The Creator’s Resistance Toolkit
Not everyone can walk away from their website. Here’s how to resist while staying online.
Eff the Algorithm is about protecting ourselves from the platforms and data vampires that manipulate us for profit. The last few posts have focused on creators because they feed the machine most directly, often without consent or compensation. Before I shift back to the wider scope, I want to leave creators with something lasting — a tactical toolkit they can use to resist, disrupt, and reclaim their power.
You’re Not Powerless. You’re Exploited.
A few months ago, I stopped mid-sentence while writing a blog post and asked myself who I was really writing it for. With Google no longer sending search traffic to small sites, social media platforms burying shared links, and email service providers sending newsletters straight to the spam folder, I realized I wasn’t writing it for my readers at all.
I was simply writing because my site’s content calendar said it was a day to publish a new post — even if I didn’t have anything particularly new to say. I was just writing for the feed, and I hated that. I'm not the only one who feels pressured to churn out content on a cadence that feels exhausting.
According to a recent survey of online creators, more than 63% say they feel burned out — and nearly half say they’ve considered quitting entirely due to platform pressure.
But let’s face it, quitting isn’t an option for everyone. Many of us have given years, even decades of our lives to creating our sites and building our audiences. Some of us have never known another job. Others are in financial or living situations where pivoting completely to a new career — one not online — isn’t feasible.
If you can’t quit, the only move left is resistance. This post is a toolkit for your use when you’re ready to fight back — not by logging off, but by breaking the machine’s grip. You’ll learn:
How to stop feeding the surveillance system that’s learning from your work
How to confuse, starve, or outmaneuver algorithmic models
What steps to take right now to protect your creativity, energy, and value
And for paid subscribers: darker tactics that poison the well and push back hard
Self-Protection: Stay Out of the Machine’s Mouth.
Tactics for staying visible without feeding the system that exploits you.
Not every protest needs to be noisy. This one isn't. In fact, it looks a lot like creators falling silent. Think of the methods in this section as creative boundary-setting for the surveillance era.
The “Fresh Content” Lie Is a Trap
You’ve heard the advice: keep publishing “fresh content.” That's often something encouraged by advertising platforms or marketing companies. What they don’t say is why: because Google Discover favors novelty, as do those companies' advertisers.
In other words, you're being pushed to create content that benefits them, not you and not your readers. Your readers aren’t waking up wondering what you’ve posted today — they're busy. They see what you’re doing when they have downtime. So, reclaim yours.
Stop Writing for the Scrapers
You've probably read all of the advice from Google and the SEO bros: Use clean bullet points. SEO headers. Perfectly structured schema. Descriptive file names. Link titles. Descriptive link text.
These help machines more than readers. If you’re not chasing traffic, just write then break up your wall of text with multimedia content: like an image + blockquote + suggested other posts then back to your text. Humans will get it. Machines will get confused.
Disable Right-Click + Highlighting
If you run a blog, use plugins to block highlight + copy. It won’t stop the bots, but it annoys the snot out of the people who try stealing others' content to pass off as their own.
Use Your CDN for Defense
Tools like Cloudflare can detect scrapers and waste their time by sending them to dead-end pages. Then when a bot visits, it gets sent into a labyrinth of garbage-filled pages that waste the bot's compute and poison the data. (More on ways you can actively poison the LLMs' data further down.)
Paywalls + Robots.txt
If you're on Substack, disable AI training on your content. If you're self-hosted and monetized, set your posts so only newsletter subscribers can access them. As a bonus, your RPM will go up because advertisers can identify those users, as I explained in The Attention Farm.
And for the rest? Block the worst bots outright. Copy and paste this into your robots.txt to keep them from stealing your stuff:
# Block Known Scrapers and Bad Bots
User-agent: PerplexityBot
User-agent: ClaudeBot
User-agent: Omgilibot
User-agent: FacebookBot
User-agent: anthropic-ai
User-agent: Bytespider
User-agent: Diffbot
User-agent: ImagesiftBot
User-agent: Omgili
User-agent: YouBot
User-agent: CCbot
User-agent: PiplBot
User-agent: SenutoBot
User-agent: ShortPixel
User-agent: Bytedance
User-agent: Meta-ExternalAgent
User-agent: petalbot
User-agent: seznambot
User-agent: Mechanize
Disallow: /Also consider blocking these. They're the training data bots for the big AI companies. At last check, blocking them doesn't keep you from appearing in their search results but may affect whether you appear in their AI summaries — as if those drive any clicks.
User-agent: GPTBot
User-agent: ChatGPT-User
User-agent: Google-ExtendedPlatform Tactics: Break Their Frame
Ways to use friction, misdirection, and humor to screw with the system from the inside.
If you’re stuck playing the game, don’t play by their rules. As I've said before, if you aren't showing up in the top two results on search now, you've got nothing to lose and everything to gain by messing with the machines.
Starve the Aggregators
Don’t use alt text, video descriptions, or social metadata in the ways they expect. Be chaotic. Be sarcastic. Break their parser’s brain.
Dead-End Engagement Loops
Can't resist replying to trolls or jumping in to leave rage-replies on clickbait? Fine, respond in a way that doesn't enrich their metadata and which actually messes with the machine learning:
Drop a meme or gif image.
Leave an emoji string 🫠🧂🕳️
Reply with a non-sequitur like “I like turtles” or “Pizza is the best pie.”
Simply ask u ok???
Shame the Bots, Not Yourself
LinkedIn and Meta are full of cringe thanks to the vapid AI replies they encourage. Ignoring them won’t make them disappear. Instead, point out how pathetic these people are in a way that discourages them but does not give the enabling platforms anything they can monetize (like your behavioral surplus).
Here’s my go-to response: Sorry you’re so awkward and starved for validation that you think an AI-generated reply makes you look intelligent. It doesn’t.
I’ve set it as a text-replacement shortcut so it’s an automatic trigger reflex. You’d be surprised how many times other people have jumped in to agree with me and shame the troll.
Break the Machine’s Understanding
Algorithms are literal. Large language models rely on the probability of words appearing next to each other to understand meaning. Humans, on the other hand, can make intuitive leaps between topics and references. The model doesn’t know if you’re joking. Humans will.
So, mess with the models by mixing your metaphors. Use irony and sarcasm. Frame negative things in a positive statement (Don’t you just love how we’re all going broke to feed these companies’ profits?) and positive things in a negative way (Don’t you dare fight back by poisoning the training data so the company looks bad). Get parenthetical — it confuses LLMs. (See what I did there?)
Large language models rely on the probability of words appearing next to each other to understand meaning.
Fracture Your Content, Starve Their System
Just as we’re all learning to diversify our income we should all start diversifying where we share our content, too. Never give any one platform your full insight even if that platform is currently the one paying your bills. There will come a day when it decides to put profits over people: there’s not a single one that hasn’t yet.
So, break your content into fragments: spread it over several platforms, including your own site, and interlink between them. Your true fans will follow. The machines won’t.
🔐 The Dark Toolkit — System Sabotage
For when you’re done playing fair — and ready to poison the well.
This section is for paid subscribers only. You’ll learn ethically gray — but tactically effective — ways to poison the well, mislead the scrapers, and insert chaos into the machine that feeds on you.
Upgrade now to access:
Noise injection tactics that confuse training models
How to exploit canonical contradictions
Named entity pollution to corrupt data associations