@adlrocha - Zero-AI Writing
To recovering craftsmanship in the age of generative AI
A lot has changed since the first time I published in this newsletter:
The first two years were great! I was publishing every Sunday. I was learning fast and was full of ideas. This newsletter was the perfect way to solidify what I was learning and share my views about the fields I was working on with the broader community. Many of you may have stuck around since those early and prolific times (if this is the case, thank you!).
I then joined the ConsensusLab team at Protocol Labs, and I had so much on my plate that writing a post every weekend was completely unfeasible if I wanted the so-called “work-life balance”. Before, I was used to draft the post throughout the week, and then review it and publish it on the weekends. With all that I had going on, I had to draft, review, and publish my articles all on the weekend. So it goes.
In the end I only enjoyed half of the weekend. This was really time-consuming and it was taking a toll personally. This is why I decided to abandon my weekly habit, and start writing sporadically instead of forcing myself to do it weekly. I was only writing when I had something to write about and worth sharing.
Until I went completely silent for over a year. As part of my 2024 new years resolutions, I decided that I wanted to recover my writing habit. I was having my first kid. My time at Protocol Labs was coming to an end. I was trying to figure out what was next for me, and seeing all the AI craze it looked like a good excuse to start learning about the field and writing about my learnings. For this new stage of the newsletter, I decided to change the format a bit: I would host my core publications on my personal blog, and share it as part of the newsletter post along with other interesting links I may have come across throughout that week. This experiment didn’t last long, from the beginning of the year to mid February.
I didn’t even have time to figure out what I wanted to do next. What I was going to do next found me in the form of an opportunity to build a really cool company with two co-founders that I really admire and that I have a lot of fun working with. Building a startup is definitely a roller-coaster, and as you can imagine, I spend all my waking hours thinking about it: from product development, to go-to-market, to people and money problems.
Reclaiming craftsmanship!
Which brings me to today. We are still heads down working on making Baselight the universal data hub for humans and AIs. In the process I’ve learned so much (including AI), and explored so many ideas, that I now feel the urge to start writing again (let’s see for how long).There is also a more selfish, philosophical, and aesthetic reason for me to want to start writing again: my increasing saturation with AI-generated content.
These days I feel like all the content that is shared on the Internet is created by stitching together a few LLM prompts, don’t you think? While easy to produce, it ends up being boring to consume (at least for me). All the content feels devoid of personality. Too structured. Too well-formed. Lacking any kind of personal touch. In many cases it is like coming across an amalgamation of vomited facts. It’s definitely useful in many contexts, but we are making readers and writers lazy in the process. I really think that they are harming our ability to think and create.
Even for technical pieces, I enjoy when you can feel the author’s pain researching the topic, their chain-of-thought and reasoning (pun intended) while writing the piece, their style. I even enjoy slight typos, incoherence, and lack of structure that end up triggering the analytical (and judgmental) side of one’s brain.
This indiscriminate use of AI is becoming even worse in the tech field, where it feels like if you are not using AI you’re staying behind (and in many cases is true). But there’s still beauty in human craftsmanship. I don’t know if there are tech nerds out there like me craving for human-written content. If they are, this new stage of my newsletter is for them. So far, all the publications from this newsletter have used zero AI (except some of the images that I use to make the posts a bit more appealing, which are AI-generated). My goal is for this to be the case moving forward.
This may even become the new normal:
If people end up reading these pieces, enjoying them, and learning something new in the process, that’s the best outcome I could’ve achieved!
If not, at least I’ve managed to experience myself feelings that may soon be taken from us by AIs: like the blank page syndrome, rewriting a sentence seven times, starting a piece without the slightest idea of where it’s going, or sketching the structure for a post and throwing it to the trash five times in-a-row.
With my full-time devoted to making the vision for Baselight a reality, I feel I have more ideas than the ones I can execute. Hence, I thought it would be useful to try to start writing periodically once again and get feedback from these ideas and hopefully inspire someone to start building them to make them a reality :)
Finally, I was lying about not writing anything since 2024. I wrote a lot in my hand-written notebook (I should publish this at some point) and about Baselight in Baselight’s “Letters from the CTO”. Unfortunately, I never found the time to share them here. It’s time to share all of that and more here once again! See you in a bit, and feel free to reach out to connect.
PS: I accept bets about when I will be publishing the next piece, and how long it will take me to become completely silent again. Hopefully, I can do my best to make you lose those bets (in a good way :) ). If someone opens a prediction market for me let me know, I want to trade it!