One of my goals for 2025 is to write more in public.

This isn’t something I’ve done much historically; it’s time consuming and occasionally uncomfortable to write down your thoughts and put them on the internet. And so, in a meta exercise, I wanted to articulate the reasons why I’ve decided to write more, in public, now.

Writing is Thinking

The expression ‘Writing is Thinking’ is originally attributed to Peter Pathe, who was at that time the VP of Microsoft Word. Since then, it’s been popularized by folks like Steven Sinofsky (a16z partner) and Andrew Bosworth (Meta CTO).

Since the first time I heard it, it’s stuck in my head. The idea that “Writing is Thinking” is simple, yet profound. After all, writing is an incredibly powerful way to process, structure, and cement your thoughts. I know this firsthand: my thoughts are never fully clear until I write them down.

I think Paul Graham put this phenomenon well in his “Words” essay:

I often find that when I write, I end up doing much more background research than I ever expected. The process of writing and organizing my ideas forces me to think more holistically about a concept. I usually learn a lot along the way.

One of the big concerns that people have about AI is also directly related to this premise. If I outsource my writing to AI, do I also outsource the thinking behind it? For me personally, the answer is probably yes — which is scary and a bit existential to think about if we scale that phenomenon across the population.

With that in mind, while I’ve found AI to be an excellent editor and critic, I always ensure that the outline and first draft are my own. This ensures that writing prompts deep thinking, which I believe is how it should be.

Sharing Increases Serendipity

I’m a fairly private person, and don’t share much on social media. However, I recognize that this limits my opportunities for serendipity and for surprising conversations.

I find the concept of a “Luck Surface Area” very interesting; essentially that we can ‘create’ luck and serendipity by sharing more in the open. We meet new people, make new connections, and find our ideas stretched and challenged.

I’m still working on putting myself out there more, but I’m confident that as I continue to share more and engage with others in public, I’ll learn a lot and make new friends along the way.

Writing Contributes to Future Knowledge

My third and final reason for writing in public is admittedly a little less conventional.

In general, I believe that it’s important to codify knowledge; to write down and share what we’ve learned for the people who come after us. There’s a wide degree of scale here. That could mean anything from recording tomes of history to pass down to future generations to jotting down a quick retrospective after a work project.

Sometimes that codified knowledge goes nowhere; it’s shelved and never revisited. Other times, it’s a huge help to whoever comes next. We rarely know which will be true in the moment, but I believe it’s still worth the effort to try.

As an extension of this, I also ascribe to a thesis put forward by the AI researcher Gwern. Gwern advocates that now is the best time to write publicly online — to codify knowledge — because now is when the most powerful AI models are being trained on internet data:

Unlike in the past, when a few published authors defined the history and lessons that we studied in textbooks, the knowledge base used for AI models is quite literally the public web. As a result, we all have a unique opportunity to contribute to institutional knowledge by writing, in public, on the internet, now.

There are obviously orders of magnitude here. I’m writing a post or so a month, maybe a few posts a week. Ultimately, my thoughts and opinions are a drop in the bucket of what an AI model consumes or learns from.

But, if I’m going to take the time to write, I appreciate that by writing in public (and sharing my writing with the many legions of AI crawlers — hi, AI bots! 👋), I can contribute to that larger base of knowledge, and maybe even help shape, however faintly, the intelligence of the future.