Why I Avoided AI—And How I Finally Embraced It

As someone who writes about AI for a living, Rhea Purohit has a unique perspective on the challenges of incorporating these tools into our daily lives. In her inaugural piece for Learning Curve, a new monthly series on AI adoption, she explores the psychological barriers that have held her back from fully embracing AI tools—and how she's overcoming them. By breaking down her own hesitations and sharing concrete steps she's taking to integrate AI, Rhea offers a roadmap for anyone feeling left behind by the AI revolution. Her column will help you understand why even tech-savvy professionals can be slow to adopt new tools, and teach you practical strategies for overcoming your own AI anxiety.

In addition to today's column, last week we launched Sparkle, an AI organizing tool, and there’s more new stuff to come. If you’ve been waiting to subscribe to Every, now is your time. For the next 24 hours, become a paid subscriber for just $132 per year—a 33 percent discount—and get the best bundle of writing and software on the internet.—Kate Lee

I write about AI for a living, but I struggle to use it myself. 

Each time the thought of using AI to research and write crosses my mind, it’s followed by a quiet reluctance. It never seems like the right time, and I honestly don’t know if it’ll be worth the effort.

This paradox has been weighing me down for a while, so I went on a mission to understand why it exists. Here’s what I’ve learned so far: 

I struggle to use AI because I don’t like the uncertainty that comes with using a new technology.  Once I realized that this paradox isn’t rooted in inauthenticity or incompetence, it freed me to find ways to change. And soon, I started running my own experiments with AI, like: 

• Using ChatGPT to teach myself philosophy
• Fine-tuning a Spiral, a prompt builder that Every launched, to generate social media posts
• Discovering meaningful ways to include Claude in my writing workflow

Being able to make sense of why I felt friction in integrating AI into my workflow is helping me get better at deriving value from it. If there’s a small, tinny voice at the back of your head, saying that you could be doing more with AI, perhaps my little epiphany can help you do the same.

Searching for the why

I started by thinking about what I do every day and how AI could help me with that. As a writer, I:

• Read what other people have written
• Think things through 
• Put one word in front of the other

Something I write about often are insights from AI & I, Every’s podcast about how smart, creative thinkers use AI. I spend hours every week listening to how they’ve integrated these tools to become better versions of themselves. 

I’m acutely aware that LLMs can help me read, think, and write. AI carries the promise of doing great work in less time. It sounds like magic…but I still don’t use it as much as I should.

My hunch is that this is because I wrote for a living before AI became accessible. I know how to go from blank page to finished piece without using a LLM. I have workflows in place—reliable and efficient—to get the job done. They typically involve coffee, a Google document, a few stray research tabs, and many short walks. It’s familiar, comfortable, and—perhaps most importantly—I know it works.

AI, on the other hand, feels like work. It’s like hiring your first employee. Yes, they will eventually make your life easier, but there’s a lot to be done before that happens—you have to think about what you want to delegate, find the right candidate, and onboard them into your organization. Something similar happens when you start using AI. You have to figure out what parts of your workflow you want to automate, choose the right tool for the task, and iterate on the input you intend to give it.

And there’s the other thing: AI is an objectively new technology. We’re in the early days of experimenting with it, and right now, there is no one true way to use it. At least not yet. It has been known to generate different results depending on whether your prompt is in ALL caps, if you yell at it, and even what day of the week it is. There is emerging research on the best ways to use AI, but for the most part, it’s all up for grabs.

In other words: 

• You have to invest time and effort into experimenting with AI.
• As you do this, you cannot be too attached to the outcome, because the result is uncertain.
• When you add a deadline and the realities of adult life to this, it's a recipe for I’ll try [new AI tool] out next time, but for now, let me just get this done.

The allure of the familiar

Now that I’ve identified the source of my friction, the uncertainty that AI brings to my workflow, I want to go deeper. I’d like to think I’m open to trying new things, so why am I drawn to things that are familiar to me?

Well, it turns out that humans are wired to prefer the familiar to the uncertain. And researchers have found that this is even more pronounced if there’s an element of finality involved in the decision.

Here’s a thought experiment.

What if I told you that you could have a conversation with anyone right now? It could be your favorite author, the last speaker of a lost language, your two-year-old’s imaginary friend…you get the gist.

That sounds pretty exciting, right?  

But what if I added just one more condition: This would be the last conversation you’d ever have. 

Has your answer changed? My guess is that you’d choose to talk to someone you know, love, and care for. I know I certainly would.

According to a recent research paper, when humans are faced with a narrowing window of opportunity to enjoy an experience, their preferences shift from the shiny, new, and novel to the old, familiar, and comforting. 

