Rohan Silva’s Post

What can the Victorian era teach us about the future of AI? One of the most interesting - and counter-intuitive - lessons from the history of technology comes from a 19th century English economist called William Jevons. He studiedwhat happened to the usage of coal as technological improvements meant factories and other industries were becoming increasingly coal-efficient. Not unreasonably, Jevons assumed that these efficiency improvements would mean that coal consumption would fall.But weirdly, across every industry, the opposite was happening.Factories needed less coal to manufacture goods, which meant those products could be sold for less. But that quickly led to consumers buying more, and that increased demand meant more coal was used overall.It was a similar pattern with steam trains. Efficiency gains meant they used less coal, so the price of train tickets fell. As a result, more people than ever wanted to travel by train, so a lot more train tracks ended up being laid, and the total amount of coal used overall went up. This counterintuitive pattern has come to be known as Jevons paradox. And if you look for it, you can find it everywhere in the tech world today. For example, cost of mobile phone data has plummeted - but that hasn’t cut bills; we’ve just invented new uses for data (like online gaming and video streaming), so overall data consumption and spending on data has massively increased. I’ve got a feeling we’ll see the same pattern play out if - as DeepSeek suggests - AI models get cheaper and more efficient to build.  That suggests investors were totally wrong to dump Nvidia stock after the DeepSeek news broke recently - and I’ve written about this in more detail for the Australian Financial Review. Link below. (The column was published a week ago, but even so, I know I’m a bit late to the DeepSeek party - apologies!) https://lnkd.in/gZ8UUc6e