AI: Hype or Not Hyped Enough?

The more I learn how something works, the more the magic fades away

It’s really hard to ignore the AI hype at this point. Everywhere you turn, there is a new “AI” tool launching or a new startup getting funded by Venture Capitalists ¹. The AI evangelists preach about how this thing is going to change the world. It is going to 100x our productivity. It is going to take your jobs. You will do a lot more with less. It is the best thing since slice bread. Blah Blah Blah Blah. The list goes on and on.

With every groundbreaking technology, there will always be the early adopters and zealots. On the other extreme, you will find the skeptics - those that are comfortable in what they already know or do and feel this new thing is overhyped and will eventually fail. In the middle are the realists ². I see these folks as the ones who bring balance to the universe. They do not care about which particular technology wins nor do they jump on every hype train. They are rooted in the fundamentals and are genuinely curious about how it works. They are the ones who truly like to understand the value this new thing brings and really does not care whether it fails or succeeds. They’ve been around for a long time and have seen many technologies rise and fall.

In finance, these are people that invest in boring index funds instead of ride the emotional rollercoaster that comes with picking the hottest new stock. Hello, Nvidia. They know that time in the market is better than trying to time the market. Now that’s the key to understanding any new technology - time in the market. AI is an umbrella term that groups Machine Learning (ML), Data Mining and Data Science etc. Large Language Models (LLM) like GPT4 just became mainstream but the concepts and algorithms that underpin these technologies have existed for a very long time. I have found that as someone who is in this space for the long haul and not here to make a quick buck, it is much better to invest your time in learning how “AI” works and what capabilities it brings. The more I learn how something works, the more the magic fades away. Over time, you naturally can tell if this is a fad or if this truly can solve new problems and improve upon older ones. Another advantage is you know the limits of the technology and can call bullshit whenever someone tells you to buy an AI powered robovacuum for example ³. Don’t get me wrong - I do see the power of large language models as a reasoning engine and a shortcut to information. There is no doubt that their reasoning capability will ultimately get better over time. We just all need to learn to separate the signal from the noise.

So, I’m I a zealot, a skeptic or a realist? Comment below.

Side rant: if a recommendation system says they are now AI powered, what does that actually mean? Recommendation systems have used Machine Learning models to power predictions for God knows how long. If I change the system to solely use a Large Language Model (LLM) or use the LLM in tandem with the previous model, I honestly see this as an improvement to the system - or a software upgrade. ML is still under the AI umbrella. Slapping on “AI powered” does not change the fact that it is a recommendation system - albeit a better one. This kind of marketing, in my opinion, is what fuels the AI hype. Products add AI as a way to gain traction and get VC funding even though the product is still what it is even without “AI”. I’m not judging but I’m kinda judging you. In fact, what does an AI powered budgeting app even mean? Anyways, let me shut up. Until next time.