The Genius With A Thousand AI Agents

If I gave you RM60,000 a year to spend on AI agents, how would you spend it?

I asked myself that question recently. Seriously. Not rhetorically.

Here's what I know about my current setup. I run about 15 agents. Some triggered by me, some scheduled and running in the background without my involvement. They handle things I used to do manually. The list keeps growing as I find new gaps to fill.

I spend about RM500 a month on the tools that power them. This is separate from the AI tools I get access to at work. This is all personal.

Now Jensen Huang recently said he expects his engineers to spend US$250,000 a year on AI tokens. Half their salary.

I cannot conceptualise that number of tokens. It is so far from my reality that I don't even know where to begin imagining what that looks like in practice.

So I scaled it down. Way down. What would I do with RM5,000 a month? RM60,000 a year?

I still struggled to answer.

And I think about this every day. Because if we believe that humancentric AI can augment and support the human, then surely spending that much of tokens would exponentially augment my output.

A few years ago, Sam Altman from OpenAI predicted we'd see the rise of the solo billionaire entrepreneur. One person, completely surrounded and augmented by AI.

But when I tried to picture what that actually looks like for me and ask myself what I would do with 10 times more agents, 10 times more compute, I hit a wall.

The one concrete thing I came up with was coding. I don't code now. With serious AI budget, I could. That told me something. My constraints aren't financial. They're about imagination.

We need to be able to imagine things that we want to do and then figure out how to get AI to do it.

And there comes the next issue.

Jim Collins wrote in Good to Great about the difference between Level 4 and Level 5 leadership. Level 5 being the ultimate in leadership effectiveness.

Level 4 has often been called the genius with a thousand helpers.

Brilliant, effective, and deeply problematic. Because everything depends on the genius. When they leave, it unravels.

Now ask the question again. What happens when the genius has a thousand AI agents?

(I am absolutely not referring to myself as the genius here. Far from it.)

The agents don't leave. They scale. They don't need succession planning. Does that solve the Collins problem? Or create a new one we haven't named yet?

And here's where it stops being a founder question.

This is the dilemma at every level of every organisation. The manager with 50 agents. The analyst whose agents hold the institutional knowledge. The team lead who is, slowly, becoming the genius with a thousand helpers.

We need to talk about this.

So I'll start with the question I couldn't fully answer:

If I gave you RM60,000 a year to spend on agents, how would you spend it?

Views are my own and do not represent the views of my organisation.