Some places do not just host innovation. They shape how it happens.

People often talk about innovation as if it only belongs to a few obvious places. London. San Francisco. Berlin. Somewhere with a polished pitch deck, a rooftop office, and a permanent camera crew filming founders drinking coffee.

That has never felt true to us.

We live and work in the North East of England, and the more time we spend building AI systems and autonomous agents here, the more it feels like one of the most natural places in the world to turn ideas into reality.

Our team has been kicking around a new feature for DMJBot, and one question sits right at the center of it:

Should an AI assistant have a configurable level of initiative?

Could the assistant do something on its own — while you're not around — and could you turn that up or down? Make it very active, or barely active, or shut off all self-started actions entirely?

We think this could be genuinely useful, so before we start building it we want to describe it here and get feedback from the community. Here's what we have in mind.

I would like to share my experience of using DMJBot to automate my daily activities on Reddit.

This is not about spam. I just wanted to save some time on the routine stuff I do anyway. Usually I open a few of my favorite subreddits, read some posts, comment on them, upvote or downvote, and occasionally publish something myself. I use only my own account for this — no multiple accounts, no spam, no bot farm. Just me, doing the same clicks I always do, except now I don't have to.

So when I got DMJBot running, Reddit was the first thing I decided to hand over to it.

My setup

I run DMJBot in Docker on my VPS.

I also connected all my laptops to it, including two old ones I hadn't touched in a long time. Both just sit with closed lids in my utility room, plugged in and connected to my DMJBot instance. One of them runs Windows, and that's the one I use for Reddit automation.

On that laptop I installed the dmjbot-bridge tool and a Chrome browser, and gave DMJBot a way to drive Chrome on it. Before automating anything, I logged into my Reddit account in that browser by hand and confirmed everything worked. From then on, DMJBot uses that browser session — the same logged-in Chrome I would use myself.

For the AI side I mostly use DeepSeek models, but I have all the top models configured too (OpenAI, Google Gemini, Anthropic, and others) and switch between them when I feel like it.

Many developers are already running multiple AI agents and assistants across multiple machines.

A common setup looks like this:

• Claude Code on a Mac for one repository
• Claude Code plus Copilot on another repository
• Codex on a Linux server for automation-heavy work
• Another Claude Code account and Codex instance on a Windows machine

You can absolutely make this work. But after a few days, the real bottleneck appears: management overhead.

You become the human message bus.

You switch between remote desktops, terminals, and chat windows. You repeat instructions. You manually track what was started, what is blocked, and what actually finished.

In other words, you built an AI engineering team, but you still do the project manager job by hand.

People sometimes ask how DMJBot compares to OpenClaw. It's a fair question — the goals look similar. But the two tools were built independently, and under the hood they could hardly be more different.

First, a clarification: DMJBot is not based on OpenClaw. We started building it before OpenClaw was released. We arrived at a similar ambition — an AI assistant that actually does your work — but the implementation is entirely our own.

Here are the two differences that matter most.

1. We don't take over your computer

OpenClaw's model is to own a whole machine. It needs a full computer, it controls it, and it does everything there.

DMJBot takes the opposite approach. The only thing that runs is the brain, inside a Docker container. System requirements are minimal — it's just the orchestrator, not a desktop it has to occupy.

From there, you connect devices to it — one or many — and you decide how much each one shares:

• full control of a device, if you want that, or

Most demos of AI assistants look the same: you type, it answers, you type again. Useful, sure — but you are still the one doing the waiting, the watching, and the wiring.

This post is different. It is a real, end-to-end story of DMJBot doing an actual piece of my work — from "is this even ready yet?" all the way to "here is the finished report" — while I was away from my laptop. No babysitting. Five screenshots, one continuous flow.

Here is exactly what happened.

The setup

I had a Jira task, SCRUM-1, waiting on me: an experiment with group communication — adding emotional status emojis to messages. But I couldn't start it yet, because the final requirements were coming from a colleague, Roman Doubush, and he was going to send them over Slack "soon."

Classic blocked-work limbo. Normally that means I keep one eye on Slack all afternoon. Instead, I handed the whole thing to my assistant.

Step 1 — "Has the data arrived yet?"

First I just asked a simple question: did Doubush say anything about the task requirements in Slack recently?

The assistant didn't guess. It went and checked — read the channel messages, listed the users, scanned the recent DMs — and came back with a clear answer:

No — nothing about task requirements. The only recent DM from Doubush says: "Will you be at the call today?" So no signals about requirements being ready or anything SCRUM-related yet.