Early Hints: On Writing with AI

Early Hints: On Writing with AI

Welcome to Early Hints.

A newsletter to help busy business leaders to stay on top of new developments.

This edition: Lessons from 12 months of Writing with AI. TL;DR

  • There's an AI hidden in the writing tools you're using.
  • Writer's block may soon be a thing of the past.
  • Drowning in a sea of content? There's a lifeline.
  • Integrating AI into your writing workflow doesn't mean selling your soul.
  • I successfully used AI tools in working with major brands to improving the quality of work.
  • There are ways to mitigate the risks in working with AI.

Writers who use AI tools

Pretty much everyone uses AI tools in their writing today, you included, whether you are aware of it or not. If you are using Grammarly, you are one of them. The current versions of Microsoft Office and Google Workplace already have predictive and generative AI embedded.

These tools already assist in improving and refining clunky or poorly written paragraphs. For fiction writers, AI can help overcome writer's block by providing alternate options. Here, the problem of the AI tool hallucinating facts has a positive effect. Claude and ChatGPT are great writing tools, while Gemini excels in online research.

I wouldn’t go as far as calling these “collaborators” as none of the current tools are opinionated. The tools can generate feedback and critical analysis from different perspectives, helping to improve the writing. But only if you think about the output and not use them as suggestions.

Humans write better than current AI systems because we write with intent. That includes breaking grammar rules, misusing words, and playing with language in a way that LLM don’t seem to be able to yet. The other caveat is that AI systems are not designed to support the actual writing process. You don’t want to try to coax the output you need out of the system while you are trying to formulate the writing. Prompting is an obstacle in some ways. In other ways, describing the context of your request helps to clarify your own thinking.

There is only so much time

Which articles to read? Which books to buy? Which people to follow online? There is only so much time and attention span, but at the same time, you do not want to miss out on anything important. While AI tools are accelerating the production of content, they also help its consumption.

For people interested in research, AI can summarize academic papers and help writers understand complex topics. Or it could view hundreds of pitchdecks and business proposals to help an investor dig deeper into the interesting ones. AI tools are already used in legal work, to transcribe and summarize meetings, and to report on sports.

There are youtube and podcast summarizers to help you get the key points and decide whether you should watch the whole thing or not. The transcripts and translations have been AI driven long before LLMs. Today, it doesn’t make sense to transcribe, translate, and summarize a conversation without using AI tools for the initial pass.

We all want to be original

“You don't need to follow anybody! You've got to think for yourselves! You're all individuals! Crowd: Yes! We're all individuals! Brian: You're all different! Crowd: Yes, we are all different! Man in crowd: I'm not...”

from The Life of Brian

Refusing to have genAI tools write everything doesn’t mean you have to throw the baby out with the bathwater. There are lots of good uses of genAI without compromising your integrity.

Many writers are online jumping on the AI bandwagon and acknowledge that their texts could not have been written in its current form without AI assistance. AI plays a significant role in the writing process.

That said, the number of people using AI tools (beyond the obvious ones like Grammarly) as part of their workflow is low. It’s still a scarce skill to be able to use these tools effectively. If you haven’t been making regular use, start today.

Integrating AI into your Workflow

Here is my treat for you this week. As a writer, I've researched how other writers use AI tools and I want to share with you a step-by-step guide on how to integrate these tools into your own writing process.

1 - Initial Conception

When you're starting a new project, I recommend using Gemini to identify keywords and themes related to your topic. If you need help brainstorming ideas, ChatGPT is a great resource, especially when you guide it with prompts to generate original output. In my experience, Claude 3 is currently the most capable tool for exploring concepts and overcoming writer's block, particularly when you feed it relevant data.

2 - Research

During the research phase, you can use Otter and Gravity to transcribe calls and meetings, while Mem or Jamie can help you organize dictated notes. I believe there's a market opportunity for AI-powered content aggregators that can quickly gather relevant information for you. I've found that Claude3 reliably condenses lengthy research papers into concise summaries, and Scite.AI is useful for finding related academic papers. Keep in mind that while AI can identify simple patterns in data, deep insights still require your human analysis. AI-powered citation management, however, is quite reliable.

3 - Writing

When it comes to the actual writing process, I find that Grammarly Pro and other AI writing assistants are helpful for sentence completion, rephrasing, and word choice. Both Claude and ChatGPT can help you tailor your writing to a specific target audience and optimize for SEO. It's interesting to note that AI tools can mimic the style of well-known authors, but they struggle to adapt to lesser-known writers.

For Example, here is the “We all want to be original” passage rewritten in the style of Ernest Hemingway:

Too much to read. Too many books, articles, people. Time is short. Focus is limited. But you can't miss what matters. AI helps make more content. AI also helps consume it.

For researchers, AI sums up papers. Explains hard ideas. For investors, AI reads pitches and plans. Finds the good ones. Lawyers use AI. It writes meeting notes. Describes sports games.

