Back to the early days of computing, but this time it’s for everyone

How AI is democratizing technology, just like the GUI did

I recently listened to a very interesting episode of a podcast called “Beyond the Prompt” (episode link: Death of SaaS and raise of what marketing folks can do with AI already with Noah Brier). This particular episode discussed the applicability of AI and how to best leverage it to your advantage. While I’d highly recommend giving the episode a listen, I’ll dive deeper and break down some ideas on how you can leverage AI in your daily life. Don’t let the title of the episode fool you- it’s not just for marketers. It’s a great listen for anyone.

The AI Revolution: We're Still Early!

It’s important to remember that AI is still in its early days of adoption. According to a survey of 5,000 people by the National Bureau of Economic Research (link to article), as of late 2024, about 40% of working individuals in the United States said they have used Generative AI, 23% of working respondents used it in the last week, and about 10% use it daily. This is rapid adoption, but we still aren’t in majority numbers. A lot of people still haven’t interacted with a large language model (LLM) before and most are still trying to wrap their arms around a strategy for interacting with them. I’ll dive deeper into strategies and methods in future articles, but for today let’s explore where AI is most valuable.

The Blank Screen Problem: Overcoming Initial Overwhelm

The value of AI is hard to see on the surface. When you go to ChatGPT, Gemini, or Claude, you’re greeted by a blank screen. Maybe it asks you “What can I help you with?”, greets you by name, or maybe it has a couple of buttons you can click on, like “search”, “analysis tool”, or “summarization”. This is an overwhelming experience. So many questions come to mind if I’m coming to one of these tools with little knowledge beyond what my friends have told me about them, like “it’s an advanced version of Google search!”, but is it really?

This experience, for better or worse, brings us back to the early days of computing before the Graphical User Interface (GUI) was invented. The GUI democratized computers by allowing anyone to click through an app and take productive actions without needing to know how to program. Just as the GUI democratized computing, AI will democratize access to information and the ability to build simple applications. The difference today is that AI is so much more accessible than a command console because you can ask for anything using natural, native, language. No special coding language needed. So that brings us to the existential dread of wondering when you can ask for anything…what in the world do you ask for?

Let’s start with the most powerful areas you can leverage generative AI in and maybe that will spur some ideas:

1. Transforming data

2. Facilitating human collaboration

3. Increasing accuracy

4. Enabling faster ideation

5. Gathering feedback

Notice above how I mention nothing of searching for information or facts? LLMs are not like Google search at all and should not be used like Google search. If you need to find an answer to something, you should still rely on Google to find quality sources, Perplexity is another option if you want your answer to be based on actual sources you can see. Although LLMs are trained on vast information from the internet, they are generally not citing their sources (unless you are using a search specific app, like Perplexity), they are very good at writing things that sound plausible, but are not true, and they tend to act as an echo chamber, confirming our own biases, rather than offering dissenting views or criticism.

Transforming data

How much of your job involves taking unstructured information and putting it into a specific format for another audience? Any time you think you could use a magic tool to convert information into another format, AI is your answer! I get my highest returns on productivity in this by far.

Maybe you just wrote a report or whitepaper for your manager and now you need to turn it into a presentation for your VP or an executive in the C-Suite. Give your paper to the LLM of your choice and ask it to create an outline of a slide deck based on your paper. Make sure you specify the main points you want to communicate in the presentation and who your audience is- both of these details are important to communicate context to the LLM. Once you have the outline, you can follow up with more prompts to focus on building out the presentation slide by slide. Normally I ask it to generate content for each slide, referencing the outline it has already produced for me to help further guide it, then once I have the presentation where I like it, I’ll go through and have it generate speaking points for me.

Another example is let’s say you have a Google Doc that has data in it that you need to convert to a table/spreadsheet. Instead of methodically going through your document and manually copy/pasting data into a spreadsheet, you should feed the document to the LLM and ask it to extract your target information and organize it into a table or CSV file for you.

In your personal life, maybe you need to turn a recipe into a shopping list. You can take a picture of the recipe in your recipe book, share the picture with an LLM and ask it to tell you what you need for ingredients. Or if the recipe already gives you a recipe list, as most already do, you can take a picture of the ingredient list, share it with an LLM and ask it to transcribe the ingredient list so you can easily copy/paste it into your grocery list on your phone. While doing this, you can tell the LLM to ignore the portion numbers and just give you the whole ingredients, like if the recipe calls for 1 cup flour, just have it list “bag of flour” or “flour” in your grocery list.

