"What About Free Tools?" — The Hidden Cost of Zero-Dollar Subscriptions

You downloaded the hex constraint PDF, ran the audit, and hit something that felt like a loophole. Half the tools on your list are free. Google's Gemini free tier. Perplexity's basic plan. ChatGPT's free access. The logic goes: if the hex is about cutting unnecessary subscriptions and reducing spending, then free tools shouldn't count. They cost nothing. They're not bloating your budget. Why would you cut something that's free?

This is one of the most common objections I get, and it reveals a misunderstanding about what the hex constraint is actually constraining. The hex isn't a budgeting exercise. It's a bandwidth exercise. And free tools consume bandwidth at exactly the same rate as paid ones — sometimes faster, because the absence of a price tag removes the last remaining friction that might make you pause before adding another tool to your life.

The Cost That Isn't on Your Credit Card

Every tool you use — free or paid — costs you three things that don't show up on a billing statement.

The first is attention. Every tool in your stack is a tab you might open, a notification you might check, a workflow you might consider using for the task in front of you. When you're writing and you have one LLM, you write in that LLM. When you have three — Claude Pro, free ChatGPT, and free Gemini — you spend 10 minutes deciding which one to use before you write anything. That decision feels instantaneous, but it's not. It fragments your focus. It introduces "what if the other one would do this better?" into every task. And it compounds across every working hour of every day.

The second is learning surface. Every tool has its own interface, its own quirks, its own way of doing things. Gemini's grounding works differently from Perplexity's citations, which work differently from ChatGPT's browsing. When you use all three, you're maintaining fluency in three systems. You're never deep in any of them. You're always at the surface level — good enough to get a result, never good enough to get the best result. The person who uses one research tool and learns its advanced features will outperform the person who uses three research tools at a basic level. Every time.

The third is maintenance. Even free tools require account management, preference configuration, conversation history, and the ongoing mental overhead of keeping track of what you asked where. I've talked to people who have the same question answered in three different tools because they forgot they already asked it. That's not efficiency. That's entropy — with a zero-dollar price tag.

Why Free Tools Accumulate Faster Than Paid Ones

Paid tools have a natural friction point: the moment of purchase. When you're about to spend $20/month on a new AI subscription, some part of your brain does a cost-benefit calculation. Not always a good one — plenty of subscriptions get purchased on impulse — but the friction exists. You feel the money leaving. That feeling creates at least a moment of hesitation, and hesitation is where decisions happen.

Free tools have no such friction. Signing up takes 30 seconds. There's no billing page, no confirmation email about charges, no credit card entry. The barrier to adoption is essentially zero, which means the barrier to accumulation is also zero. I've worked with people whose honest tool count was 14, and when we ran the audit, eight of those tools were free. They'd never made a deliberate decision to adopt any of them. They'd just... signed up. One Twitter recommendation here, one "try our free tier" prompt there, and suddenly their workflow involved five different AI tools that all did roughly the same thing at slightly different quality levels.

The free tier is also a deliberate product strategy, and it's worth being honest about what that strategy is. Companies offer free tiers to get you using the product so that when you hit the limit — the usage cap, the context window restriction, the feature gate — you upgrade to paid. The free tier is the first hit. It's designed to be useful enough that you keep coming back, but limited enough that you eventually want more. When you add a free tool to your hex without counting it, you're not outsmarting the constraint. You're participating in a conversion funnel that's designed to turn free users into paying users. The $0 on your bill today is the $20/month on your bill in three months, once you've built enough workflow dependency that canceling feels like a loss.

The "Second Opinion" Trap

The most common way free tools infiltrate a constrained stack is the second opinion pattern. You have Claude Pro as your primary LLM. It's your hex slot for thinking and drafting. But you also keep ChatGPT's free tier open "just to check Claude's answers sometimes." And maybe Gemini too, because it's connected to Google Search and "sometimes you want grounded results."

