Welcome back to the second episode of the year and maybe an important one to look back on, this is hopefully going to get pretty interesting. Also apologies that this is slightly later than usual, just landed in Istanbul!
After some reflection it became clear to me that there's one problem which has persisted throughout my life, which is universal, and which if (properly) solved would not only measurably improve the lives of everyone around me but also positively contribute towards building a competitive advantage for large scale corporate enterprises. I'm talking about knowledge acquisition, and specifically the ability to capture and retain information in a way that unlocks it for long term use and decision making. We're going to cover a lot of ground but I promise it all fits together.
Some background; I recently watched a video summary about the Alien movie franchise (100% relevant I promise). If you're unfamiliar it centers around a collection of ventures by an interstellar mega corporation called Weyland-Yutani. This was the result of a merger between Weyland Corp (a robotics research and planetary colonization company) and Yutani Corporation (a cybernetics research company). This got me thinking about businesses in terms of decades and centuries, and about the types of problems companies like this would face (and therefore perpetually open problems in the world). Even in the technologically advanced world of Alien, with near-AGI androids capable of scheming and deception, humans still drive the majority of decision making in these organisations. So what might these futuristic companies be doing differently which we aren't doing today? What does their onboarding look like? What does their company Confluence/Wiki/Intranet look like?
In the age of AI, cultivating human intelligence within an organization is crucial to long term success and a core competitive advantage (I believe) for a few reasons. Knowledge systems built using language models (e.g. enterprise ChatGPT) are useful, but what really matters is human decision making (from broad impact C-suite decisions through to day-to-day role specific micro decisions). If we look more closely at human decision making, it's based* (roughly) on fluid intelligence and crystallized intelligence. Fluid intelligence refers to generic and abstract problem solving, in the context of decision making this means 'can someone make the best decision based on available information'. Crystallized intelligence refers to the amount of information available to the person in their mind , in decision making it means 'does someone have the most available information'. From this framing it's clear that the best outcome both at an enterprise and personal level is to have people who rank highly in both fluid and crystallized intelligence nearby, and have them call the shots. Yet it's also clear that if you had to choose between someone with more of either one, someone with higher crystallized intelligence would give higher quality decisions since 'seeing more' context around a problem is more valuable than better reasoning within a limited context. In other words, a great chef with four ingredients can make an okay meal, a good chef with ten ingredients can make a great meal.
Also important to consider is that language model based knowledge systems are cool, but they don't directly contribute towards decision making since the information needs to be in the mind of the person making the decision in order for them to be able to consider it (the ingredients need to be in the kitchen, not at the supermarket). Of course, you could type 'what do I need to consider when deciding between X and Y' into your company chat bot, which is probably reasonable, but is it reasonable to do that every time and for decisions at every scale? In engineering and in management, I would prefer to work with someone who is highly knowledgeable in person and defers lower level tasks to AI, than someone who is for lack of a better phrase, an AI powered idiot 👀. So how do we make people more knowledgeable in a systematic and reliable way, and how would Weyland-Yutani do it?
Lets say I was a hiring manager or HR executive and saw the value in upskilling my teams, what could I do? Knowledge acquisition is a very broad field, with language 'learning' companies like Duolingo, Babbel, Rosetta Stone, and others doing their best to deliver shareholder value, but I would argue that they're more interested in edutainment and serving ads than in delivering robust education. There are also more generic tools like Anki, Quizlet, and others which are okay for the individual but each have their own limitations. Then there's the elephant in the room which affects all of them, and it's a modern word I'm a fan of; enshitification. It's a natural evolution of every piece of software, more users means more feature requests and a stronger push for monetization, which means paywalls, ads, and a diluted user experience as the core purpose of 'education' slowly but inevitably shifts to 'user retention'.
There are also of course the domain specific quiz apps which you'll probably have used and thoroughly enjoyed. Compliance and corporate policy things like 'Jane took a briefcase of cash to help land her client's son an internship at your company, is she a) acting responsibly and transparently, b) exploring alternative recruitment strategies, or c) about to be in big trouble'. These tools are fine, but if we're interested in cultivating human intelligence in a way that leads to high quality decision making within an organisation, we'll need something a little more powerful and flexible.
What I think would be useful is a tool with all of the powerful learning tricks like spaced repetition, free recall, gamification, and all of those things, but generic enough for any domain and secure enough for commercially sensitive knowledgebases. Imagine being able to upskill yourself and your team in a way that isn't completely mind numbing, and that gives you insights into things like which people are strongest in which areas, or onboard new joiners to your culture, jargon, and core information without burning several weeks on pair programming and endless presentations. And the applications are broader still, for individuals it could be used for interview preparation, exams, PhD vivas, important presentations, and more. For enterprises it could mean much more, in the trading world for example it could bring everyone up to speed on historically significant trades, modelling approaches, confounding factors, tickers, competitors, and things like that, all culminating in the highest possible decision quality at the human level which is ultimately all that matters. I think having smarter people in the world is a valuable and worthwhile pursuit, so over the last few days I've been building a prototype which is everything I would want in a knowledge acquisition platform, so if you're interested in what that looks like please let me know, and if you're in the learning and development team in your organisation I'd love to talk about what features you'd look for in something like this.
Just to close, I'm ultimately building this tool for myself, since I can feel a slow but constant sense of atrophy in the breadth and depth of my knowldege. And it's not started recently or as a result of this solo entrepreneurial adventure, it's been happening since I graduated, and before then from the second the exam or coursework finished in all of the knowledge I acquired up to that point. Learning things is fun and fulfilling, but it's the repetition, recall, and usage of these things which commits them to long term memory, and ultimately allow them to inform (consciously or subconsciously) your decision making. You only have to get that right once to permanently add to your crystallized intelligence, so there's everything to gain and nothing to lose.
Anyway I hope this has been interesting, I'm off to get the most authentic Turkish Doner I can find and will report back how it goes.
Thanks and all the best,
Oliver
* if you subscribe to Cattell's early model of human intelligence
Previous Episode: Vietnam In Review