Transforming data is by far the most powerful way to increase your productivity using AI. So much of knowledge work is about shape-shifting information. Handing these tasks off to an LLM allows you to spend more time on higher-level tasks.

Facilitating human collaboration

How many times have you stopped listening and talking in a meeting so you could focus on writing down a crucial point accurately? Most people cannot multi-task, so when you can offload your note-taking to a robot, you can be that much more engaged in a conversation.

Aside from letting robots take notes for us, AI can be leveraged to summarize meeting notes, discover misaligned priorities or contradicting decisions, and help you assign tasks out to proper subject matter experts. One big opportunity is to organize all of your documents for a specific project into a notebook in NotebookLM and presto, you have a chatbot trained on all of your information regarding that project that you can now query to surface crucial information quickly or generate new information and content.

All of the above helps you focus on communication. When you can think less about collecting and organizing the information you need to communicate, you can think more about how you are going to communicate that information. What’s the best way to frame it? Is a presentation the best format? Who is going to be consuming this information? What do I want to advocate for when providing this information to my audience?

Increasing accuracy

Just like we use the “Find and replace” tool in a Google Doc to find instances of certain words and change them to best suit the doc we are writing, so they are accurate for that specific audience, we can use LLMs as much more nuanced find and replace tools.

With an LLM, we can update a document by saying “remove all instances of numbers and replace them with a high level reference to the number that is still contextually relevant, but does not disclose the exact number, for example, I would replace ‘65’ with ‘more than 50’”. Or maybe you want to quickly remove all personally identifiable information, instead of specifying the names or addresses you want to remove, you can just tell an LLM to remove personally identifiable information. It should understand well enough to scrub the document of everything that could be used to identify someone.

In this respect, LLMs give us the ability to provide vague but specific directions and have it take direct action.

Enabling faster ideation

Brainstorming and ideation is another super power of LLMs, because ideas can be generated quickly and cheaply. Brainstorming can give you such a high return because the stakes are so low. Ask it to generate 10 ideas on X and it will riff with you until you can’t take any more. The ability to play with and bounce ideas around without taking up a coworker's time allows us to flesh out ideas longer before pulling other people in.

Gathering feedback

So many people use LLMs as writing machines- why don’t you use it as a reaction machine? I loved this point by Noah! Instead of asking LLMs to talk for you, why don’t you ask them to act as your editor? They can tell you how you come across and if what you’ve written is clear and understandable.

For the next week, any time you’re about to shoulder tap a colleague to get their opinion on something, like how you’ve written a memo, email or the project plan/proposal you’ve built, I want you to ask an LLM for a reaction instead. Just make sure you ask it to be brutally honest to override its system instructions to be a helpful and pleasant assistant.

Wrapping up

The benefit to using AI really comes when you leverage it for vertical-specific knowledge you’re already very familiar with. I can talk to you all day about methods and strategies, but the real power is going to come when you use AI to help in areas of knowledge that you are already very knowledgable about. This allows you to verify if what an AI has written is accurate or not that much faster. In addition, you can leverage your industry knowledge to access the specific parts or “experts” inside of an LLMs knowledge base.

LLMs are basically like super brains full of individual plausible experts. They are trained on vast swaths of the internet and have knowledge about so much information that it’s impossible for any one person to be an expert in how to use LLMs for every part of a business. In other words, it takes an expert like you to get the most out of AI for a particular role or industry. By learning to integrate AI tools into your work life, you can become an industry leader by figuring out the best ways to leverage AI in your specific domain.

If you have yet to try using an LLM in your personal or professional life, I highly encourage you to give it a try today. You have nothing to lose and so much to gain by adopting this technology early. I’ll link some of the most popular frontier models below.

• Google’s Gemini (strong all-around tool, deep Google integrations)

• ChatGPT (strong all-around tool, great for quantitative analysis)

• Claude (excellent for writing and coding)

• Perplexity (replace your Google searches)

I’ll plan to write more articles about leveraging AI for productivity. Let me know in the comments about what you think of this one and if you have any burning questions about how to best leverage AI in your personal or professional life you’d like me to dig into in a future article.