On paper, this sounds reasonable. Different models have different strengths. Cross-checking is good practice. But in practice, what happens is this: you ask Claude a question. You get an answer. Then instead of evaluating that answer and moving on, you paste the same question into ChatGPT. You get a slightly different answer. Now you have two answers and you need to figure out which one is right. So you paste it into Gemini. Now you have three answers. You spend 15 minutes comparing them, decide the truth is probably somewhere in the middle, and write something that's a synthesis of three mediocre responses instead of one response you actually engaged with critically.

The second opinion pattern feels like rigor. It's actually abdication — outsourcing your judgment to a consensus of language models. If you can't evaluate whether Claude's answer is good without checking two other LLMs, the problem isn't that you need more LLMs. The problem is that you're not engaging critically with the output of the one you have. The fix is learning to evaluate — checking sources, testing claims, applying your own domain knowledge — not adding more voices to the committee.

How to Audit Your Free Tools

The same three-question audit from the "But I Need More Than Six Tools" article applies to free tools, with one addition.

For each free tool, ask:

What specific output does this produce that my paid tools can't? If free Gemini does research and Claude Pro also does research, you have overlap. The fact that one is free doesn't make the overlap productive.

When did I last use this to produce something I shipped? Not "opened it," not "tested a prompt" — produced something real. If the answer is more than 30 days ago, the tool is occupying mental space without producing value.

Can any other tool in my stack do this at acceptable quality? Not identical quality — acceptable quality. If Claude handles 90% of what you use Gemini for, the 10% delta isn't worth the bandwidth cost of maintaining a second tool.

And the additional question for free tools specifically: Am I using this because it's the best tool for the job, or because it's free? This is the honest one. If a paid tool does the same thing better and you'd choose it over the free one in a head-to-head comparison, the free tool isn't earning its hex slot on merit. It's earning it on price — and in the hex, price isn't a valid criterion.

The Exception: Free Tools That Fill a Genuine Gap

Not every free tool is a trap. Some free tools genuinely fill a capability gap that nothing else in your stack covers. If your hex includes Claude for thinking, Cursor for code, and n8n for automation, and you add Perplexity's free tier because none of your other tools do source-grounded web research as well — that's a legitimate hex slot. The tool does something distinct, you use it deliberately, and the fact that it's free is incidental to its value.

The test is the same test you'd apply to a paid tool: does it produce unique output, do you use it regularly, and does it earn its slot based on capability rather than convenience? If yes, count it. If you'd keep using it even if it cost $20/month, it belongs in the hex.

But if the honest answer is "I'd cancel it if they started charging" — that's your answer. If a tool isn't worth $20/month of your money, it's not worth a hex slot of your attention. The hex values your attention more than your money. Your attention is the scarcer resource.

The Real Constraint

The hex constraint says six tools. Not six paid tools. Not six expensive tools. Six tools — because six is the number that forces you to choose, and choosing is the entire point. When free tools don't count, the hex becomes a budgeting exercise: "how do I minimize my AI spending?" That's a fine question, but it's not the hex question. The hex question is: "what are the six tools that produce the most value for my specific work, and how do I go deep on those six instead of skating across twenty?"

Free tools count because the cost of a tool isn't measured in dollars. It's measured in the attention you give it, the decisions it forces you to make, the learning surface it adds to your life, and the workflow complexity it introduces. A free tool that you use daily and that shapes how you work is occupying exactly as much of your bandwidth as a $200/month enterprise subscription. Possibly more — because at least the enterprise subscription triggers a quarterly "is this still worth it?" review. The free tool just sits there, uncounted, unexamined, consuming bandwidth nobody's measuring.

Count your free tools. Run the audit. Most people discover they have two or three free tools doing work that one tool — free or paid — could handle. Cutting those redundancies doesn't just simplify the hex. It gives you back the decision energy you were spending on "which free LLM should I use for this?" and lets you redirect it toward the work the tools are supposed to help you do.

This article is part of the Hex FAQ series at CustomClanker.

Related reading: What Counts as a Tool, The Hex Explained, But I Need More Than Six Tools