The Algorithmic Bridge in 2024

3 stats, 3 lessons, 3 changes

Here’s a toast for more digestible end-of-year post formats. I won’t review the state of AI (there are better resources for that) or what I published on the Algorithmic Bridge (I did once, almost died of tedium; and if I kill you, who’s going to read it anyway?)

Instead, I’ll look at three stats (everyone loves numbers), revisit three important lessons that marked me, and share three plans for the future.

I should remind you that the ongoing discount ends in three days. You should get it before the high of the new year makes you forget the best deal you'll ever get in 2024:

---

REMINDER: The Christmas Special offer—20% off for life—runs from Dec 1st to Jan 1st. Lock in your annual subscription now for just $40/year (or the price of a cup of coffee a month). Starting Jan 1st, The Algorithmic Bridge will move to $10/month or $100/year (existing paid subs retain their current rates). If you’ve been thinking about upgrading, now’s the time.

---

Three stats

Subscribers

This has been a big one for me because it’s been a small one for The Algorithmic Bridge: ARR grew by $8.5K and paid subs by 160. For most people, this is a lot. But it is not compared to TAB’s 2023 growth: ARR grew by $42K and paid subs by 830. This slowdown shattered my expectations and our ape brains hate nothing more than having their expectations shattered.

So I’ve had a stressful year. I got one patch of alopecia areata in my beard, which hadn’t happened to me before, not even during the screening exams of the aerospace engineering degree (a degree I never took advantage of except to be uncomfortably aware of all the weird noises during takeoffs).

I had to accept the “non-growth” of this newsletter, scrambling to find the root cause(s). I’ve managed more or less satisfactorily (I think I know what happened, both externally and internally, which I may share next year if my suspicions prove correct) and I’m fully focused on my work again.

Looking back I realize I had to go through this. Too much success too early can be counterproductive; you risk it getting to your head. Too much failure can kill your dreams so that’s not necessarily good either—but just the right amount saves you.

It was a humbling year for me and I’m stronger for that.

Articles

I’ve published a lot this year. 112 articles and essays. 35 weekly reviews (I stopped for a while but recovered it). And a few more bonus posts. A (rough) total of 280,000 words—or three and a half novels. Not bad. Some curiosities with takeaways:

• My most viewed article was one I published last week: OpenAI o3 Model Is a Message From the Future: Update All You Think You Know About AI. Takeaway: be quick to the things that matter.

• My two longest articles, OpenAI o1: A New Paradigm For AI and GPT-5: Everything You Need to Know (7k and 14k words, respectively) are, perhaps surprisingly, in the top 5 most viewed of 2024. Takeaway: quality trumps brevity.

• My most liked article was a rant: You Guys Have No Idea Just How Much People Hate Generative AI. Takeaway: people want their beliefs echoed back to them.

Writing

I’ve written every day. That’s 366 days. I didn’t publish as often but I worked on the newsletter as a full-time job. I either explored a new idea, reworked an early draft, did research on a topic that interested me, posted on my socials, or edited-polished an almost-done piece… I’m still figuring out a daily schedule though.

I’m doing this all by myself. No other help (either direct or indirect) besides the monumental support from both friends/family and you, my dear readers. The reason I’ve been able to do this consistently for three years without a formal background or previous freelance experience is that I love writing. I also love thinking and reading, which are the other two corners of The Blogger Trinity.

I understand the appeal of doing this part-time: the security of a 9-5 job, the stability of two independent sources of income, and the reassuring effect of having a cushion to fall back on if it doesn't work out.

I just know that it wouldn't have worked for me. I needed to jump into the abyss, blindfolded, for the adrenaline to kick in and bring out the best in me. I'm glad I did it.

Three lessons

The balance between being first and being good

For as long as I write about AI—a technical but fast-paced field—I will face a difficult trade-off: being first vs being good. I noticed this a lot during 2023-2024 when so many people decided to start an AI newsletter.

People who work on covering news, events, releases, announcements, etc. should strive to be quick with their reporting. When I waited too long to make something really good I often ended up with a polished piece about a topic that had gone out of fashion two weeks prior. However, haste is often the primary enemy of quality—they are inversely correlated. When I was quick I seldom added to the conversation, which prompted people to unsubscribe: Crap that goes out early is still crap.

Big outlets like the NYT have the privilege of embargo but independent bloggers like me don’t always do (it’s a rare thing). The only way to compete at the highest level is to be quick while maintaining quality. That’s how you grow and how you keep growing. I’ll strive to be both first and good but I need to keep in mind when to optimize for one or the other when doing both optimally is not possible.

Sometimes it’s more important to be first, sometimes it’s more important to be good.

