One year on Substack

I started Chronicles of the Future on 21 December 2023. Since then, I’ve written 44 posts containing 53,115 words, covering developments in AI, climate change and demographics. I thought it was a good moment to reflect on what I’ve learned from one year of Substacking.

A couple of years ago, I read Gal Beckerman’s fascinating book The Quiet Before: On the Unexpected Origins of Radical Ideas. Combing through history, Beckerman showed how some of the world’s most transformative ideas and movements – whether that be the Enlightenment, Chartists, or third wave feminism – started off with a small number of cranks debating away from the crowd, where they couldn’t be overheard.

In an age of viral social media content, Beckerman worried we were losing the quiet before, with new ideas being exposed to the light before they were ready, derailing the mechanism driving social and intellectual progress.

Now, I’m always sceptical of theories suggesting that change is now impossible, whether that be because of capitalism, neoliberalism, or, in Beckerman’s case, social media. It’s lazy, cynical conservatism that fails to imagine a future different from the present, despite that being one of the only constants in history.

As much as I enjoyed Beckerman’s book, one of the gratifying experiences of this year has been witnessing how he got it wrong. There are new ideas gestating, and even if they sometimes receive attention before their time (and get – sometimes fairly – whacked by hit pieces in traditional media), they’re often able to weather a flash of scrutiny. Beckerman’s concern may have said more about the temporal moment of cancel culture rather than something intrinsic about how we will communicate on social media going forwards.

The best example of this has come in AI. In June, former OpenAI employee Leopold Aschenbrenner wrote a 100-page essay, Situational Awareness, on the state of AI, which I covered at the time, arguing that superintelligence is inevitable, a national security issue, and that the US government needs to take control of its development similar to the Manhattan Project.

Aschenbrenner spent the next few months talking to people deeply embedded in debates about AI, such as when he appeared on the Dwarkish podcast. Fundamentally, he shifted the debate among AI development experts from ‘fast versus slow’, to ‘fast is inevitable, it’s just about whether we go fast through central government or not’ (with the likes of Marc Andreessen favouring a decentralised approach).

In November, the United States–China Economic and Security Review Commission (a US quango) has published its annual report. Its top recommendation was for Congress to “establish and fund a Manhattan Project-like program dedicated to racing to and acquiring an Artificial General Intelligence (AGI) capability.” That would be Aschenbrenner’s Manhattan Project in action.

The quiet before still exists, history is still being written and change is possible. Things just move faster now.

In addition to this blog, I do some political commentary in the UK. Perhaps this is idiosyncratic to a country in decline, but the more I’ve learned about policy debates, the more obvious the solutions appear. How can we fix the housing shortage, for example? Just build more houses, duh! The problem is that these solutions are blocked by various forms of institutional failure. My naïve assumption going into this process was that learning more about my three chosen topic areas would lead to me discover similar ‘truths’. Needless to say, I’ve been disappointed.

I’ve come to realise that AI development really is a black box, even for the people building it. We don’t know if scaling laws will continue, and because of that, the range of scenarios to plan for are almost uselessly large. As such, I’ve spent most of the year looking at the changing consensus, both in what is expected to happen and how the technology is currently being used. Anything else, particularly from a non-expert, seems like baseless speculation.

Climate change is different. The equation is simpler: Xbn tonnes of carbon = Y degrees of warming (feedback loops complicate this, but so much of that science – at least to an amateur – seems contested, that I’ve generally decided to ignore it). The central debate of whether to reduce emissions through degrowth/reductions targets or green energy abundance hinges on three important variables. Can technology save us? Despite being partial to boosterism, this remains unclear. Can politicians make people stomach lower living standards. Again, unclear. And thirdly, will countries work together to reduce emissions? Anyone should be sceptical.

Lastly, throughout the year I’ve despaired at the lack of workable solutions to global fertility collapse. The issue is still far from public consciousness, conventional natalism likely doesn’t work, and all other suggested solutions are unfortunately just crackpot. Maybe this is strange, but of the three areas, this is the only one which I’m fatalistic about.

But then again, my uncertainty could just be a byproduct of following subjects without the necessary technical expertise to know the solutions when I see them.

I started this Substack as a way to hold myself accountable for reading deeply about topics that interest me. The idea was simple. Knowing I had to write something coherent by the end of the week made me more likely to commit time to reading. I wasn’t aiming to influence anyone, which is why there’s been very little marketing involved.

Even so, on the rare occasions when I felt I’d stumbled upon something resembling an original idea, I shared it with those who I thought might be interested. It’s been truly heartwarming to discover how many people are open to engaging with strangers who show a genuine interest in their work.

There’s a lot of talk about the growing toxicity of the public sphere, especially since Musk’s acquisition of X. But in my experience, if you seek out people with interesting ideas, engage deeply with their work, and respond thoughtfully, most will engage in good faith. Here’s to letting the best ideas win.

*

I’m taking a break from weekly posting to focus on a few other projects. But I’ve found this exercise incredibly useful, and I may return to it later in 2025.

Thank you for reading and enjoy the holidays!