First Principle - Respect Lindy
I love heuristics. Simple rules that get you most of the way there. However, we are not content with simple, powerful, and effective. Instead, we love to complicate our lives with the "latest and greatest": ChatGPT, Virtual Reality, Twitter 𝕏, etc.
Whenever shiny new things compell you, stop, take a breath, and think back to your Grandma's "Rules of Thumb" - more technically known as "The Lindy Effect".
The Lindy Effect is a heuristic that states:
The robustness of an item is proportional to its life! ¹
That is, for any non-perishable, you can expect it to be around in the future for roughly as long as it has lived to today. A book published 50 years ago will be around another 50. A book published 2000 years ago will be around another 2000.
Nassim Taleb popularized The Lindy Effect in his 2012 book, Antifragile. Since then, Taleb has gone on to prove the Lindy Effect mathematically in his paper "Lindy as a Distance from an Absorbing Barrier".²
Example - Nubank
What does this have to do with a startup or software engineering? Our industry obsesses over "the new"; but your best friends are the tools, processes, and practices that have been around a while.
Nubank is one of the largest Fintech companies in the world. Though young, Nubank distinguished itself with valuable products. How did it do this? By building out a Double Entry Accounting system on top of a LISP dialect: that is, implementing a 600+ year old technology using a 60+ year old programming language.
Obviously, Nubank does a lot right outside of these two "non-perishables", e.g. product-market fit, timing, focus on cost-conscious engineering, and innovation mindset.
The point isn't to say that success requires older technologies. But you must always research them first.
Exercise - ChatGPT
ChatGPT dominated the first half of 2023. Almost all of my curated non-fiction podcasts, programming or pop-culture, drafted in its considerable wake. Though ChatGPT's hold on the zeitgeist is waning, let's use this as an opportunity to critically analyze this technology for "Lindyness".
ChatGPT itself was launched in November 2022, so - as a whole - it's only been around 8 months.
However, we can dig deeper. ChatGPT is a "large-language model" with a "chatbot" interface.
Large Language Models (LLMs) are relatively new because of the massive compute required to get to the "large" qualifier; however, "Language Models" have been around a LONG time. For example, the Language Models in Speech Recognition appear in commercial products in the 1980s and trace their origins to the 1950s and Bell Labs.
Chatbots also seem to be a new technology since their wide adoption coincided with social media and synchronous business messaging tools like Slack. However, chatbots go back as far as the 1950s when Alan Turing published his famous "Computing Machinery and Intelligence" paper.
I'll stop here for today. Suffice it to say, were I to make a prediction for ChatGPT based on this analysis, I would say:
- ChatGPT itself will be replaced by something new within the next 8 months.
- That replacement will be larger, use language models, and have a human-language interface.
tl;dr
Often we dance into the great beyond with the hubris that no-one but we know have ever tried to solve the problem before:
G.K. Chesterton authored this principle, referred to as "Chesterton's Gate" (translated to Twitter 𝕏 Quote Syntax):
There exists [...] a fence or gate erected across a road.
THEM: I don't see the use of this; let us clear it away.
YOU: If you don't see the use of it, I certainly won't let you clear it away.
YOU: Go away and think. Then, when you can come back and tell me that you DO see the use of it, I MAY allow you to destroy it.
Advice
Before choosing the latest and greatest, first consider the ideas that have stood the test of time.
NOTA BENE: An old technology must have some usage to prove robustness, i.e. don't pick a 100 year old technology that no one has used in 90 years.
For example, when picking a library for your programming language of choice, scour the documentation for references to prior art. The best libraries are written by authors who've researched what exists.
Lindy is a Rule of Thumb and Rules of Thumb are Lindy.
It's OK to choose boring technologies.
Notes
1
https://archive.org/details/isbn_9781400067824
2
Few philosophers have stimulated me as much as Taleb in the past decade. I don't agree with everything he says, but he has made me think in new ways. Speaking of Lindy, my two other favorite philosophers of the past decade are Hannah Arendt and Ursula K. Le Guin, whose books are 50+ years old.