Hey {{first_name | there}},

A San Fransico man spent $6,000 advertising an AI chatbot. 

People used it. Asked it random questions like you would with ChatGPT, got an answer and it was all good.

Except there was no “AI” behind it. 

It was just one tired guy named Tucker Bryant behind that ”AI chatbot” sitting at his computer, reading messages one by one and typing back answers for up to 10 hours a day.

This is ChatTJB. And here AI stands for "Average Individual." 

Most people had no idea they were talking to a human.

Bryant calls this "cognitive surrender." The idea that we've gotten so used to trusting AI outputs that we stopped questioning them entirely. I think he's onto something. 

But while Tucker is poking fun at our blind trust in AI, the rest of the world is doing something opposite:

Anthropic quietly changed how Claude works. Every new model now has invisible watermarks into generated text. Images get signed metadata.

It’s because of the EU AI Act's Article 50 that says all AI output should be detectable. Anthropic was the first major provider to roll this out globally. 

On paper, this sounds reasonable. Transparency is good. People should know when they're reading AI-generated content.

But the watermark doesn't really affect meaning, quality, or readability. It just adds a machine-detectable signal hidden in the words. 

Copy and paste it somewhere else, and the watermark might survive. Edit it heavily, and it degrades. 

Which brings us to the obvious question. If someone really wants to pass off AI content as their own, can't they just paraphrase it? Yes. They can. And many already do.

But this isn't just Anthropic. The entire internet is moving against AI slop.

YouTube wiped 4.7 billion views from AI slop channels. Channels got demonetized. Some got terminated entirely. 

LinkedIn blocked billions of AI-generated comments. Their chief product officer said they catch hundreds of thousands of automated comment attempts every single day. They also removed the AI writing prompt that used to show up when you drafted a post. Now it just proofreads. 

Substack partnered with Pangram, an AI detection tool, to let readers scan posts and see how much was written by a human. 

Snapchat joined too. They acknowledged that fully AI-generated content is typically "low-quality" and "repetitive." 

So yes, the crackdown is happening. And I get why.

But here's where it gets messy.

These AI detections are not accurate. Not even close.

YouTube mistook a popular science channel as AI slop and flagged them. But all of their videos were human-made. They had to suffer the consequences of false AI detection.

Infact a 2026 study tested commercial AI detectors on writing and found false positive rates between 43% and 83%. 

That means in the worst cases, 4 out of 5 human-written content got flagged as AI. 

And here's the thing that should worry all of us.

If you use AI to help with your writing, which most of us do now, your content might get flagged. Light editing, paraphrasing, even using AI for grammar checks. All of it can trigger detection systems.

So while platforms are trying to protect us from AI slop, they might end up punishing the people who use AI as a tool, not a replacement.

First, know that using AI doesn't make you a fraud. The problem is not using AI. The problem is relying on it completely without adding your own thoughts, voice, and effort.

Second, be transparent. LinkedIn rewards this now. Substack gives you a space to explain how you use AI. Use it. Readers don't hate AI. They hate seeing bad, generic, and repetitive content.

Third, build your own voice. AI can help you write, but it shouldn't replace how you think. The creators who survive this crackdown are the ones who use AI as a starting point, not a finish line.

Tucker Bryant proved something interesting with ChatTJB. People don't just want answers. They want connection. They want to know that a real person cared enough to respond. 

That's something no AI, no watermark, and no detection tool can replicate.

The question is whether you're building your work around that truth, or just generating content and hoping no one notices.

- Aashish

Nobody became a developer to write documentation. But the docs still need to get written — PRDs, README updates, architecture decisions, onboarding guides.

Wispr Flow lets you talk through it instead. Speak naturally about what the code does, how it works, and why you built it that way. Flow formats everything into clean, professional text you can paste into Notion, Confluence, or GitHub.

Used by engineering teams at OpenAI, Vercel, and Clay. 89% of messages sent with zero edits. Works system-wide on Mac, Windows, and iPhone.

Try Wispr Flow free