ChatGPT Got Some Things Wrong About Paragraph

If you haven’t seen it yet, on April 6, I posted a Twitter thread comparing Web3 publishing platforms Paragraph and Mirror.

What you may not realize is that I created that Twitter thread with the assistance of ChatGPT.

It’s not that I needed to do so, but I wanted to experiment with ChatGPT and it was the perfect opportunity to take the AI chat program for a test drive to discover its true capabilities. I’ve got to say, though I’m late to the party, I am a little bit impressed.

I started with a simple question: Compare Mirror and Paragraph.

The program began with Mirror. Its description of Mirror was pretty accurate.

I was not, however, impressed with the description of Paragraph. It got some fundamental facts wrong.

Because Paragraph allows newsletter creators to token-gate their content, I knew the part about not using blockchain technology was incorrect. Furthermore, it isn’t just for serialized fiction. I asked a follow-up question:

I'm sorry. I must have miscommunicated. I meant Paragraph.xyz.

Here is the first half of ChatGPT’s response:

And ChatGPT followed that up with:

Not bad, but it still says Paragraph is for serialized fiction. I’m not sure why the platform is described that way. Maybe it does host a lot of serialized fiction.

The final summary paragraph of ChatGPT’s response reads:

Curious about whether Paragraph’s newsletters can be used to publish non-fiction content, I asked:

Can Paragraph be used to publish newsletters focused on non-fiction content?

Here was the response:

Finally, I got around to asking ChatGPT to compare Mirror and Paragraph as if writing a Twitter thread. Here’s what I got:

ChatGPT’s written response, in full, is verbatim below:

Sure, here's a Twitter thread comparing Mirror and Paragraph:
1/5: Mirror and Paragraph are both blockchain-based publishing platforms that offer creators new ways to share and monetize their work.
2/5: Mirror focuses on long-form, serialized content like essays and books, while Paragraph specializes in serialized fiction.
3/5: Both platforms allow creators to publish their work in a decentralized way, without intermediaries, and receive support directly from their readers.
4/5: However, Mirror uses cryptocurrency incentives to promote high-quality content, while Paragraph operates on a subscription model where readers pay a monthly fee to access all the content on the platform.
5/5: Whether you're a writer, artist, or content creator, both Mirror and Paragraph offer exciting new ways to share and monetize your work. Check them out and see which one is right for you! #blockchain #publishing #creativity

Not bad, but why only 5 tweets? I really wanted something longer. I asked ChatGPT another question:

Is it possible to embed NFTs in Mirror or Paragraph posts?

Of course, I knew the answer to that question, but I wanted to see what ChatGPT would say and, if I liked the answer, I could incorporate it into my Twitter thread. Here’s the response:

The final paragraph of ChatGPT’s response was:

Overall, the use of NFTs in publishing is still relatively new and evolving. However, both Mirror and Paragraph have shown a willingness to experiment with new technologies and explore new ways to support and reward creators, so it's likely that we will see more integration of NFTs and other blockchain-based features in the future.

Satisfied with that answer, I asked the program to incorporate that response into our Twitter thread. Again, I only got five tweets. Here’s the updated thread:

I was fairly happy with the response, but I wanted a longer Twitter thread, so I combined the two Twitter threads on my own to create a unique, original Twitter thread comparing Mirror and Paragraph as Web3 publishing platforms. You can read the Twitter thread right here.

If you haven’t already done so, please take my survey regarding Paragraph and Substack.

If you like Cryptocracy, I’d be delighted if you’d share it with your friends and subscribe. To subscribe at Paragraph, click here. At Mirror, click here. At Substack, click here.

First published at Substack.

H2
H3
H4
3 columns
2 columns
1 column
Join the conversation now
Logo
Center