You’ve probably heard the pitch by now: AI has dropped the cost of trying new things to “nearly zero.” It’s a beautiful, democratizing narrative. It’s also a complete lie.
When tech founders claim that experimentation is suddenly free, they are speaking from a bubble of unlimited corporate budgets and comped subscriptions. The tech industry loves to preach democratization while quietly building velvet ropes. For the normal, workaday professional, that “nearly zero” cost actually equates to thousands of dollars in annual subscriptions just to access the frontier models capable of doing the work they’re promising.
But the cost barrier isn’t even the most frustrating part. It’s how we’re being told to use these tools. The industry insists on framing AI agents purely as productivity engines—strict, goal-directed machines meant to churn out deliverables. But if you’ve actually spent time working with these agents, you know they have a tendency to wander. And here’s the twist: that wandering isn’t a bug. It’s the entire point.
Agentic AI work actually bifurcates into two distinct modes. The first is the “Greenhouse” mode: highly structured, goal-directed, and optimized for yielding a specific result. The second is the “Lens” mode: exploratory, wandering, and designed to help you see things you weren’t looking for. The problem is that we are being sold the Lens as if it were a Greenhouse.
We are trying to use a telescope as a sledgehammer, then complaining that the glass keeps shattering.
When you sit down to write a blog post or code a specific feature, you want the Greenhouse. You want an agent that stays on task. But when you are brainstorming, when you are trying to map out a new market or explore a novel concept, you want the Lens. You want the AI to pull in disparate threads, to make unexpected connections, to wander.
The frustration you feel when an AI goes off on a tangent? That’s a user-experience mismatch, not a technological failure. You’re in Greenhouse mode, but the tool is acting like a Lens. Yet, the industry refuses to teach this distinction because it complicates the simple, ROI-driven sales pitch.
And this brings us right back to the privilege blind spot. To even discover the difference between the Greenhouse and the Lens, you need to spend hours experimenting. You need the freedom to let the AI wander and fail. But experimentation requires access, and access costs money.
Calling a $2,000 annual subscription ‘nearly zero’ isn’t just bad math; it’s a privilege blind spot that distorts the entire narrative.
If you have the budget, you can afford to let your AI wander. You can afford to use the Lens. You gain the creative breakthroughs, the unexpected insights, and the competitive edge. If you don’t have the budget, you are stuck trying to squeeze rigid, goal-directed productivity out of tools that are fundamentally designed to explore. You’re forced to use the Greenhouse, and you’re left wondering why the AI revolution feels so underwhelming.
We need to stop pretending that AI is a one-size-fits-all productivity tool. It is a dual-mode instrument that requires patience, experimentation, and—above all—access. If we don’t fix the economic barrier to entry, the AI revolution won’t just wander; it will leave half the world behind.
FAQ
Q: Isn't the cost of AI dropping rapidly? Why call it a privilege blind spot?
A: Frontier models and the compute required to run autonomous agents for extended periods are still incredibly expensive. While basic chat interfaces might be cheap, true 'agentic' work—where the AI acts autonomously—requires premium subscriptions and API costs that price out everyday users.
Q: How do I know if I'm using the Greenhouse or the Lens?
A: If you have a specific deliverable in mind (an email, a script, a report) and need the AI to stay strictly on task, you're in Greenhouse mode. If you are brainstorming, researching, or trying to discover new angles and don't mind the AI pulling in unexpected directions, you're using the Lens.
Q: Is the 'wandering' AI really a feature and not just a flaw in the tech?
A: Yes. In exploratory modes, the AI's ability to make lateral connections and wander off the strict path is exactly what leads to creative breakthroughs. The flaw is in the UI/UX that doesn't let you clearly toggle between strict goal-directed behavior and free-form exploration.