- Published on
The Matrix: Golden Age of Machine Learning
- Paul Mestemaker II
The Matrix is upon us
I remember watching The Matrix when it first came out. I was on spring break with my family in Newport Beach, CA and we were told that it was home to one of the largest screens in the western hemisphere. So instead spending more time at the beach, we went to The Big Newport to see The Matrix on a 71-foot screen.
I know Kung Fu
I played a ton of sports growing up and even took some martial arts classes. But I had to practice for thousands of hours. Needless to say, I thought it was the coolest concept that Neo could just download moves directly into his brain and not have to physically practice.
Kung Fu vector embeddings
I've spent the past few months playing around pretty extensively with several machine learning frameworks. No joke: I've repeatedly felt like I'm Neo from The Matrix. Every week it seems like I can leverage my existing traditional software engineering knowledge and put in minimal effort and "download" a completely new ML capability into my brain. Here is a smattering of the things I've been able to build prototypes with:
- Image generation
- Image upscaling
- Image classification
- Object detection & tracking
- Audio transcription
- Audio cleaning
- Natural language processing
- Text vectorization
- Semantic search
- Recommendation engines
- Reinforcement learning
- Text generation
This is the golden age of machine learning. I've been able to build a few things that I thought were limited to elite teams of people just a few years ago.
If you have an idea but do not know how to code, don't worry! There are going to be 10x more no-code tools and 100x more low-code capabilities (thanks to large language models) in existing tools than there were a year ago.
If you have an idea and do know how to code and have the appetite to learn Python, then you're just a few hours and a couple of
pip installs away from prototyping some absolutely incredible $#!t!
p.s. while I was writing this, it was not lost on me that a lot of these capabilities are powered by matrix operations under the covers. 😅 What a time to be alive!