ChatGPT vs Claude: I Ran a Gordon Ramsay Test to See Which AI Actually Wins
Plus a 5-step playbook you can run in 10 minutes to test it yourself.
Just like most people, I started using AI with ChatGPT for everything.
Brainstorming. Writing. Strategizing. Creating PRDs. If it needed AI, ChatGPT handled it.
I didn’t question it because I didn’t know any better. Everyone was using ChatGPT, so I used ChatGPT.
But I write a lot. Newsletter posts, social content, strategy docs—writing is the core of what I do.
And I kept seeing the same thing pop up on X:
“Claude is better for long-form writing.”
So I switched. Not because ChatGPT was bad, but because I needed an AI that could actually write, not just generate content.
The difference was huge. Claude’s writing felt more natural, more human, more like something I’d actually publish without heavy editing.
I moved everything over: writing, coding, my entire personal AI agent system. Claude became my default for everything.
But then I noticed something interesting.
When I needed to brainstorm—really explore ideas, generate options, think broadly—Claude felt constrained. It gave me good ideas, but narrow ones. Like it was trying to give me the “right” answer instead of showing me all the possibilities.
I tried ChatGPT for brainstorming on a whim, and the difference was clear. Broader thinking. More directions to explore. Better at generating the raw material I needed before narrowing down.
So now I run a split system: Claude for writing, coding, and being my AI agent. ChatGPT for brainstorming and early-stage exploration.
But I stumbled into this setup by accident. I didn’t test it systematically. I just noticed patterns and adapted.
Today’s guest post is from Timo Mason🤠, who writes Write Your Way To Wealth—tactical AI content systems for Substack creators. This is his second appearance on The AI Maker. In his previous post, he showed you how to repurpose social content like Alex Hormozi. Now he's back with something even more tactical.
If you want to dig deeper into what Timo writes about, check out these three latest posts:
How I create Substack articles 5x faster — My creation framework for authentic ai-assisted writing
CustomGPTs for Substack Creators — Why, when and how to build GPT’s to level up your Substack
How One Post Got Me 5,000 New Readers In A Day — The collaboration strategy that beats organic growth by 50x and why most Substack writers fail at it
Timo actually tested ChatGPT and Claude side-by-side across different content types to see which AI wins where.
Turns out, my accidental split system matches what he discovered through deliberate testing. Each AI has its lane. And forcing them outside of it wastes time and produces worse results.
This post shows you exactly which AI to use for what. It’s backed by real side-by-side comparisons. Plus a 5-step playbook you can run in 10 minutes to test it yourself.
If you’re still using one AI for everything, or you’ve stumbled into a split system like I did, this will clarify exactly what’s happening and how to optimize for it.
Here’s Timo.
Howdy, Wealth Gang 🤠
I wasted one year being loyal to the wrong AI, then I finally tested both side-by-side and discovered something that changed my entire content strategy.
Let me back up…
When I started writing seriously, I had one rule: Keep it simple
Just pick one AI and go all in.
So I naturally chose ChatGPT.
Everyone used it, the interface looked clean, and it got the job done fast.
I built my entire X content system around it.
Idea → Viral Writing GPT Starter Kit→ edit → post.
I racked up 14 million views on Twitter and my threads went giga-viral.
ChatGPT became my co-pilot, and we flew high.
I didn’t question it because the results spoke for themselves.
Why would I switch tools when this one already printed wins?
The minimalist in me loved the simplicity.
One tool, one process, zero friction.
Looking back, ChatGPT wasn’t perfect, but it didn’t need to be.
It got me from zero to 14 million views, and I was happy with it.
But you never know what you’re missing until you try something better.
And I was clueless about what would await me outside of the ChatGPT bubble…
So in this article, I’ll show you exactly which AI wins for what content type—backed by real side-by-side comparisons—plus a 5-step playbook you can run in 10 minutes to test it yourself.
The Claude Discovery
I’d heard about Claude content writing for months.
It even became this internal meme in my head. Everyone said the same thing: “I prefer Claude for writing.”
Like, they just didn’t want to say “Claude is the best writing AI” so they turned it into a personal opinion. :D
But I ignored these “preferences” because, again, ChatGPT worked well for me. :)
Then I had a conversation with Mia Kiraki 🎭 about it and I just respect her expertise so much that I finally gave Claude a try…
I pasted in a prompt I’d normally give ChatGPT and hit enter.
The difference hit me instantly.
The writing didn’t just look “good.” It read natural, more fluent, more human. Full sentences that actually flowed instead of robotic bullet points that turn people off instantly. It had rhythm, variety, and personality baked into every line.
