One of the reasons that there are always so many new ad formats popping up is that our brains identify them as invasive, and we develop an immunity to them over time. Creative advertisers find ways of concocting new signals to assault our sentience, to get at our wallets and perform a Money.
Much in the same way we have set up mental barriers to protect ourselves from advertising, I find that I have erected some communication defenses from AI-generated content.
Let me illustrate that by asking AI to write the opening paragraph for this blog.
I suppose a trigger warning is warranted here.
Trigger warning! AI-generated text:
Advertising has always been an arms race between persuasion and pattern recognition…
A new format appears and works brilliantly for a while. Newspaper ads looked like information. Banner ads looked like navigation. Native advertising looked like journalism. Then audiences learn the pattern, build resistance, and the format loses much of its power.
We may be watching the same thing happen with AI-generated writing.
The first wave of LLM prose felt remarkable: polished, articulate, and seemingly effortless. But exposure creates familiarity… Readers are beginning to recognize the rhythms, structures, and telltale phrases of machine-generated text in much the same way they learned to spot clickbait headlines and sponsored content.
It is not a question of whether AI can write anymore… it is a question of whether readers can tell that it did.
It is not an arms race for generation… it is an arms race for authenticity.
And prompting is rapidly becoming more than getting an AI to write well. It is becoming the art of producing writing that can survive an audience increasingly fluent in the fingerprints of its creation.
I am sorry for shitting directly into your brain. I don’t feel good about doing what I just did, but I think you get the point. There is something about my warm and friendly Szeder-esque opening, and how it differs from the schlip schlop emitted from an LLM that you can spot from the International Space Station.
You might also be momentarily confused… Because generally I am here with the “hooray LLM” flag while sitting at the “I heart LLM” fan club officer table.
Yes, I do love me some LLMs and tools to help make things faster and better. At the same time, you might notice me posting obnoxious jokes about the random cold call attempts I get, often on behalf of LLM tools, on LinkedIn.
Let me help you hold these two apparently opposite ideas in your head at the same time. While I love me some “let’s get this done thirty percent faster” technology, I want that extra time to be spent at a higher-quality level.
I do not want to spend that extra time sifting through someone’s promptables. I have been on the receiving end of documents like this, and I will confess that I am about to give Sam Altman and his point zero one percenters a massive boner. The only way to deal with someone who uses a big-assed old prompt to send you a voluminous document of fewmets and digital offal is to do one of those reverse ju-jitsu things. Take their output from an LLM that has plopped itself into the bowl of your inbox, feed it back into an LLM, and ask for it in summary form; We must fight fire with fire.
I have developed a reasonably good feeling for when something coming my way is authentically written and also when it is “absolutely right!”
Cold calling is not the best thing for an LLM to do, and my inbox is littered with deliberately misunderstood replies as I have attempted to use humor to preserve my sanity from the onslaught. I find that being authentic online is best when you are akshully being authentic, not having an LLM pull a “how do you do, my fellow kids?”
Sam Altman ain’t getting paid for making fetch happen. LLMs and cold calling go together like peanut butter and dynamite.
I am not going to hop on a soapbox and attempt the moral high ground argument around the environment, or the cost, or the … whatever it is people cry about as they tattoo “FIRE MY ASS” on their foreheads as they pop off on LLM technology.
I will say “Yes, this is one thing it does poorly” and then move on like an adult, capable of agreeing to disagree on the internet’s vast ocean of (mostly wrong) opinions. I won’t scream “SOURCE?!” in your face or hold up your arguments and counter assault with the fifth grade’s best ad hominem attacks.
I mean… “I know you are, but what am I?” In 2026? We can do better.
All I want to do is make sure you know that some people love communicating with other people. And I am one of those people. Attempting to impress me by a “save as” PDF from Markdown with Claude colors draping the document is not going to get very far with me if you want to impress me or do a Money from me. Just no. Stop.
At the very least, I want you all to know these people are out there, and it is okay to take what they send you and feed it into an LLM to ask “what means?” or equally toss it in the garbage and pick up the phone.
That is it. That is the whole post.
Speak to you all again soon.
