Last week, Artificial Ignorance hit 10,000 readers.
When I started writing about a year and a half ago, I wasn't sure what all this would become. I can't say I had many expectations, and I certainly didn't think I would reach this point. But I wouldn't be here without all of you reading, so first and foremost: thank you.
Memories
For much of 2022 and 2023, I was living under a rock. I was vaguely aware of GPT-3 and Stable Diffusion, but I couldn't have told you much about how they worked or even who made them.
And then came ChatGPT. I, like billions of other people, was shocked. I had studied computer science and machine learning (though my ML was quite rusty), and I thought we were at least five to ten years away from something like ChatGPT.
So I started digging. As I went, I kept finding incredible discovery after incredible discovery - Midjourney, ElevenLabs, Claude, and eventually Sora. It felt like being a kid in a candy store with new technology. It reminded me of when I first learned to code, and every week brought some new skill that I could now wield.
One night, I turned to my partner and said, I think this generative AI thing will be a big deal. I don't know exactly how big, but I’ve got a feeling that if I don't learn it and understand what it’s capable of, I’m going to regret it for a really long time.
That’s been the motivation for this newsletter from day one - not to grow an audience or make money, but to discover and learn for myself what’s actually happening at the frontier of generative AI. In turn, to distill and teach those lessons to others.
Momentum
The single most common piece of advice that Substack gives you as a new publisher is: “Be consistent.” They even send you a weekly email by default to get you to keep up the streak.
When I started writing, I took their advice to heart (plus, I do well with deadlines). So, for the last 73 weeks, I’ve published (at least) twice a week. Usually, that’s in the ballpark of 2000-3000 words each and every week. It hasn’t always been easy: I’ve done it through getting sick; I’ve done it through family vacations; most recently, I’ve done it through interviewing for a new job (while holding down an existing one).
At this point though, the schedule feels more constraining than beneficial. Sometimes, I want to cover a piece of breaking news on a Sunday, but it doesn't "fit" the schedule. Other times, I have content (like interviews) that I've held onto to have a fallback post to hit the schedule. And there have been many, many times when I'm scrambling late into the night on a Tuesday, wondering if what I have is good enough to publish.
Moving forward, I want to be less hard on myself. I still intend to publish twice a week - sometimes even more than that. Friday roundups will stay the same (at least for now). But otherwise, I want the flexibility to work on a less rigid schedule, and to work on bigger and more thoughtful pieces that might take longer than a few days to put together.
Modifications
Besides a scheduling tweak, I still want to try more things with this publication. None of my ideas are very concrete, but I think they would make sense if done well: I have little desire to launch a paid Slack community just because everyone else is doing it.
Interviews. I enjoy talking to AI builders and thinkers in an interview format, but I can't help but feel that there's a deeper way to engage with guests. The current format is fine, but I wouldn't call it exciting. Can I test a product, build an app in tandem, or otherwise go deeper with my podcast guest? I also think a lot about how to surface interesting ideas and AI perspectives outside the norm.
Community. When done well, communities are great. When done poorly, they're awkward ghost towns. But building a great community usually means being focused - and I (still) don't know what the "focus" of this newsletter is. It's definitely "AI," but "Slack/Discord/Substack Chat for people interested in AI" doesn't seem narrow enough to feel compelling. Alternatively, I could do one-off community events like workshops, presentations, or Q&A over Zoom.
Education. I love teaching; it's one of the biggest things that motivates me to write. Yet I'm left wondering whether there's a way to teach something "bigger" - an email series or a video course, for example. A million people are trying to teach AI - from Google to OpenAI and everyone in between. Is there a piece of education that I'm uniquely suited to creating and that I have the resources to make a reality?
Monetization. As I said, I didn't start this for the money, and I still don't. But I believe that if you do something well, it is worth figuring out how to monetize it. I'm going to continue experimenting with different monetization strategies, starting with the 50% discount on a paid subscription that’s running through the end of the month.
I strive to make as much of my content available for free as I can, with paid subscribers being "power users," so to speak. Many of these ideas I'm considering will likely have additional benefits for paid subscribers, whether it's joining office hours or previewing new educational content.
Collaboration. Right now, this is a one-man operation. I would love to work with other folks to try and make more of the above a reality. If you're reading this, I want your feedback, comments, thoughts, and critiques. If you like one of these suggestions, if you want to collaborate in some way, or if you have an idea for a guest post, I would love to start a conversation.
Once again, thanks for being a part of the journey - and here's to another 10K!
wow my 3 fave substack authors in one chat :) Charlie, when I first started writing a year ago you were my first motivator on Substack as I - coming from tech and similarly gobsmacked by AI - couldn't learn enough. @whytryai came soon after.
fwiw coming from a smol pub I don't think Substack's beat you on the head consistency message is all that critical. As long as you keep putting out good content weekly your audience and impact will continue to grow.
One thing I wonder for you guys as I look at the increasingly crowded and noisy space of AI pubs is how you stand out or differentiate? I think that's both your oppty and risk to grow the newsletter into something more.
10,000 subscribers is fantastic, Charlie! Congratulations, even if gaining subscribers or getting eyes on your stuff initially was not the main goal.
Very much like you, I learn by writing and researching, and I'm really motivated to get the facts right! Even better, a consistent group of folks who are willing to read my stuff and give me some feedback (or, more often, just bring the conversation further forward) is out there ready to help me every time. We both have that from what we're doing, and that's so cool.
Keep on rockin', and if there's anything I can help with, just gimme a shout! Very well done.