I read it out loud and thought, “Wait, this sounds like something I’d actually write.”
That same day, I switched everything over to Claude.
I’m a minimalist, I hate subscriptions. I avoid paying for tools unless they’re absolutely necessary.
But Claude felt like a must-have, so I signed up for the subscription without hesitation.
From that point on, Claude became my writing partner for every piece of content.
But switching to a new AI came with a whole new set of challenges…
The Messy Transition
I went all in on Claude.
I migrated my entire content system over. Every writing process I’d built with ChatGPT got rebuilt in Claude.
It felt like moving to a new country where everyone speaks a slightly different language.
Sure, it’s still AI. But Claude had its own way of interpreting prompts so I had to relearn how to communicate with it.
I entered discovery mode.
I tested everything. Article intros, full newsletters, Substack Notes, hook writing, email sequences. I wanted to know exactly where Claude shined and where it didn’t.
But then I hit a wall I never experienced with ChatGPT.
Token limits.
Imagine you’re in the middle of a heavy content day, writing multiple articles, then suddenly, Claude taps out. “You’ve reached your usage limit. Try again in a few hours.”
Brutal.
It didn’t happen every day, but when it did, it killed my momentum. I’d be mid-flow, deep in the zone, and then.... “You’ve reached your usage limit.”
I had to either wait or switch tools, so I adapted.
I kept ChatGPT around for the lighter stuff. Analysis, strategy, brainstorming, quick tasks that didn’t need Claude’s writing magic. Basically, I used up the tokens I already had sitting there instead of letting them go to waste.
Claude became my content weapon. ChatGPT became my strategist.
I assumed Claude crushed everything writing-related, but I never actually tested it side-by-side with ChatGPT on the same content types.
So little Timo got curious :)
What if Claude wasn’t the universal writing champion I thought it was?
What if ChatGPT still dominated certain things, and I’d just been ignoring it?
Turns out, I had good reason to be skeptical…
The Real Test
I decided to settle it once and for all.
I wanted to see real outputs to compare side-by-side, so I made the next logical step…
I set up a cooking showdown with Gordon Ramsey as a judge.
The challenge: Turn this exact article you’re reading right now into an appetizer-like short-form piece for a Substack Note.
The judge: Gordon Ramsay. Brutal, honest, and zero tolerance for mediocrity.
The arena: My Article-To-Post AI workflow that I duplicated in both AIs
Let’s see who survives Hell’s Kitchen…
First Up: ChatGPT Presents Its Dish
ChatGPT steps forward with confidence and serves this:
I wasted a year being loyal to the wrong AI.
Not because it was bad.
But because I never questioned it.
When I started writing seriously, I made one rule:
Keep it simple.
Pick one tool.
Go all in.
So I chose ChatGPT.
Clean interface.
Fast outputs.
Everyone was using it.
I built my entire X content system around it.
Every tweet.
Every thread.
Every viral post.
Idea → system → edit → post.
The results were hard to argue with:
14 million views.
Multiple giga-viral threads.
A repeatable workflow that just worked.
ChatGPT wasn’t perfect, but it didn’t need to be.
It was good enough to take me from zero to real traction.
And that’s the dangerous part about tools:
You don’t know what you’re missing until you try something better.Gordon Ramsay takes a bite:
“Right. Let me taste this.”
[reads through]
“Simple. Clean. Gets straight to the point. No bullshit. You didn’t overcomplicate it, and I can actually read this on my phone without losing interest halfway through.”
“This is exactly what short-form content should be. Punchy. Skimmable. Ready to post. Well done.”
ChatGPT exhales. One down.
Next Up: Claude Presents Its Dish
Claude approaches the judges’ table with a full serving platter and presents this:
When I hit 14 million views on Twitter, I thought I had content creation figured out.
It was simple.
ChatGPT. Every single day. Every single post.
One year later, I switched to Claude and realized I’d been leaving performance on the table the entire time.
Here’s what that stubborn loyalty cost me:
I built my entire X content system around ChatGPT because everyone was using it.
The workflow was clean. Idea to post in minutes.
14 million views later, I had zero reason to question it.
Why switch tools when this one was already printing wins?
But when I finally tested Claude side-by-side, the difference was instant.
The writing wasn’t just good. It was natural. More human. Full sentences that actually flowed instead of robotic bullet points.
I read it out loud and thought “this sounds like something I’d actually write.”
That same day, I switched everything over.
Here’s the lesson: success with one tool doesn’t mean it’s the best tool.
You don’t know what you’re missing until you actually test the alternatives.
