Read time: 5 minutes | Subscribe to this newsletter

Hey there! Moritz here. Thank you for joining me on another post of The Prompt Warrior, where I help you grow your business by leveraging AI.

- Sponsor this newsletter -

A big thank you to this week’s newsletter sponsor: tl;dv.​

Automatically record and transcribe your Zoom and Google meetings with the power of AI.

1,000+ customers save countless hours every week by using tl;dv.

Try out tl;dv here (for free)

In the last month my usage of ChatGPT has declined significantly.

For most of my tasks, I’ve been using Claude (the LLM by Anthropic) instead.

If you’re like me and use LLMs mainly for content creation, copywriting and brainstorming, you’ll want to seriously consider Claude.

In today’s post I’ll do a direct comparison between the two, and see who comes out ahead when it comes to copywriting tasks.

(Spoiler alert: It’s Claude)

We’ll cover 5 use cases:

1. X (Twitter) hooks

2. Email subject lines

3. Cold DMs

4. SEO-optimized blog posts

5. Slogans

Note: I’ve used very simple prompts for each category, in an attempt to make the comparison fairer.

(Btw, Claude is currently only available in the US & UK. If you’re outside you can use a VPN to sign up and then use it without the VPN henceforth.)

Let the battle begin…

1. X (Twitter) hook

Writing the Twitter hook is the most important part of a thread. And ChatGPT is notoriously bad at anything tweet related:

There are so many reasons why this hook is horrible. Including emojis and hashtags for example is a very bad idea. Also the whole hook just sounds like something from a 70s infomercial.

Claude also does a bad job, but at least it sucks a bit less. It doesn’t include any hashtags and the wording is a bit more enticing.

→ Round 1 winner: Claude

2. Email subject line

Next, let’s see how well our candidates do at writing email subject lines.

ChatGPT:

Claude:

I like how Claude makes many suggestions by default, so that you can pick the one you like. This makes it much more likely that you’ll end up with a good result.

The subject line ChatGPT came up with sounds kinda tacky and I pretty much favour all of the ones Claude suggested.

→ Round 2 winner: Claude

3. Cold DM

Next up are cold DMs. Cold DMs are supposed to feel casual and their goal is to get the prospect to reply.

I do like the brevity of ChatGPT here.

But Claude does something else that makes it better here imo: It includes how you plan to help the prospect. This is pretty important.

Both cold DMs aren’t great, but if I had received both, I’d be more likely to reply to Claude than ChatGPT.

→ Round 3 winner: Claude

4. SEO-optimized blog post

Next up are blog posts. I only included the beginning of the posts for readability:

Interesting to see how both articles are actually very similar.

I find it hard to judge this round though, I’ll give it a tie.

→ Round 4: tied

5. Slogan

Finally, let’s see how well they can come up with Slogans.

That’s really cheesy. I’m imagining a cheap Thai TV commercial showing slow-mo’s of lips here.

I again like how Claude gives us several options. I’d say “Nature’s Best for Lips and Skin” is my pick here.

→ Round 5 winner: Claude

Summary

Overall, Claude is the clear winner here.

This definitely matches my experience in my day-to-day tasks. Claude is a much better companion for me as a content creator.

What’s also clear though is that for both ChatGPT and Claude you wouldn’t look at this copy and say: “Wow, great copy!”

This shows again how:

1. LLM outputs are usually not good enough yet to be taken without any human editing

2. Engineering your prompts to produce better outputs is important

Thanks for reading!

If you found this helpful, please support me by forwarding this newsletter to a friend or colleague.

See you in the next one!

Moritz