📖 Token Saving Series (4 parts): Part 1: Why your AI gets dumber the longer you chat · Part 2 (This post) · Part 3: Clearing a clean desk for your AI · Part 4: I built a budget dashboard for my AI
In the previous post, we talked about how longer AI chats get more expensive and dumber, along with three ways to save right now. This post takes it a step further to discuss a more fundamental divide:
Some jobs simply shouldn’t be given to AI.
Sounds a bit contrarian, but this is exactly my core mindset for using AI cheaply and accurately.
Two tools: the abacus and the supercomputer
There are actually two completely different paths when handing things off to a computer.
One is hardcoded scripts. You figure out the rules upfront, write them as steps, and then it just runs them. Like an abacus, the rules for moving the beads are fixed. Calculate ten or a hundred times, the answer is always the same.
The other is making AI think (LLMs). You describe what you need, it “understands,” and then generates an answer. Like a supercomputer, it can take on anything vague or requiring judgment, but every time you boot it up, it burns power.
Many people’s instinct is: AI is so strong now, just throw everything at it.
This is exactly the root of burning money and time.
One table to understand the tradeoff
| Aspect | Hardcoded script (Abacus) | Making AI think (Supercomputer) |
|---|---|---|
| Accuracy | 100% certain. Same input always yields the same result. | Drifts. Asking the same question twice might give different answers, or it might hallucinate with a straight face. |
| Handling ambiguity | Can only do things with clear rules. | Can handle semantics, judgments, exceptions, and that “you know what I mean” gray area. |
| Cost | Practically zero. | You pay for tokens every time it runs. |
| Speed | Milliseconds. Done in a blink. | Seconds, and it gets slower the longer the chat. |
| Prep work | You have to think through and write the rules correctly first. | Just speak and use. Zero setup. |
See it? Neither is inherently better; it’s just about which fits the task.
Things with clear rules that repeat a lot go to the abacus: fast, accurate, and free.
Things that need judgment, are vague, or differ slightly every time—that’s when it’s worth booting up the supercomputer.
A real example: how I read my medical textbook PDFs
I have a lot of textbook PDFs, and I often need to fish content out of them.
If I just threw every page straight at the AI, that’s treating the supercomputer like a printer: paying for “looking at the image plus reading the text” on every single page. It’s terrifyingly expensive (as I’ll explain in a later post, throwing a PDF directly at AI can burn a thousand or two tokens per page).
My approach is layered:
- First, use a free little script to extract the text directly from the PDF. This step is pure abacus, zero cost.
- If it doesn’t extract cleanly (scans, handwriting), only then do I upgrade to using optical character recognition (OCR) on my local machine. This runs on my own computer and doesn’t cost a dime in AI fees.
- I only send the image to AI when I truly need it to “understand what this diagram is saying.”
What’s better is that the judgment of “did this page extract cleanly?” is also done by a script: count the word count, look at the gibberish ratio, check the fonts. The rules are clear, no need to bother AI at all.
The result: 99% of the workload gets eaten by the abacus. The supercomputer only shows up for the 1% that genuinely requires judgment.
So what exactly did I turn into scripts?
Take a few of my real-life examples, and you’ll find they have something in common: clear rules, repetition, or a need for precision.
- Those “interceptors” are scripts themselves. In part three, I’ll mention the little gadgets I attach to AI, like “discard the output for successful commands” or “ban throwing PDFs to AI as images.” These judgments are entirely hardcoded scripts. AI doesn’t intervene at all.
- Math and stats must use scripts. This is crucial: AI is actually really bad at precise calculations. Ask it to calculate a long string of numbers, and it might get it wrong while sounding completely authoritative. My return on investment (XIRR) and various spreadsheet calculations are all handed off to scripts. The answers are precise and free. Never ask AI to do mental math for things that need calculating.
- File conversions. Converting PDFs to text, Word docs to my notes, extracting audio tracks from video, batch resizing images—these all have standard workflows. Write a script to run them, zero AI cost. (Yes, even the four cover images for this series were resized at once with a one-line script.)
- Scraping data on a schedule. I have a few little bots keeping an eye on rehab course registrations, new papers, and even typhoon day-off announcements. They run automatically on a schedule and only notify me if there’s a change. These are pure script-based scheduled jobs, no need for AI to stay awake and stare.
- Boring batch chores. Renaming a bunch of files at once, moving folders, tidying up lists—anything where the rules can be explained in one sentence gets hardcoded.
See the line there? As long as a task has “clear rules,” “repeats,” or “requires precision,” it should be a script, not AI. Only after handing all of this over to scripts does AI have the bandwidth to focus on what it’s truly good at: judging, understanding, and generating.
“The cheapest call is the one you don’t have to make”
I really like this saying. Every time you make AI think, it’s an inference tax.
GitHub published some data themselves: they pulled out a “clear rules, no AI judgment needed” step from an automated workflow and used a script instead, and their token usage dropped by 62%. The exact same task, done the exact same way, just giving the abacus parts back to the abacus.
This is also why, when people ask me “Will AI replace a lot of jobs?”, my feeling is: the strongest parts of AI actually make me see more clearly “what things really don’t need AI at all.” Giving the right job to the right tool—that’s the real skill.
Summary
- Clear rules and repeats? Use a script. Fast, accurate, and free.
- Vague and needs judgment? Then use AI.
- A complex task should usually be broken apart: hardcode what can be hardcoded, and only hand the small chunk that needs brainpower to AI.
Starting in the next post, I’ll walk you through my “heavy usage” setup on my computer, starting with how to clear a clean desk for your AI. That’s where saving tokens really shows its mettle.
—
⬅️ Previous: Why your AI gets dumber the longer you chat ➡️ Next: Clearing a clean desk for your AI
References
- GitHub Blog (2025). Improving token efficiency in GitHub Agentic Workflows.