How AI is Used in Blogging and Writing

Lessons and tips from professional blogging have been shared over the years, including insights from a decade of experience. Recent significant changes in the blogging workflow are now being discussed.

Pre-AI Blogging Workflow

The pre-AI blogging workflow, established over many years, typically involves these steps:

• Interesting conversations at work or with friends/colleagues often spark ideas.
• Blog post titles are frequently sent via email from a phone.
  • These titles typically contain strong opinions and can be easily expanded into full blog posts. Examples include:
    • “what needs to happen for web3 gaming to work”
    • “lessons learned from launching games that apply to web and mobile apps”
    • “why startups should ignore social media haterade”
    • etc
  • An app like ‘Email Me’ can be used, offering a simple text box and one-button send, which is also useful for saving articles or to-do list items.
• When processing emails later, one of two actions is typically taken:
  • If motivated, a blog post is started immediately using the title.
  • Alternatively, titles are added to a Notes app, accumulating a backlog of over 100 potential posts.
• Writing usually occurs on weekends, aiming to complete a full essay in a single 2-3 hour session.
• Proofreading or rewriting is rarely done after the initial writing session.
• Essays are typically scheduled for publication on Monday/Tuesday mornings at 9 am.
• On publication day, a light proofread is performed, followed by writing a tweetstorm to drive traffic.

This established process has proven effective for years. Ultimately, its success lies in dedicating a few hours on Sundays for writing, despite numerous emails, entertainment, and professional research tasks competing for attention.

However, over the past year, the workflow has been adapted to incorporate new generative AI tools, especially ChatGPT. This article aims to share experiences regarding its strengths and limitations.

Post-AI Blogging Workflow

Following the release of ChatGPT, various experimental uses have emerged. Several beneficial applications are outlined below.

Starting with a blank page can be challenging. Generating a first draft, even if imperfect, can often help stimulate creativity. A prompt similar to the following might be used:

I’m writing a blog post about how Gen Z consuming more video, playing games, and other forms of rich and interactive content will mean that they will read less long-form text like newspapers and books. Write me a first draft of a 5 paragraph essay that explores this topic, but only use 2-3 sentences for the opening and closing so that it’s terse and punchy. Be opinionated.

ChatGPT often produces a seemingly coherent response. However, closer examination of the results reveals a lack of specific examples, storytelling, and statistical data. For instance, Paragraph 3 might miss an opportunity to discuss YouTube’s educational applications. The writing style can also appear stiff and formal, often not aligning with a personal voice.

Overall, the initial output may not be ideal.

Nevertheless, by identifying flaws in the generated text, iteration becomes straightforward. A follow-up prompt can be used to revise a specific paragraph and link it to a different idea. Alternatively, the text can be edited and rewritten entirely; even if 90% of the final version is different, it remains more productive than starting from scratch.

Brainstorming Outlines, Topics, and Questions

While ChatGPT may be less effective for direct writing, it excels as a non-judgmental brainstorming partner, unconcerned with the quality of ideas or writing style. It proves highly valuable for generating blog post outlines, compiling topic lists, and formulating inspirational questions. In this context, the AI’s content does not require a 100% hit rate; if 20% of its generated lists inspire actual work, it is considered successful.

For instance, if the goal is to write a blog post about new app opportunities in VR, a query such as the following can be posed:

I’m writing a blog post about new app ideas in VR. What are 15 questions that someone might ask me about this topic? Make them spicy, and if helpful, cite statistics to make the questions more interesting

The AI generates a diverse set of questions, some of which may be uninspired (e.g., simply asking about useful VR apps). However, it also includes many intriguing questions related to areas like esports, content creation, and corporate training.

This functionality can also be a valuable asset for new podcast creators interviewing guests, or for drafting essays that critique existing paradigms (e.g., “Give me 10 questions from someone who is skeptical of of the value of AI safety”). And once an interesting question is identified, outlining an entire blog post around it becomes straightforward.

For example, crafting prompts for outlines that specify particular topics and ideas is often beneficial.

Take the concept of K-12 education and VR – can you write an outline expanding on the promise and also skepticism around this idea? In particular touch on the concept of blending entertaining and gaming and learning, as one of the points. Create 6 sections, and put 3 sub-bullets under each. Use strong opinions and ask tough questions.

The resulting outline is generally effective and serves as a solid starting point.

ChatGPT is also highly effective for generating outlines for long-form writing. A book on network effects, for instance, was written three years ago without AI assistance, utilizing an outline structured in various stages with examples from iconic tech companies. A prompt similar to the following could be used with ChatGPT:

Create an outline for a book on network effects for mobile apps and websites. Stage out the outline so that it starts on launching a product from zero, and the issues there, versus the stages of scaling the product, then hitting market saturation. Make sure you are able to weave in the issues facing network effects-driven products like social networks, marketplaces, dating apps, and collaboration tools. Write a 10 section book, with 3-5 sub-bullets. The first sub-bullet in each section should always be an example of an iconic tech company, like Uber, Airbnb, Dropbox, Tinder, Slack, or others. Please make the comprehensive outline for a book focused on network effects targeted at professionals working in the tech industry

The generated outline is quite useful. While it may differ from a personally crafted structure, often categorizing by product type rather than broad concepts, it still provides a solid foundation. This tool could readily be employed for brainstorming and prototyping, allowing for iteration through various conceptual descriptions. It could facilitate working at a higher conceptual level in early stages, avoiding immediate immersion in sentence-level details.

Cleanup — At the Beginning and At the End

A helpful tip for those struggling with writing is to verbalize ideas aloud. Many individuals are more verbally inclined, making interesting connections when speaking, but facing difficulty when confronted with a blank page. The voice feature in ChatGPT’s mobile app, along with products like Oasis AI, offers an intriguing solution: users can speak on a topic in a stream-of-consciousness manner, and these tools can significantly refine the output. While the initial output may not be perfectly usable, it provides a foundation that can be further added to and edited.

This approach becomes powerful when combined with AI-generated outlines. By providing ChatGPT with a topic or a strong opinion (e.g., “open source products have bad consumer UX!”) to generate an argument outline, one can then verbally elaborate on the argument, record it, process it through an AI app to convert to text, and subsequently edit the result.

AI-driven cleanup is also beneficial at the final stage, ensuring the content conveys a professional or expert tone. ChatGPT can rewrite content to adopt a desired tone or incorporate specific examples for citation and development. This capability to refine ideas and text both during and after the creation process is quite compelling.

Current Limitations of ChatGPT

Generative AI holds significant promise for enhancing the writing process, and its current limitations are expected to be addressed in the future. Several noticeable limitations include:

• The desire exists for AI to access real-time data, enabling it to use the most current examples and statistics.
• Some generative AI tools can produce images, which is beneficial for creating charts, figures, concept sketches, and lead images for essays.
• Integration with social media platforms like Twitter would be ideal for easy content sharing, including AI-generated tweetstorms.
• The ability to train AI on a personal writing corpus, as well as on the styles of respected writers, is desired. Currently, customizing the tone of voice requires effort.
• The capability to feed data, documents, and other information into ChatGPT for use in arguments and suggestions is sought. Tasks such as summarizing a 10,000-word chapter into a 100-entry tweetstorm would be highly beneficial, but are not yet possible.

While there is still significant development ahead, the potential is incredibly promising.

In current blogging practices, it is common to have WordPress open in one window and ChatGPT in another, frequently switching between new chats and copying content. This represents the first significant workflow change in over a decade, and future developments are eagerly anticipated.