There’s a lot of noise around AI agents right now. Tools like Moltbot, (formerly Clawdbot) an open-source agent that runs on your own machine and connects to WhatsApp, Telegram, or Discord… are genuinely impressive. And genuinely concerning!
The security advice tells you everything you need to know about where we’re headed: run these agents on a separate machine. A Mac Mini, a VPS, a Raspberry Pi. Keep them away from your main system because they’ll have access to your private data.
That’s sensible advice. It’s also a massive red flag about who gets to use these tools.
The Setup Problem
Getting Moltbot running isn’t a quick download. You need command line knowledge, an understanding of API keys, the ability to configure servers or local environments, and a separate Claude or OpenAI subscription on top of everything else.
For people comfortable with that world, this is already transformative. They’re using agents to manage email and calendars, check themselves in for flights, control smart home devices, build websites from their phones, and run autonomous coding loops that work while they sleep.
For everyone else? The door is closed.
This Isn’t About Intelligence
I want to be clear about something. When I talk about the gap between people who can use these tools and people who can’t, I’m not making a statement about capability or worth. Chefs, nurses, carpenters, teachers… none of them are less intelligent than software developers. They’ve just spent their time mastering different skills.
Technical literacy is a specific kind of knowledge. Most people don’t have it because they’ve never needed it. Their jobs, their lives, their interests pointed them elsewhere.
The problem is that AI agents are about to make that specific kind of knowledge incredibly valuable and leave everyone without it further behind than they already were.
Two Classes
We’re heading toward a world with two kinds of workers.
One group can deploy AI agents to multiply their output. They can automate the tedious parts of their jobs, build custom tools on the fly, and wake up to find that tasks have been completed while they slept. Their productivity compounds.
The other group watches this happen from the outside. They use the same consumer apps they’ve always used. They don’t fall behind in absolute terms, but the gap between them and the first group widens every month.
That’s not a comfortable future to think about.
What Technology Is Supposed To Do
The promise of technology, especially open source technology, has always been democratisation. Lower the barriers. Give more people access to more capability. The internet did this. Smartphones did this. Free software did this.
AI agents, right now, are doing the opposite. They’re creating a new barrier that sorts people into haves and have-nots based on whether they can navigate a terminal.
What Needs To Happen
We need user-friendly interfaces that hide the complexity. No-code solutions that let anyone build agents for their own needs. Companies investing in accessibility as much as they invest in features.
This isn’t about dumbing things down. It’s about meeting people where they are.
Because if these tools stay locked behind technical gatekeeping, we’re not building a better future. We’re just building a more stratified one.
The Question
I don’t know how this plays out. Maybe the interfaces get better fast enough that everyone can use them. Maybe they don’t, and we end up with a permanent productivity gap based on who happened to learn the right skills at the right time.
What I do know is that the people building these tools have a choice. They can optimise for power users and let everyone else figure it out, or they can treat accessibility as a core feature from the start.
The technology is exciting. The possibilities are real. But so is the risk that we leave most people behind.
Is that the world we’re building? Or just the one we’ll get by default if we’re not careful?
These are interesting times. Exciting, yes. But also worth thinking about carefully.
Leave a Reply