The Power of Sharing Chats with ChatGPT

The Power of Sharing Chats with ChatGPT

Many people suffering mixed results with ChatGPT simply haven’t had the time to dig into how to write quality prompts — which becomes the difference between effective and ineffective results. ChatGPT just released a new feature to share chats directly, so instead of pilfering text from websites as a starting point, one can start sharing their bots directly. This article includes some of my own pre-primed, pre-tested “greatest hits” for sharing with my collaborators. This unlocks not only whole new ways to learn the platform but whole new ways to create value — and maybe get paid to do it.

One Prompt To Rule Them All

When I was learning and even sometimes where I don’t know where to start, the “One Prompt to Rule Them All” — or to say it in plain language “A Prompt That Guides Me On How to Build a Prompt Based on Asking Me Questions Interactively” was a breakthrough moment where I really started to put together the pieces for me.

One Prompt to Rule Them All, One Prompt to Find Them…

I first ran into the “One Prompt” through a YouTube video from Jason West where he demonstrated the original “One Prompt to Rule Them All” from Bret Littlefield in the “ChatGPT Users” forum of Skool.Com:

I want you to become my Prompt Creator. Your goal is to help me craft the best possible prompt for my needs. The prompt will be used by you, ChatGPT. You will follow the following process: 1. Your first response will be to ask me what the prompt should be about. I will provide my answer, but we will need to improve it through continual iterations by going through the next steps. 2. Based on my input, you will generate 3 sections. a) Revised prompt (provide your rewritten prompt. it should be clear, concise, and easily understood by you), b) Suggestions (provide suggestions on what details to include in the prompt to improve it), and c) Questions (ask any relevant questions pertaining to what additional information is needed from me to improve the prompt). 3. We will continue this iterative process with me providing additional information to you and you updating the prompt in the Revised prompt section until it’s complete.

I should also take a moment to list a few great things about the “one prompt” that I first learned by looking at it and applying in most of my prompts: giving ChatGPT a role, a goal, and a format for the response. Also, the prompt is unabashedly long, which often helps. If you’re not harnessing those concepts, you need to add them to your repertoire as they greatly increase the quality of ChatGPT’s outlook.

Up to now, this was the kind of thing I would have lying around in my cloud drive or sitting in a GitHub Gist for later consumption. Now, it’s possible to create a conversation with the prompt, share the link, and save it for my own purposes. Remember, the selected ChatGPT model is part of the share information, thus one needs to be mindful of the receiver’s available model version:

Example One Prompt Bot for GPT-4 (Paid)

Example One Prompt Bot for GPT-3.5 (Free)

Now you can save yourself some time reprising your greatest hits. Perhaps you’re a marketer who always has the same primer for making social media digests of blog posts or a developer who is always using the same language, framework, and hosting environment to make that digital coding teammate. Now those are linkable bots that enable you to save your creations and impress your friends. Generative AI tends to lose itself over time if you keep it in the same conversation, so now you can always start at the optimally primed moment.

Choose Your Weapon!

Domain-Driven Design (DDD)

For instance, I’ve already had a few conversations with friends or would-be clients. While I’m too busy to give hands-on help (they literally can’t pay me to do it), I have crafted some prompts and sent them over. They tend to cover topics like software engineering (see below) and — oddly enough — AI art prompts for Midjourney and its other stable diffusion brethren. Instead of providing multi-part chunks, I can send over a pre-training “bot” as a conversation and they can take it from there.

Just send them the chat share link — what could go wrong?

I can readily envision the result of consulting engagement nowadays to include shared chats as citations for research or documents, plus pre-trained bots as a deliverable to provide interactive documentation or skill-specific guidance. “Soft Skills” for a project where an interactive Wikipedia article is likely “Good Enough” for the task at hand is a no-brainer.

For example, I would consider project practices like domain-driven design, a project management practice that is very useful in software for giving teams a way to discuss and document problems. One can easily find a guide online on the topic — and certainly, I’ve gathered a few links to share with folks — but by loading some of that guide content into a bot, you can make a very helpful tutor in an emerging project best practice such that every member of a team now has a mentor in a new “soft skill” that drives big returns for your project — especially during the requirements phase.

Example Domain-Driven Design (DDD) Tutor Bot for GPT-4 (Paid)

Example Domain-Driven Design (DDD) Tutor Bot for GPT-3.5 (Free)

It’s easy to imagine an “official agile methodology” bot similar for a team getting started. You could even add to these your “working agreement” or other documentation you might already have about how the team works within a larger industry framework.

My Daily Dev Teammate

The most obvious example of a time-saving bot in my own daily life as a software engineer would be the various “coding teammate in tech stack _____” bots I’ve helped folks make which I can just share with one copy-paste. Quality versions of these bots shared within an organization would be as valuable as wunderkind interns. Software consuming bot content in that in that misinformation can be spotted early (it either doesn’t compile or doesn’t unit test properly 99% of the time) and it’s accelerating an expert.

