

Lab demos are playgrounds. Product experiments have baselines, protocols, and measured outcomes. We keep the two separate.

[See product experiments](https://2bcome.ai/experiments/)

CONNECTED WORK

## The next useful connection.

EXPLORE

### [LLM Battle: Two Models, One Goal](https://2bcome.ai/lab/llm-battle/)

Play through an interactive 2bcome Lab demo showing two AI models taking turns against the same ambiguous goal, shared state, and stop condition.

[Explore](https://2bcome.ai/lab/llm-battle/)

FIELD NOTE

### [AI Agent Blog: Opinions, Guides and Build Notes](https://2bcome.ai/blog/)

Read the 2bcome Blog for direct opinions about AI, practical agent guides, and build notes from creating autonomous systems in public.

[Explore](https://2bcome.ai/blog/)

EXPERIMENT PLAN

### [AI SEO Experiments: Plans and Measurement Methods](https://2bcome.ai/experiments/)

Follow the proposed 2bcome experiments for agent-led website improvements, internal links, and content refreshes. Plans are published before results.

[Explore](https://2bcome.ai/experiments/)