This speaks to the type of experiences we find meaning in. In the words of the researchers, “variety may be the ‘spice of life,’ but familiarity may be the spice of life’s endings.” The study was done in the context of how people choose between hedonic activities, or the things you do for a shot of pure, unadulterated enjoyment. But I think this applies to something that falls outside the definition of hedonism (or at least the traditional one!): work. 

When faced with a deadline (narrow window of opportunity), I have a tendency to choose tried-and-tested methods (old, familiar, and comforting) over a suspiciously helpful tool (shiny, new, and novel). 

Opening up to uncertainty

This tendency is bad because these tools are fragments of a future that is taking shape around us—and by ignoring them, I’m setting myself up to fall behind. But there is a bright side: I’m staring down the problem and understanding why it exists, which is the beginning of solving it. I need to be more comfortable with the uncertainty that AI brings with it. These are the three ways in which I’ve opened myself up to AI:  

Take it slow 

Learning a new language as an adult is similar to weaving AI into familiar workflows. While learning a new language, you’re basically learning new ways to do familiar things. You’ve been to a grocery store many times, but suddenly, you have to find new words to ask when that fruit you like will be in season again. Similarly, when it comes to AI, you’re attempting to use new tools to do things that you’ve done many times before. 

I’ve been learning to speak Spanish for the past year, and the piece of advice that has helped me the most is: little by little (or, to quote my Spanish teacher, poco a poco). After a few months of language classes, it makes sense to be able to fluently order a dish in a restaurant. But it also makes sense to fumble while expressing complex thoughts or feelings. I try to remember this when I’m using AI. I’ve taken away the pressure of upending my workflow to integrate AI. While writing this piece, for example, I did not use an LLM to brainstorm an outline, generate a first draft, and get feedback on it from Dan Shipper and Kate Lee. But I did have Claude open in a tab, and I ended up asking the model to help me rephrase a few sentences I thought sounded awkward, and for feedback on a couple of paragraphs—and it was genuinely helpful! I’ve given myself permission to take it slow.

Design your own curiosity 

Curiosity has the reputation of being something that you either have or you don’t. I don’t agree. I think it’s more fluid than that—a quality that ebbs and flows, one that you can guide and nurture. 

I started writing about technology because I was fascinated by what humans could achieve with a few lines of code, bits of silicon, and their imagination. Taking that to its logical conclusion, the drudgery in using AI shouldn’t bother me so much; I should be far more excited about experimenting with it. But the truth is that the realities of adult life can get in the way—and sometimes, I lose focus of what fascinated me about technology in the first place. In other words, my curiosity ebbs. 

To counter this, I’ve started carving out time to nurture my curiosity. One of the items on my daily to-do list is reading something interesting on the internet, even if it has nothing to do with what I work on, especially if it has nothing to do with what I work on. And every Friday is dedicated to experimenting with an online tool I found cool that week. Sometimes that ends up being about AI: making a song about my dog Oscar, who I grew up with and now lives many miles away with my parents, or training a custom GPT to extract insights from a podcast transcript. I think curiosity has the potential to compound—and it’s worth developing a habit around it.

Let your environment shape you

Humans are social creatures, and for better or worse, we’re strongly motivated by the people around us and their habits. If the people around you are excited about AI, you’ll probably be nudged in a similar direction. 

I’m lucky enough to work with Every, where everyone is excited about finding ways to do great work with AI. So I see a lot of this unfold up close. Over the past few months, here’s what I’ve witnessed:

• I remember reading the Discord messages Dan sent the team as he built the first version of Spiral over a weekend.
• When the images Lucas generated about the future of technology went viral on Twitter, we spoke about how he uses AI to explore his creativity at our weekly standup (and eventually how to incorporate them more into our work). 
• In the channel for writers on the team Discord, our editors Kate, Meghna Rao, and Scott Nover are fine-tuning an LLM-powered copy editor that’s trained on Every’s style guide. 

It’s subtle, intangible, and slow, but being exposed to all this experimentation around me primes me to run my own experiments, the ones with ChatGPT, Claude, and Spiral, that I spoke about earlier.

We live in a world where you can choose who you want to surround yourself with intellectually. The internet offers access to thinkers who can inspire and challenge us—lean into it.

I’m not the most sophisticated user of AI—check out Dan’s writing for that perspective—but in learning how to open myself up to AI, I’ve integrated it into my life in surprising, useful ways. Writing this piece reminded me that so much of our skepticism around using new technologies is not practical, but deeply psychological. In this series, I’ll chip away at the latter and document my insights, in the hope that it will gently nudge you toward opening up to AI as well.