I wrote 20 articles in 1 day with the help of free AI tools (The Results)

Today I wrote 20 articles with the help of AI to help kickstart the blog for my new WordPress Support business WP Master. The plan was to do 4 blocks of 2 hours straight and aim to do 5 articles in each block, check out this morning’s post here to see the full plan.

Here’s a quick video of me making a start and the process I started following. I made this video first thing in the morning, the results are further down in this post.

Tools

First off here are the tools I used:

WordPress (obviously)
ChatGPT – free version
Google Bard
Apple Preview – I’ve been using this a lot for shrinking images it’s quick and handy.
Photopea – Awesome free image editing tool for stuff that was slightly more complicated. Basically a free online version of Photoshop.
Unsplash – I got most of the blog images from here and some from Google Images with creative commons license on to make sure I wasn’t stealing. Interestingly I thought I would use AI more for images but Dalle-e 2 just gave weird output most of the time, I can no longer generate images on Mid Journey without signing up for a paid plan and I searched the Lexica AI image search engine a few times but the pics were all kind of the same style and didn’t really work.

Plus after about 20 years of working on a laptop full time I just started using my iMac which I’ve had for about 6 years and never used. Very silly, it’s so much better haha!

Making a start

I was hoping to start at 7am but life got in the way so I started just after 8am. I kicked off making the following video on how I was going about it.

20 Articles, 21,345 Words

After around 11 hours including breaks, perhaps 10 or so hours of work, I finished up the final post. All up I made 20 articles, totalling 21,345 words. To put that into perspective, my first book The 7 Day Startup was just over 25,000 words.

Also I should point out that I haven’t proof read any of these articles!!

Article list

I came up with a list of 30 articles using AI to help (see the process in my article about the plan for today). And throughout the day I went through them mostly in order, but I skipped ones I didn’t feel like attacking. I’m glad I had 30 because in the end there was only about 20 there that I really wanted to pull together. Here are the 20 posts I wrote and further down is the 10 I had left over.

Here are the ones left over (I came up with 2 more through the day).

Why Hire Someone to do WordPress Updates
WordPress website help Australia – why choose WP Master
Where to find the best WordPress support in Australia
WordPress support – The best options around Brisbane and the Gold Coast
Tips for Optimizing WordPress Website Performance
Essential WordPress Maintenance Tasks for a Healthy Website
How to choose the right WordPress hosting provider
What building websites was like before WordPress (skipped for now I feel like I’ve covered this enough in other posts but may come back to it).
What trends might help or hinder WordPress in the future (skipped for now, I’ve covered this in other posts but may come back to it).
Comparing the trends of today’s most popular website builders (skipped, not sure about this one)
How to deliver outstanding WordPress customers service (will come back to this when i have more time ill have to write most of that myself)
Common WordPress Issues and How Support and Maintenance Services Can Help (skipped for now)

Updates throughout the day

I popped back into this post to make some updates throughout the day.

8am – First 2 hour block, let’s go!

Published the plan here, let’s start this first 2 hour block. Goal 5 articles!

11:15 update (2 working hours in)

So I started an hour late (was planning on starting at 7) and the first 2 hour block too me about 2.5 hours, and my break went longer, so I’m about 2 hours behind oops.

I couldn’t help myself writing a lot of the content myself, I like writing about my own experience so the generic content wasn’t really doing it for me. That said, the AI tools were super useful at filling in the gaps.

All in all I’ve done 6 articles, not a bad start but I’ll have to pick up the pace!

1:37pm update (4 working hours in)

I’ll be honest this is a bit harder than I thought. I’m a bit behind but I’ve done 10 articles. I’m a bit exhausted so having a break for lunch. I’m still finding that I’m writing most of the articles myself, I struggle a bit to use the content from the AI, although I’m definitely using bits and pieces. Some articles like this one on setting up a Staging site I’ve written 100% myself. Others I’ve used AI a bit more although so far those are rare. 5 recommended plugins for every WordPress install I used AI as the starting point for the first 4 items, but I chose the specific plugins and I re-wrote pretty much every line to make it my own.