Audience capture is silent (pursue your curiosity)

Audience capture is a common phenomenon for people who write online. As you grow big, you risk going from a place of “I do what I want” to a place of “I do what my audience wants” to avoid losing what you’ve earned. If the Venn diagram of those two categories stops overlapping, you will resent yourself for doing something you dislike and resent your audience for forcing you to keep them content. You’ve been captured.

The popularity AI got after ChatGPT sent bloggers like me down this slippery slope. You could get tons of engagement, nice metrics, and money just by turning into an AI influencer. Even if you didn’t go that far, you could still write about the hot topic of the week instead of the things you cared about. 2024 was my audience capture year: Seeing the numbers dwell, I stopped thinking about what I wanted and began worrying about how to fix it. So I leaned into what I thought interested my audience.

I believe now that the first mistake I made when growth plateaued was not doubling down in pursuing my curiosity. After all, that’s what brought my readers here in the first place. You don’t want to chase your readers but chase your instinct and intuition, and they will follow alongside you. To fix mishaps, be more honest with yourself, not less. Or you will betray yourself for nothing.

(You may want to read this article by Gurwinder Bhogal.)

Focus on consistency, not growth

I’ve always been obsessed with stats. Everyone is. But when they trace a nice upward curve we don’t notice the addiction. We’re just happy. Unfortunately, 2024 drew across my subscriber dashboard the horizontalest line I’ve ever seen. I couldn’t get it up. Lol. I started checking my growth stats every day. Even on days I hadn’t published (so it was unreasonable to expect the number to change at all).

I felt worse. My obsession materialized as an annoying inability to concentrate on my next article. I saw, helpless, how my craft degraded along with my trendline. At some point, as if I were in a time-agnostic dream, I started questioning whether my quality dwindling had been a consequence and not actually the cause of the slowdown.

I made a drastic decision: I would enforce a complete avoidance of stats. I bookmarked my drafts folder, ensuring Substack’s green-turned-gray monthly deltas couldn’t worm their way into my brain. I dodged every temptation fate threw at me to sneak a glance. I even removed Notes from my home feed. At last, I could face my unwritten words on a blank page and let the rest of the world fade away.

No obstacles. No anxiety. No distractions. It worked.

Three changes

I plan to publish more often

With a three-times-a-week publishing schedule, I have many more things to say than time to share them (dozens of could-have-been-good drafts are gathering dust, half-written, in my too-late-for-this folder).

I’ve experimented with publishing four pieces a week and didn’t get the expected return, but I believe it was due to an unintentional reduction in quality, not the increment in quantity. I will only increase quantity if I can keep quality intact.

I know readers don’t like to be bombarded with more things than they can read. At the same time, I find myself wanting to read more from my favorite writers however much they publish. I commit to staying among your favorites so this works out.

I plan to explore styles and lengths

I’ve been doing reviews, timely news pieces, and long deep dives since I started.

I want to do more storytelling. Not strict fiction, but real stories intermixed with philosophical takes and abstract themes; like a good New Yorker essay would read but without the pristine craft. Also satire/humor. I’ve already published a couple in a new section I’ve entitled “Algorithmic Jest.” I firmly believe AI needs more jest right now.

I want to keep exploring different lengths as well. One deep dive after another gets tiring for me to write, for you to find excuses not to read them. I will go as low as 300-400 words (e.g. light humor or quick newsy briefs) and as high as 10k-15k words for the deepest dives (e.g. overviews of important AI models like GPT-5, o1, o3, etc.)

Nah, some things don’t change

At some point, I’ll explore other topics. Four years of moderately successful writing have taught me non-AI lessons others might find useful. But I’m almost certain 2025 won’t be the year for that. AI is still the most important subject, with too many pivotal events expected next year. (And the years after that.)

Models o1 and o3 reshaped the landscape, introducing a paradigm that OpenAI and its competitors will continue to advance. China has caught up to the US, heating AI geopolitics; governments will issue statements beyond policy and regulation. Social media is deteriorating, with Meta setting up plans to encourage AI-generated bots on Facebook and Instagram. Meanwhile, Google DeepMind’s Veo 2 has achieved a new state-of-the-art in AI video, now indistinguishable from human-made content.

Expect plenty of coverage from me on all things AI in 2025. As ChatGPT fades from the spotlight, staying vigilant will matter more than ever. Before ChatGPT, there was little of note happening for most people. During its heyday, keeping up was effortless; AI progress was all anyone talked about. But after ChatGPT, a fog is quietly seeping into our feeds, their newsrooms, and those polished statements by AI labs. A dense fog of uncertainty, doubts, and fear.

I happen to be a fog lifter. At your service.

— Alberto

A real picture of me with my fog-lifter cloak lifting some fog