I wasted a year being loyal to the wrong AI because I never questioned what was working.
Don’t make the same mistake.Gordon Ramsay reads through it slowly:
[long pause]
“Claude. Come here.”
[Claude nervously steps forward]
“What the fuck is this?”
“You’ve given me a bloody three-course meal when I asked for a snack. Look at this! It’s beautiful, don’t get me wrong. Full narrative arc. Setup, tension, resolution. Reads like a proper story.”
“But I asked for punchy SHORT-FORM content. Something people can read in 30 seconds while they’re taking a shit.”
“This? This is a mini-essay. It’s like you’ve served me a Michelin-star tasting menu when all I wanted was a burger. Wrong format, wrong audience, wrong bloody kitchen.”
[shakes head]
“If this was for a full article? Brilliant. But for what I asked? Completely missing the point.”
The Verdict
Winner: ChatGPT.
Gordon Ramsay wipes his hands on his apron and delivers the final judgment:
“ChatGPT understood the assignment. Short-form content needs to be fast, punchy, and ready to post.”
“Claude is an absolute brilliant writer. But you tried to make a five-course meal when I needed a damn appetizer.”
The cooking showdown taught me what months of assumptions couldn’t: Each AI has its lane, and forcing them outside of it is a waste of time.
And once I realized that, I adapted it to my content system…
My Split System
After that experiment, I built a split system based on what each AI actually does best.
Claude handles all my long-form content.
When I’m writing a 1,500-word article, I don’t want bullet points dressed up as paragraphs. I want writing that flows like a river. Claude delivers that without me needing to rewrite every other line.
ChatGPT handles all my short-form content.
When I need a Substack note, hook, or headline, ChatGPT gives me the core idea without over-explanation.
The rule I follow now
Long-form (Articles and emails) → Claude
Short-form (Notes, hooks, headlines) → ChatGPT
But…
Claude is my priority AI because long-form content is harder to create, more valuable to my audience, and has a bigger impact on Substack growth. If I had to pick just one LLM, it would be Claude.
But I don’t have to pick just one and neither do you.
Each AI stays in its lane and because of that, both perform at their peak.
But that doesn’t mean my system also adapts to your workflow…
The 5-Step Comparison Playbook
You’ve seen the results I got, but that doesn’t mean that these are conclusions that work for every creator..
You need to test it yourself.
Your writing style, your content format, your audience, heck… even your prompting techniques—it all plays into the final AI output.
So instead of just trusting my experience, run your own experiment in 10 minutes or less:
Step 1: Pick one AI workflow you’re already using
This could be writing Substack articles, turning articles into Notes, creating hooks, whatever.
Important: Pick something you’ve done multiple times and generally feel confident about.
Step 2: Duplicate this workflow in both AIs
Take your workflow—your prompt and your knowledge base (any documents you feed the AI)—and set it up identically in both tools.
In ChatGPT: Create a custom GPT. Add your prompt as the instructions. Upload your knowledge base files.
In Claude: Create a project. Add your prompt as the custom instructions. Upload the same knowledge base files.
Step 3: Run the exact same input through both
Now that both workflows are set up identically, give them the same input.
If you’re writing an article, use the same topic, same angle, same context in both ChatGPT and Claude.
If you’re repurposing content, feed them the same article to convert.
Step 4: Save both outputs and compare side-by-side
Don’t edit anything yet. Just save the raw outputs from both AIs and put them next to each other.
Ask yourself:
Which one fits the format better?
Which one sounds more like me?
Which one would I actually use without heavy editing?
Your gut will tell you which one works.
Step 5: Make a decision and commit to it for that workflow
Once you’ve compared, pick the winner for that specific content type.
If ChatGPT crushed the short-form content, use it for that. If Claude dominated the long-form writing, stick with it for newsletters.
You don’t need to be loyal to one tool. You just need to know which one does what best.
Takeaway
Stop being loyal to tools. Be loyal to results.
I wasted a year assuming ChatGPT was the only AI I needed because it was working. Then I tested Claude and realized I’d been leaving quality on the table the entire time.
Your next step:
Pick one content workflow you’re already using. Duplicate it in both AI’s. Run the same input through both. Compare the outputs. Then commit to whichever one gives you better results for that specific content type.
Don’t guess. Test it and build your system around the AI that gives you the best results.
If you want more tactical AI content systems like this, Subscribe to Write Your Way To Wealth.
Every week, You get the exact workflows, tools, and strategies I use to grow and monetize Substack hassle-free.
See ya soon
Timo Mason🤠
 | A guest post by
|
I love this, thank you!
Ramsey would be "swearBOT" 🤣