What If Money Was No Object?

The CEO said on the earnings call my team's big roadmap item was coming this quarter. The team's estimate didn't match the promise. The CFO asked me: "what if money was no object?"

It was about 10 years ago and I was on the hot seat in front of 30+ Directors, VPs, and execs in a monthly operational review. I did my best then to explain what we learned 50+ years ago in the Mythical Man-Month, which is that we couldn't just helicopter in more teams and get it done faster. Adding more people to a late project makes it even later.

Now, you can throw agent loops, subagents, and swarms at problems and burn lots of tokens very quickly to get things done. If you're good at directing them, you can get big projects done much faster and maintain a high standard of quality. It is possible - I've got the receipts.

I've fully rebuilt the architecture of an existing product in 6 months, mostly by myself. That's a two year, full team project in the before times. I'm wrapping a full new integration in days - something that took a month+ in 2024.

I've also spent tens of thousands of dollars worth of list price tokens - $50K+ this year, by myself - but on subsidized plans. Real money out the door has been about $400/month. I don't expect that to last forever.

In places where the subsidies aren't available, getting good at forecasting and monitoring the impact of what we invest our time and tokens in is the way to create leverage.

You can now spend money to get something done 5, 10, 100x faster and then pile on more tokens to increase quality... but not many things are going to deliver 5, 10, or 100x the results.

I've been fixated for years now on better understanding and measuring the software development process.

Today I think serious operators should directly track time and tokens going into the work they're doing and the impact that happens afterwards, whether it's building software or not.

The hard part is choosing and measuring the leading indicators like activation rate, pipeline growth, and other key customer behaviors that correlate with what you want to see on the balance sheet when what you do now takes months or even years to show up there.

Start getting good at this now, or you're going to be stuck managing per team member token budgets and not knowing who's able to drive real results with them. Or worse, blowing through your whole AI budget and not having a clue what mattered.

I'm consulting on directing agents and measuring impact with founder-led companies if you would like some help.

Join the conversation on LinkedIn