Tools sum up videos and podcasts. What to watch. What to skip. AI has transcribed and translated for a long time. Now it doesn't make sense not to use AI. Let it take a first pass on speech, words, and meaning.

4 - Editing

During the editing stage, grammar and spell-checkers are certainly helpful, but I've noticed that overusing them can lead to generic-sounding text. AI can help you maintain consistency in style across a document, but be careful as it might reinforce bad writing habits if not used thoughtfully. I find readability analysis to be extremely useful for assessing clarity and complexity. Plagiarism detection is also important, though keep in mind that AI models are trained on others' work to some extent.

5 - Publication

When you're ready to publish, I strongly recommend automating your formatting with AI tools. While AI-driven content distribution platforms are not yet common for publishing, sentiment analysis of reader feedback can suggest necessary adjustments for you. There are also many AI-driven marketing tools available for creating targeted ads and social media posts to help promote your work.

AI in my work

I used AI tools in working with Adidas, Mercedes, NN Group, Standard Bank, and various others customers in researching and analysing a topic, and in writing everything from concept papers to product visions to project proposals. The quality of my work improved to the extent that I received glowing reviews from senior executives in those companies.

That’s because I use AI-driven tools at every step in my process. I do not take the output as is, but use it to refine my thinking and writing. Unless they ask me, I don’t tell them what tools I have been using. That’s not to say that I am hiding anything, it’s just that they do not care.

In a few years, different AI models (genAI, predictive AI, narrow AI) will be built into all the existing tools that people are familiar with. I can’t imagine any executive in a software company thinking “Yeah, let’s not do that.” Whether someone used AI in their work or not will be a moot point.

Risks of using AI

That said, I think people are still hesitant in fully utilizing what’s possible because they’re afraid of taking a hit on their reputation. They might have played around with the tools a bit, then gave up disappointed. Or they see the data and privacy risks and refrain from using them in the first place.

I don’t want to you to take unncessary risks, so here is are 3 best practices you can use:

  1. Document your workflow and the tooling that you are using publicly, like I have done with this newsletter. That way, people will see your tech stack and your process. You could even share intermediary steps to let people see how your output evolves through the use of these tools.
  2. Try the damn thing. All of the tools mentioned here have a free version. You can just spend a few hours every week exploring and experimenting. It’ll be much faster than reading up on them or eying them critically from the sidelines. The amount of time and money you will save by using them is ridiculous.
  3. Don't let AI tools take over your writing process. Supplement your own creativity, but review and edit carefully to maintain your unique voice. To stay sharp, set aside time to write without any AI help at all. Journaling helps in that respect.

If you liked this…

This is the 8th issue of the newsletter. I appreciate any feedback you could give me. I’d love to have a conversation on this in the comments with you.

Also, if you liked this, please share it and let other people know.

Cheers,

Ahmet

Impressive insights on AI-assisted writing, Ahmet—your experiences over the past year are incredibly valuable for anyone looking to enhance their writing process with technology!

要查看或添加评论,请登录

ahmet acar的更多文章

  • The Data Paradox

    The Data Paradox

    Welcome to another edition of Attain. In this article, you'll read about: How organizations are drowning in data but…

    1 条评论
  • Forget Innovation. Focus on Improvement.

    Forget Innovation. Focus on Improvement.

    Companies across industries obsess over innovation. Design consultancies, startup entrepreneurs, and corporate…

    9 条评论
  • inconsiderate: why customer experiences fail

    inconsiderate: why customer experiences fail

    I got soaked by the tropical rain rather than staying another minute at the restaurant. We all have horror stories of…

  • The Role of AI Tools in Modern Work

    The Role of AI Tools in Modern Work

    People have been trying to enhance productivity since the beginning of civilization. The plow, the loom, the wheel.

    2 条评论
  • Embracing the Product Operating Model

    Embracing the Product Operating Model

    While at AWS, I helped 3 different companies transition into that model while training their product managers…

    2 条评论
  • Why Product Management doesn't work.

    Why Product Management doesn't work.

    I recently read Pawel Huryn's breakdown of Marty Cagans 20 product model first principles. They are well structured;…

    10 条评论
  • The Rise of Service-as-Software: Revolutionizing Professional Services

    The Rise of Service-as-Software: Revolutionizing Professional Services

    A seismic shift is underway in the world of professional services. The advent of generative AI is ushering in a new era…

    2 条评论
  • On Quality

    On Quality

    Easter Edition: On Quality and why we don’t have it As I am writing this, I am sitting at tthe white Diani beach in…

    1 条评论
  • Product Management in 5, 4, 3...

    Product Management in 5, 4, 3...

    This edition is an article on Product Management in 5 levels of complexity. I'll explain why this is important at the…

  • On metrics and analytics

    On metrics and analytics

    Insights on using data TL;DR: How to work with data without expertise in data AI and NoCode Tools Data Collection and…

    3 条评论

社区洞察

其他会员也浏览了