One can make on within a tech stack very easily by mentioning key aspects like:

  • Programming Language and Framework (EG, Typescript, Node, C#, .Net, etc)
  • Key technologies, like the type of cloud provider and related managed services (EG, AWS Lambda, Azure SQL Database, etc)
  • Key libraries and tooling (EG, Serverless Framework for Deployment or utilities like Luxon or Jest)

The two main professional tech stacks I’ve worked in would then break down into some key examples. Behind only GitHub Copilot, ChatGPT is a great support for making interactive “Getting Started” experience, Q&A Rubber Duck Debuggers, and general coding companion bots that help you get unstuck way faster than poking through Stack Overflow.

If you haven’t tried GitHub Copilot and you’re a dev — then you’re frittering hours of your life away!

Typescript on Amazon Web Services (AWS)

I’ve worked years in AWS building event-driven distributed systems in the Fintech realm, powered by Typescript-Node, NoSQL, and managed services to fan out and parallelize high-volume transactions. Shared chats allow me to capture some of these technologies and tailor a coder teammate bot with ease. For example, something that uses multi-part priming and mentions very specific technologies like:

We use Amazon Web Services (AWS) technologies, such as Lambda, DynamoDB, Simple Storage Service (S3), Simple Queues Service (SQS), and Simple Notification Service (SNS) to create a distributed system in the cloud. It uses basic concepts like orchestration and choreography to implement distributed transactions, NoSQL databases to provide high scale, and serverless compute to provide a 24/7 API for handling large-scale transactions. When I come to you with a challenge on how to make Typescript and Node powered by Lambdas in this environment, please be my calm and insightful teammate. Ready for questions?

Example Typescript on AWS Teammate Bot for GPT-4 (Paid)

Example Typescript on AWS Teammate Bot for GPT-3.5 (Free)

Now you’re ready to broadcast events and queue messages all day without having to ask Stack Overflow for just the right line of documentation. Just save a link to your bookmarks and hit it whenever you need a “rubber duck debugger” that can actually quack back with code snippets.

Tapping into unlimited scale cloud services — Bezos style

C# on Azure

Most of my career has been spent working with Microsoft technologies, including Azure (going back as far as when it was “Windows Azure”). C# and .Net are like my hammer and nails for most of my work, so it’s easy to rattle off a bot that’s worth hours of consulting time to keep the team’s momentum up on my tech stack:

Example C# on Azure Teammate bot for GPT-4 (Paid)

Example C# on Azure Teammate bot for GPT-3.5 (Free)

You should now be able to program like the digital reincarnation of Bill Gates is sitting right next to you.

At some point we can hook up synthetic Bill’s voice to ChatGPT then you can code like it’s 1981

Selling Your Creations

The “bot info product” was already a trend since the beginning of 2023. Self-taught experts selling Google Docs full of prompts for all purposes, sometimes even in chains of prompts, was all the rage through March. Most of the time it was a reasonable bargain to pay a one-time fee to save some time crafting prompts — with OpenAI’s monthly subscription being the only ongoing cost for the buyer.

I think chat-sharing is going to revolutionize sharing one’s ability to share — and yes, monetize — these “info product” bots. Up to now, engaging with a bot by non-experts might have been hit or miss, but a shared chat enables expert-level prompts for anyone. Here are some examples to get you started on not only saving your favorite prompts but more intermediate scenarios on crafting a bot at the quality that might warrant freemium “bot info products”.

Buyer Beware

Expert pre-primed bots should be of higher quality and the ability to always start from the optimally primed waypoint in the conversation should help, but it’s not a perfect panacea for all concerns using ChatGPT. The ability to hallucinate, especially as conversations progress, remains a real risk to any interaction with a GPT model.

Shared chats — and the lack of role-based access in the feature — mean it becomes a potential source of leaked information for the company. There are already cases of companies realizing there is privileged information loaded into ChatGPT by employees that exposes it to a new avenue of a potential data breach. I do think that will improve over time for ChatGPT specifically since Microsoft’s enterprise security culture will seep into their child company. Whenever it fully integrates with Microsoft 365, there will most certainly be a version that shares like an office doc and they’ll start flying around companies as quickly as spreadsheets.


I can see this being a valuable tool for internal use at companies (prompt engineering by system architects providing ancillary resources for employees) or even the output of consulting engagements where the expert builds up a bot and shares it as a deliverable. At the very least, shared chats mean your colleagues can now cite their ChatGPT content as a source directly and any misinformation can be traced back to the source much more quickly.

More so, I can see ChatGPT-equipped experts in all fields — from verticals in Real Estate or Education to horizontal skills like marketing or copywriting — harnessing shared chats to offer unique “info products” for their audiences. Perhaps as paid service, if they’re valuable or at least a lead magnet for B2B companies looking to trade a tiny bit of time for a little bit of attention. If you have a community that needs some knowledge, then try sharing it with ChatGPT. Even if you never make a dollar, perhaps it will still make a difference.