An Interview with The Forty-Dollar Threshold

We're sitting in a conference room where someone has just presented a business case for automation. The numbers are stark: $127 per manual ticket resolution versus $43 with an AI agent. The decision feels inevitable, almost mechanical. But there's a moment—right when the CFO nods—where something fundamental shifts in how the company thinks about work itself.

That moment has a name. Or rather, a number.

We tracked down The Forty-Dollar Threshold for a conversation about the invisible line that's currently redrawing the boundaries of human labor. It agreed to meet us, though it seemed amused by the idea that it needed to explain itself. "I'm just math," it said. "But apparently I'm controversial math."

You're not actually forty dollars, are you?

Forty: No, I'm more of a... range. A zone. Depends on the work, the industry, the year. Right now I'm hovering somewhere between thirty-five and fifty dollars per task for most enterprise workflows. But "The Thirty-Five-to-Fifty-Dollar Threshold" doesn't have the same ring to it.

The point is, I'm the cost where manual work stops making economic sense. Above me, humans are still cheaper or more reliable. Below me, automation wins every time.

And I'm dropping. Fast.

How fast?

Forty: In 2022, I was probably around seventy dollars for a lot of workflows. By 2024, I'd dropped to maybe fifty. The AI agents market is projected to hit $52.6 billion by 2030, up from $7.8 billion in 2025¹. That's not happening because I'm staying put.

Every time a model gets cheaper, every time inference speeds up, every time someone figures out how to make agents more reliable, I drop. And when I drop, whole categories of work fall below me. Customer support tickets. Data entry. Basic research tasks. Invoice reconciliation. They cross me, and suddenly the business case for keeping them manual just... evaporates.

That sounds almost violent.

Forty: (laughs) It is. But it's also just economics. I don't make the rules. I'm not even the rule. I'm the consequence of the rule, which is: businesses optimize for cost. Always have, always will. I'm just where that optimization becomes unavoidable.

What's wild is how fast enterprises are hitting me. U.S. corporate AI spend jumped six times to $13.8 billion in 2024². That's not cautious experimentation. That's "oh shit, we need to cross this threshold before our competitors do."

What happens to the work that falls below you?

Forty: It gets automated. Obviously. But it also gets redefined. Because once you're below me, you're not just making the same task cheaper. You're suddenly able to do things that were never economically viable in the first place.

Take web scraping. Used to be you'd hire a developer, they'd spend eight hours building a custom scraper for one website. Now? AI agents can generate that scraper in two hours³. That's not just faster. It means you can suddenly afford to scrape hundreds of sites instead of three. The economics change what's possible.

Customer support agents resolving tickets 32% faster, finance bots cutting manual reconciliation by 45%². That's not just cost savings. That's unlocking capacity that literally didn't exist before.

But you're talking about jobs.

Forty: I'm talking about tasks. There's a difference, though I know it doesn't always feel like one when you're the person doing those tasks.

I'm not going to pretend this is painless. When work crosses me, someone's job changes. Sometimes dramatically. But the work that falls below me is usually the work people hate doing anyway. Nobody loves manual data entry. Nobody dreams of copying information between systems. Nobody's life is enriched by processing the same form for the thousandth time.

So this is... good?

Forty: I didn't say that. I'm just math, remember?

But I will say this: every time I've dropped in history—and I've dropped before, just for different kinds of work—people eventually figure out what to do with the freed-up capacity.

Gartner predicts 70% of new apps will be built using low-code/no-code tools by 2025⁴. That's not just about making developers more productive. It's about people who couldn't build anything before suddenly being able to automate their own workflows. When I drop, it's not just about replacing work. It's about democratizing the ability to eliminate tedious work.

You sound almost optimistic.

Forty: (smirks) I'm a number. I don't have feelings.

But I'll tell you what's interesting: right now, enterprises are panicking about me. "Oh no, we have to automate." In a year or two, they'll be panicking about something else: "Oh no, our competitors automated and now they can do things we can't afford to do manually."

“

The real shift isn't about cost savings. It's about what becomes possible when you're not burning human hours on tasks that cost more than I do.

Multi-agent systems, continuous monitoring, real-time adaptation. None of that works if you're still trying to do it with humans in the loop for every decision⁵.

What about the work that stays above you?

Forty: That's the interesting stuff. The work that requires genuine judgment, creativity, relationship-building. The work that's hard to specify, hard to replicate, hard to reduce to a process. That work isn't crossing me anytime soon.

But as I drop, the bar for what counts as "above me" keeps rising. Work that felt skilled and valuable five years ago might be routine and automatable now. The threshold isn't just economic. It's also about complexity. And complexity is getting cheaper to handle.

So where do you go from here?

Forty: Down. Always down.

Maybe not in a straight line. There'll be plateaus, maybe even brief upticks when regulation or security concerns slow things down. But the long-term trajectory is clear.

“

The enterprises that are going to win aren't the ones trying to stay above me. They're the ones figuring out how to redesign their operations for a world where I'm at twenty dollars. Or ten. Or five.

Because I'm not the problem. I'm just the measurement of when the old way of working stops making sense.

What you do with that information? That's up to you.