Faster Writing Iterations With ChatGPT

I’m careful about jumping on technology bandwagons too early. With ChatGPT though, I’ve been finding powerful ways to incorporate it into my daily life.

Below I’ll talk about how it helps me to take my rough ideas, and iterate quickly to come up with a strong result.

Key ideas below:

  • Pulling out the most convincing ideas, and emphasizing those
  • Re-framing the content for different audiences

Below, we’ll use a real example by building a paragraph explaining TypeStream to various audiences.

Get help building the core ideas first

Just like we all did in high school, we need to build an outline or even just bullet points of the key things we want to get across for our writing. By nailing this down first, we can later ask ChatGPT to structure it in different ways in our next step.

To build a first outline, throw all the following in the prompt:

  1. Give as much context as you can - ChatGPT does well with as much unstructured text that has the context that you think is relevant to your input.
  2. Make your ask clear:

ChatGPT: What are the core benefits you see based on our pitch (content below). You might have to infer and make assumptions.

I removed the text with (...) for clarity.

Result:

Simplicity: (…)

Speed: (…)

Data Privacy and Security: (…)

Cloud Deployment: (…)

Real-Time Data Enrichment: (…)

Ease of Interaction with Kafka: (…)

I really didn’t think these were very compelling reasons for developers to use TypeStream. Two thoughts:

  • the context I fed the AI didn’t actually paint a clear picture on why I think TypeStream is powerful.
  • we also got feedback others thought this was a dev tool. Maybe the AI is confirming that for us.

First Draft

Let’s help ChatGPT understand what we’re hoping TypeStream can do:

Prompt: One of the key benefits that we haven’t highlighted is how easily one can prototype locally, then when it looks good, it’s a single command to deploy that into production. I think that would save software developers a lot of time. 1000x faster was an exaggeration. Can you emphasize that.

I was hoping it would continue building an outline with bullets again (though I didn’t specify for it to do that so I shouldn’t be surprised). The response:

Certainly, this additional benefit is very important and can be emphasized as follows:

Prototyping to Production with Ease: One of the standout features of TypeStream is its powerful prototyping capability. (…)

(…) It’s not just a step forward in data pipeline creation; it’s a giant leap.

Iterating quickly

Okay! Now we’re getting somewhere. Two things I noticed:

  • The marketing lingo is terrible
  • We need to paint a picture of how TypeStream is better than the status quo.

Prompt: Can you restate it but with without “marketing lingo” like: " not just a step forward in data pipeline creation; it’s a giant leap". What would be the status quo of doing that today? Say: Comparing with building a app that processes the Kafka events.

ChatGPT came back with an emphasis around maintenance and the value is in the development speed. I also think the language it used isn’t commonly used for software engineers

Prompt: Instead of “reconfiguring settings, redeploying the application, and making sure the entire pipeline is production-ready” can you break this down and emphasize building it from scratch. Use the terms “data processing” instead of “data pipelines”.

It came back with a version that was pretty good but I still spent five minutes re-writing it to be in my voice and some other minor details.

Total time to come up with a first draft I’m decently happy with: 10 minutes:

If product engineers want to build a new feature or product with the data on Kafka. Today, engineers have to build an app service to consume the data from Kafka, test it, containerize it, deploy it and monitor it. This could take days or weeks.

What if this same development lifecycle could be done in just a few minutes? With TypeStream, this entire process is simplified.

TypeStream is a Kafka backed application development platform that is very easy to interact with. To build a web application that does data processing, it’s as easy as writing Unix pipes. When you’re ready to deploy to production, it’s just a matter of issuing a single command - no need massive deployment pipelines.

This way, TypeStream significantly reduces the time and effort typically required in the data processing life cycle, offering an efficient and more straightforward way for your engineers to work with data.

Reframing it for different audiences

While the tool is mostly geared to engineering teams that have Kafka in their engineering organization, I wanted to explain it to people with other contexts:

  1. Engineers who may have heard about Kafka, but have never used it.
  2. Tech investors in case we want to secure some funding.

Audience 1. Software Engineers without Kafka experience

Let’s start with developers. I wanted to be very specific with the type of develop that we need to explain it to:

  • they aren’t familiar with Kafka
  • they do understand relational databases and traditional queues.

Prompt: Can you take the same value proposition above and make it understandable for a software engineer who isn’t familiar with Kafka or event streaming but who are very comfortable with queues and relational databases.

ChatGPT’s response:

Imagine you’re a product engineer, and you’re tasked with building a new feature or product using data from a queue or relational database. (…)

Whoa there partner.. my bad. TypeStream only works when the data is in Kafka already. So maybe we have to explain what Kafka is and why a team might want to have data in there.

Prompt: Not quite. TypeStream requires the data to already be in Kafka. Please revise with this consideration:

  • briefly explain what Kafka is and why it is powerful for software engineers who aren’t familiar with Kafka or event streaming and why it’s useful. But they ARE very familiar with relational databases and queues.
  • then since TypeStream depends on leveraging the data that has to be in Kafka, explain how the value proposition is valuable for people when the data is in Kafka. Here is my closest pitch for other audiences so far: (I pasted in the draft above to be sure it works for that version, and not what is just generated a second ago)

ChatGPT’s response:

Kafka is a distributed event streaming platform, often seen as a more powerful alternative to traditional queues. (…)

Now, if you’re a product engineer intending to build a new feature or product using the data available in Kafka, you typically face a development process that can be quite lengthy. It involves (…)

However, what if there was a way to cut this extensive development cycle down to mere minutes?

(…)

Okay! Way better.

Audience 2. Tech investors

Now I wanted to take that same explanation for tech investors. I know they are busy and have to read pitches all day long so I thought bullet points could be a helpful way to organize things:

Prompt: Now, can you rephrase the above explanation:

  • audience: venture capitalists who may not be technical themselves and especially don’t understand data streams or Kafka.
  • make it more bullet points
  • emphasize on business value: accelerating development, increasing revenue, reducing cost

Response:

Absolutely, here’s a simplified version suitable for venture capitalists, emphasizing on business values, and structured in bullet points:

  • Introduction to Kafka: Think of Kafka as a very high-performance pipeline that can transport, store, and process streams of records or events (like orders, log updates, etc.) in real time. It’s more robust and flexible than traditional databases or queue systems.
  • The Current Challenge: Developing applications to interact with this stream of data in Kafka is complex and time-consuming. It requires creating, testing, preparing, deploying, and monitoring an application service, which could take days or even weeks, slowing down product development and potentially delaying revenue.

The TypeStream Solution: TypeStream is a platform designed to work with Kafka, turning the often complex and lengthy process into a simple and efficient one. – Building applications with TypeStream is as straightforward as writing simple commands, significantly accelerating the initial development stage. – Transitioning the application into a production environment is as easy as issuing a single command, simplifying deployment and reducing the time taken to launch products or features. The Business Value:

  • Accelerated Development: By making the process of building and deploying applications faster and more intuitive, TypeStream helps teams shorten the time to market, enabling quicker product iterations and feature releases.
  • Increased Revenue: Faster development and deployment times mean that new features and products can start generating revenue much sooner.
  • Cost Reduction: By streamlining and automating complex processes, TypeStream reduces the man-hours needed for these tasks, leading to significant cost savings in the long run.

Conclusion

And there you have it. Hopefully you now have some more ideas on how you can get your ideas organized quickly, and working with ChatGPT as a writer and you are effectively the editors.