In his 2024 Black Hat talk, Moxie Marlinspike (the cryptographer and founder of Signal) spoke about some ways our world is different to the ones portrayed in fantasy stories like JK Rowling’s Harry Potter books.

Moxie starts by noting that these books ask us to imagine how exciting life would be if only we lived in a world where all one needed to do magical things was a magic wand and access to a library. And yet, in the Harry Potter franchise, most the students seem quite bored to be at the school where they teach you all the magic, and seem quite unmotivated to learn about all the amazing knowledge they have access to.

A natural reply, and one I have seen before, is that the real world is also like this. We have access to all sorts of exciting knowledge, that can allow us to do amazing things, and yet, often, we are not excited to make use of it. Moxie’s reply to this is that the real world and these fantasy stories differ in a notable way.

there’s something that feels different about our world and the Harry Potter world... To me the big difference is that in the Harry Potter world all they need is knowledge of the spells and a wand. They don’t need to assemble and motivate large groups of people they don’t need permits they don’t need to figure out how their magic fits into a larger economy they don’t have to allocate large sums of capital in order to cast their spells. In our world if we try to cast a spell we’re going to need more than a wand and that is what makes us Muggles

He goes on to talk about an exception to this, which is what it felt like to be a kid learning to code and growing up with access to a computer. Instead of a magic wand and a library all you needed was a laptop and access to the internet.

In his talk Moxie also talks about how these kinds of dynamics can fundamentally change the shape of a field, beyond just making it easier to start out. Experienced computer programmers benefit from these effects too - they can very cheaply throw together little experiments, and, perhaps even more essentially, they can develop a unique kind of deep understanding through deep cycles of repeated tinkering. Breakthroughs seem much more likely to arise in fields that work this way, and indeed we see this in the explosion computing breakthroughs that resulted from the tinkering “hacker culture” of the 80s for example with early innovations in computer graphics and AI.

In as much as the individual hacker is constrained by a lack of resources, they are also rewarded with an abundance of creativity and understanding. Their speed of iteration can be much faster, and they are able to be driven purely by passion and curiosity. It has been many centuries since a similar approach was even thinkable in the natural sciences.

There are fields outside of science which have this blessing, for example writers, philosophers, and some kinds of mathematicians. Every now and then a mathematical breakthrough will still be made by an amateur, purely driven by their own curiosity.

A thing I have been thinking about lately, and which I’ve been confronted with at pretty much every stage of my career has been that even in computer science this “hacker” mindset can only get you so far. At some point assembling and motivating large groups of people, getting the necessary permits, figuring out how your work fits into a larger economy and how to attract and allocate large sums of capital become essential to doing things that matter. This has become even more true in the field of computer science with the rapid scaling of AI training from experiments that can be run on a desktop to massive megaprojects costing sizeable fractions of GDP.

Moxie’s conclusion is that contact with these kinds of problems of scale risks losing the magic that made computer science such an amazing field the first place. However, the more I read about the history of the natural sciences in 17th century Britain, the more I think that perhaps this contact goes both ways - and that there is a synthesis to be found between the unconstrained individuality of curious tinkering, and the carefulness and coordination that is required to do potentially dangerous things at huge scales. Computer Scientists needed to learn these skills at one point, before the explosion of commodity computing hardware, and maybe they will benefit from learning them again.

More to come on what I think that looks like.