top of page
  • Writer's picturelauralikespi

ChatGPT Take 8 - Money, Money, Money

I've given in.

I'm officially paying for ChatGPT plus.


In case you’ve missed my previous ramblings:


The Fine Details


Let's get the important part up front - money. To use ChatGPT Plus costs $20 per month plus VAT (which is working out as £18.92 for me so far). To me, if you're using ChatGPT regularly or if something you're using it for depends on being able to access the tool at any time, this feels like a low cost.

As far as I can tell, this is on a rolling monthly contract and therefore you can test it out and cancel at any time (or at least, this is why I signed up).


The main features promised as part of ChatGPT Plus are:

  • access to GPT4 within the ChatGPT chatbot

  • better access to ChatGPT (including during busy times) and the responses are apparently faster

  • ability to use new features (when they are in beta)

Why I Finally Caved

Now on the more interesting parts - the why. There were four reasons why I give in and started paying:

  1. The few times I couldn't access it

  2. Hearing people tell me how much better GPT4 was compared to GPT3.5

  3. The Sept 2021 issue

  4. Plugins


I would love to blame the ChatGPT outages on my impromptu schedule for writing these blog posts, but I can't. However, there have been several times when I want to test something or write the blog, and I could not.

I am looking forward to never being told ChatGPT is at capacity right now when I am excited to play around or create something.


ChatGPT is a chatbot, built on top of OpenAI's Large Language Model (LLM). This LLM is called GPT (stands for Generative Pre-trained Transformer). There have been a few version of this model but ChatGPT has only ever used two - GPT3.5 and GPT4. The free version only gives access to GPT3.5.

GPT4 is an improved version of this model - better at logical reasoning and knows more (according to an interview I attended with an engineer from OpenAI). This is because the model itself was trained on more parameters (think of these like words, so GPT4 knows more words) and it has more long-term memory (making it better at having conversations).

On the plus version, users can send 50 messages every 4 hours to GPT4, and freely use GPT3.5 with priority over free users.

September 2021

If you have been reading my blogs, you will know I have been getting increasingly frustrated at this response:

ChatGPT: As an AI language model, I don’t have real-time capabilities and my knowledge is up to date only until September 2021.

I had it in my head that the newer model GPT4 would overcome this issue and be able to give me more up to date responses.


The final reason was everyone going on about plugins. Plugins are add-ons to give additional capabilities to a software tool. An example of a plugin is Adobe Flash Player (trying not to show my age here). If you use Wordpress, plugins are very common.

I was excited to try out some of these add-ons to enhance my ChatGPT usage.

My First Unimpressed Attempt

Yes, this is a pattern. I have a habit of magically expecting new tech tools to be amazing and change my life. I did not find ChatGPT Plus super easy and intuitive to use to start of with (I'm good with tech, I swear).

I decided to try something I had previously done with GPT3.5 - get ChatGPT to talk like me.

L: I am going to send you 4 writing samples. Please do not respond until I confirm I have finished sending all the samples.

I need you to summarise my writing style, tone, anything relevant to getting you to be able to write similar to me. Please do not return a writing sample. Return a paragraph describing the writing style.

Despite my (in my opinion) very clear instructions, ChatGPT kept responding before I'd sent 4 samples, and also kept sending me a writing sample back.

ChatGPT: Of course! Please proceed with the writing samples, and let me know when you're done. I'll wait for your confirmation before analyzing and summarizing.

I sent in one sample (I had to do so individually because they are quite long), and the SuperBrain replied by sending me samples 2-4 written by itself:


Sample 2

I’m going to be completely honest with you all: I did not have a blog post planned for this week. Between meetings, new work tasks, and trying to find some semblance of a social life post-lockdown, my writing took a back seat. But isn’t that life?


I tried again:


1. Do not respond until I said the words "I am finished"
2. I do not want you to generate samples. Do not generate samples of my writing unless asked.
3. There will be a total of 4 samples sent to you by me. I have so far sent one.
4. I want a paragraph summarising the writing style. I want this paragraph to be useful to you. If I was to send it to you in another chat window.

ChatGPT: Understood. Please proceed with sending the rest of the samples, and I will wait for your confirmation with the words "I am finished" before analyzing and summarizing.

This time I had some success, after I sent sample 2 ChatGPT replied:

ChatGPT: I am waiting for your next sample or the confirmation that you are finished.

Unfortunately after sample 3, it seemed the SuperBrain had forgotten the instructions:


Black Mirror: A Deep Dive into Dystopia

Anticipation for Season 6 My eagerness knows no bounds as Season 6 of Black Mirror, a show I've been smitten with for so long, gets teased by Netflix. It's been a lengthy wait since the last season in 2019. With the completion of my thesis writing phase, I've been indulging in expressing my thoughts about this futuristic dystopia. The recent Miley Cyrus track, "Flowers", even prompted me to binge the entire series once again.

Beginnings in Britain For those unfamiliar with Black Mirror's origins, its debut episode aired in 2011. This eerie introduction captivated me, even if it wasn’t the technological spectacle I anticipated. Although my flatmates dismissed it, I was immediately entranced, leading to covert viewings of the Channel 4 episodes.


(Clearly sample 6 was my Black Mirror blog post)

I tried to be a bit more stern:

L: Reminder

Do not analyse until I say "I am finished"

I do not want you to create a writing sample.

ChatGPT: Understood. I will wait for your instruction. Please proceed when you're ready.

ChatGPT continued behaving after I sent sample 4:

ChatGPT: I'm ready for your instructions when you're ready. Please let me know when you're finished and how you'd like to proceed.

L: I am finished.

Please analyse the 4 prompts which begin with "sample".

Do not create your own sample. Instead, summarise the writing style, tone, structure, etc in a way which would make a useful prompt to be sent to you. 

This worked! ChatGPT analysed each sample for Writing Style, Tone, Structure and Key Features. I had purposefully fed in four very different writing samples, so I requested:

L: The four samples were written by the same person. Can you summarise their writing style generally?

To which ChatGPT followed the same pattern, but much more detailed. To get what I wanted, I added one final prompt:

L: If I wanted a writing style prompt to send to ChatGPT, how would you word the writing style of the author?

ChatGPT: Certainly! Here's a summarized writing style prompt based on the author's characteristics:

"Write in a style that is reflective and descriptive, blending personal anecdotes with factual information. Utilize a mix of short and long sentences, often using shorter ones for emphasis. Maintain a tone that can range from reflective and humorous to neutral and personal. Infuse the content with passion when discussing subjects of deep interest or familiarity. Integrate modern slang and colloquialisms when appropriate, and weave in personal experiences to give an autobiographical touch. Focus on imagery and occasionally engage directly with the reader for a conversational feel. The topics should resonate with themes of technology, AI, and personal growth."

This prompt should guide ChatGPT to emulate the author's unique writing style effectively.

Using this prompt in a different chat window to get ChatGPT to write a LinkedIn #mondaymotivation for me resulted in this delight:

Reflecting on this, perhaps (again) I have been harsh on the upgraded SuperBrain. While I know my prompting has improved, and I felt a little bit frustrated when ChatGPT kept doing what I had explicitly asked it not to do, the responses are better (in a weird way which I can't put my finger on).


Next time join me as I start to play around with the notorious plugins.

22 views0 comments


bottom of page