Trying to Make Sense of the A.I. Hysteria
Whether or not you are in the technology or startup space, you’ve heard about Artificial Intelligence. You probably don’t know exactly what it means and what to make of it, but you know it’s all the rage. Every investor wants it in their portfolio, every entrepreneur wants it in their elevator pitch, and the average consumer is terrified of it either taking their jobs or killing mankind someday, or both.
My assistant randomly selected 100 companies on Crunchbase and found 87 of them directly reference AI or a related term (machine learning, deep learning, neural network) in their company description. This is causing a lot of confusion out there and diluting those that are truly advancing this field, advancing these technologies, and finding truly innovative applications.
The truth is, they are all probably telling the truth, and that’s because AI in its broadest sense has been around for a while. Search algorithm’s such as those used by Google are a form of AI. It is constantly scraping the web, learning about what is relevant and strong, learning your own behavior (search history, browser cookies, everything in your Gmail account), and many other signals to make decisions on its own, with zero human input, on what to show for each keyword at any given time and location. This is machine learning.
Without going through every use case out there today, you get the point — AI has been around for a while and now any company using the most basic form of machine-learning is plastering it all over their business to be part of this crazy trend. Even the more advanced forms of AI — neural networks, brain emulation, natural language processing, have also been around and studied and worked on for decades. (Side Note: Nick Bostrom’s book Superintelligence is a great overview of this world and technology for those like me with only surface level understanding)
So why then is it all the rage? What is happening at a broader level that is allowing it to advance so rapidly now vs. before?
3 Driving Forces
I should make it clear I am not an expert in this field. My interest and fascination with it is similar to many others in the tech and media space, but on a 1 to 10 scale, I’m probably a 2 to 3 at best. From what I can see, in the last couple years and in the 5–10 years ahead, I’ve notice 3 macro forces that are driving this current, massive movement in AI
A. More Data
There is more data then ever before. The advent of the smartphone 10 years ago made us humans connected everywhere we went. Facebook has built an empire out of social data. We are now connecting devices throughout our home. Our cars are becoming data machines and we are just at the beginning of that. We are heading towards a world where everything is connected and both providing and using data in some way shape or form. Pretty soon this sandwich I’m eating as I type this will be signaling neurons to my tastebuds into my cortex onto Elon Musk’s Neuralink and then back to the sandwich shop owner to tell him to ease off the mustard just a tad next time I’m in.
B. More Data Storage
Cloud computing has advanced the ways we store and hold data and process data. No longer bound by hardware to send and receive, which has both capacity limits and physical damage, everything happens now in the cloud in terms of data transfer and processing. As for the physical storage itself, that’s changing too. There have been breakthroughs in recent years to store as much as 215 petabytes of data onto DNA, and companies such as Microsoft are taking this very very seriously. Its expensive and a bit slow for now, but that will change.
C. Advanced Data Processing
The shift from CPU to GPU, thanks in large part to Nvidia and now being copied by everyone else in the chip and semi-conductor world, has been a massive breakthrough for AI. Think of it this way. Traditional CPU processing is sequential or linear in nature. Whereas GPU chips can process multiple sequences simultaneously — much like the human brain at a basic level. They were made popular for video games due to the processing and rendering needs, but now they are being used for advanced processing of all kinds.
AI vs. The Internet
Many of compared true AI as the next major revolution of, drawing to historical eras in the past such as The Internet itself, The Industrial Revolution, and The Agricultural Revolution. Major forces that forever changed the course of human history and everything that we do.
I say yes and no. There is some truth to this, but it is not apples to apples IMO. Here’s why.
- Like the Internet, AI will become ubiquitous in everything we do. Nearly all aspects of our lives will be impacted by these breakthrough technologies. So the ubiquity of it is very similar.
- Unlike the Internet, AI is not bringing a massive change in consumer behavior. It is not creating a new medium altogether for consumer behavior. More specifically, AI will make existing behaviors more efficient. Nearly all of them. Yes, there will be completely new applications here and there made possible by AI that could not exist without it — for example, wearing a translation device that allows you to orally interact in nearly any language in real time — but for the most part, these will be few and far between. This means AI is not going to bring about completely new, disruptive business models across the board like the Internet did, but those who invest in it will be able to make existing consumer businesses much stronger than those that don’t.
- True machine superintelligence- which we are very far off from, at least 40-50+ years at best according to most experts, in many ways will have a bigger long-term impact than any revolutions prior. This is because once we get there, continued advancement is not 100% in our, human, control. No, that does not mean the machines are going to turn on us (although it also doesn’t rule it out). It just means that in theory, getting AI to unsupervised learning and then to human intelligence where it can start to learn on its own and then to superintelligence where its advancing beyond our comprehension may very well be the final human created Revolution we ever see. From there on it will be machine led or machine-human collaboration, such as habitable living outside our solar system. By this standard, AI in its most extreme form outpaces all of these revolutions combined.
So What Does It All Mean?
As someone who is not directly advancing this technology at a company like Google or Facebook or Baidu or OpenAI, these are the takeaways for “regular folks” like myself:
- Learn the basics. Whatever business you are in, you will fall behind competitors in the decade ahead if you are not familiar with how these technologies can be applied to your line of work.
- Stop bullshitting that you are an “AI” company just because you use an algorithm somewhere in your business. Cause everyone does. And it would be easier for all of us t0 stay on top of the latest advancements if there was less abuse of this term.
- Data is everything. Your AI is only as strong as the data pool you have to work with. This is my biggest worry personally — that the tech giants will build walls around their data that nobody can compete with. Folks like OpenAI are helping provide parity on the technologies themselves, but not the data. This will have to come in the form of government regulation. Even more terrifying in many ways!
- For entrepreneurs, the great news is that many of the core underlying technologies, from Natural Language Processing software to advanced machine learning from folks like TensorFlow, are publicly available. So the art comes then in designing the applications. This is similar in a way to how in the last 10–15 years, the underlying technologies that built the world wide web became open source and fairly easy for all to use and access, making the application development the key in breakthrough businesses. We are entering a similar state with AI.
Most of all, enjoy it! AI I think will bring a lot more good than harm. The bigger threat will always be looming out there — but lets be honest, humans will always be the biggest threat we all face. A human is corruptable, a machine is not (except by another human hacking into it). Who knows, things may actually be even better and safer once we reach SuperIntelligence and most innovation is out of human hands :)