Because I’m a bit behind on time I’m also not doing some of the things I’d usually do like adding links to every mention in the article (I’m doing some), and adding images inside articles. Again I’m doing some but not many because I’m rushing a bit.

Still, this is fun and I think these articles will be useful to some people, and for me to use with clients in the business, so I’m feeling like it’s a worthwhile effort.

4:24pm update (6 working hours in)

Well I’ll be honest I’m kinda exhausted haha. I’m using AI a lot for small ideas for the content, but haven’t been able to shake the habit of pretty much writing the whole article myself. I also don’t want it to be shitty generic content, so I’d rather it take a bit longer and have content that’s worth reading.

I’m 14 posts down with 1 block remaining, although it’s already almost 5pm when I was planning on finishing so I’m a fair way behind. Having a quick break and then I’ll get back into it for the final stretch (hopefully).

One more point, Bard is damn good. I’m finding myself using it more than ChatGPT. It includes image and links which is awesome, it has the ability to browse the web which is a big plus and it has a cool feature to export to a Google Doc that I’ve used a few times for larger bits of content.

6:43 Finished! (4 blocks of writing over 11 hours total).

All done, 4 writing blocks complete. All went overtime. All in all just under 11 hours in total time, I’d say around 10 of actually working at the computer.

Beer O’clock!

Some lessons

I’m a bit tired so I might need more time to reflect on some lessons but here are a few:

I thought I would use the AI content a lot, in the end I didn’t use much of it. I like writing about my own experiences so most of the content was re-written. Some posts I entirely wrote myself, very few were majority AI content. That made the whole process take longer but hopefully the content is better as a result.
I used only free tools, if you consider the pretty damn cool paid tools out there you’d have to wonder what was possible. Even just GPT4 with browser and plugin access would have been probably a huge improvement.
AI was unbelievably useful. I haven’t counted how many prompts I put into ChatGPT and Bard but I had them both open all day and probably asked it various things every 15-20 minutes.
I used the AI differently than I thought. I thought I would ask it for an article on a topic, copy all the content into WordPress and go through and edit it. Instead the behaviour was generally “thought dumping” into WordPress immediately and only using AI when I got stuck. For example for a post like Shopify vs WordPress: Pros and Cons, I started writing it and when I needed to know details I could get them super quickly from the AI tools. I was cheating by writing about things I knew about of course, it wouldn’t work this way if it was things I didn’t know about. I asked things like “when was WooCommerce released”, or “when was WooThemes acquired”, or “what was the market share of Shopify in 2015” and “How many extensions are there for Woo Commerce”. Sometimes I would ask it the question I’d already answered on the blog to see if I missed anything obvious (“what’s good about Shopify” or “what’s bad about WooCommerce”). In the end, I’m pretty confident these articles are let’s say 95% of what they would have been with me doing them the old school way, but there’s no way I could ever produce that much content on my own in one day.
Bard is awesome! By the end of the day I found myself using Bard as the default and ChatGPT if I wasn’t happy with the Bard results. Pretty remarkable given how we were all thinking about AI and ChatGPT a few weeks ago. ChatGPT has some serious competition on its hands. Bard benefits included access to the internet, producing nice neat tables which was awesome, it includes images and links and it has a sweet feature to export the results to a Google doc, how good!
AI sent the efficiency through the roof. l did a similar experiment about 10 years ago to see how much content I could create in a day. I made 13 articles and around 10,000 words. Today I doubled that and I think the content might be better overall. The AI makes the whole process unbelievably efficient. There is no thinking time. When I’m normally working I’d say a good chunk is stopping and thinking – maybe even half the time. That’s gone with AI so there is no stopping. It’s exhausting but unbelievably productive.

Time will tell how useful any of this content is. If answers are so easy to get from AI then why do we need articles on blogs? I did have other reasons why I wanted the content but I fear the days of creating content just for “SEO”, even if it’s good content and even if it’s useful, might be very nearly over.

If you have any thoughts, hit me up on Twitter @thedannorris.

Featured image from Lexica