GPT-4
Recently, GPT-4 was released and the first hints of its potential have become apparent. What is even more disturbing is how much of an improvement in understanding it has over GPT-3.
Any rational person observing this is thinking, “What comes after this?”. As someone interested in software development, I have seen countless videos of people dismissing GPT3 and now 4 as “not sophisticated enough to take our jobs”. Yes, this is true, obviously. Making remarks like this and not pointing out the clearly exponential growth of this technology is a brain-dead take in my opinion.
The future of AI
So, where do I think this is going? AI language models and other specialized AIs will almost certainly displace 90% or more of any job that relies on pure information. Software engineers are almost exclusively working by thinking and then writing code. THAT is what produces results. These don’t require robotic implementations, like if a Plumber or Electrician were to worry about AI. (They should worry too, robotics is also exponentially growing)
5 years ago, AI was a pipe dream for being useful. However, GPT-3 less than 1 year ago already showed practical use cases. With GPT-4, it’s already clear how much progress has been made over 1 year. I am not even going to mention the implications of MidJourney, DALL-E and other image generation AI technologies.
Exponential folds
Exponential growth means that the progress you make in the next fold, is equal to the progress made in ALL previous folds. In fold 1 to 2, this is like going from 1 to 2. Obviously not a huge jump in comparison to what existed. What happens when you’re on the 30th fold? It’s 1,073,741,824. It took 30 doubles to get there and in ONE more to be ~2 billion. This doesn’t ever slow down, it just gets faster.
If Moore’s law being proven over the last few decades is anything to go by, we have seen how exponential growth in computing has impacted our accessibility to cheaper and faster computers, phones, tablets, and the like. In my opinion, AI will, in my lifetime, take most of the world’s jobs. That’s a good thing. People used to know how to make computers, and now that is a niche skill. Programmers today write code but, people of tomorrow probably won’t need to or be able to do that as widespread as we have now. There are many types of work that existed 100 years ago that don’t today.
Why would you want to work when you can get everything you need for cheaper and cheaper every year? That is perfectly possible with exponentially improving technology. The caveat is inflationary money. If your money buys less, then you NEED to work. With Bitcoin, this wouldn’t be necessary.
AI singularity insurance
Bitcoin is the perfect money for a world with exponentially increasing technological improvements. Bitcoin has an exponentially decreasing supply with a fixed cap. Money with those properties will be able to absorb all the progress and preserve your purchasing power into the future.
When, not if, AI becomes autonomous to the extent of buying its own materials needed for manufacturing, operating costs, repairs, and service fees, it will require digital money that has no friction. Bitcoin/lightning offers this solution. There is also a case to be made that this AI will interoperate globally and possibly even have a preference for what money it wants to use to increase efficiency and minimize risk unto itself. This makes Bitcoin the logical choice.
Bitcoin is the only monetary path forward that allow us to accrue the value of AI on society at a personal level. Bitcoin as a thermodynamically closed system is a topic deserving of its own article and has been discussed by others (Michael Saylor most notably). The more progress we cram into 21 million units, the more valuable the units become. AI will almost certainly expedite this process for humanity. Don’t hold money that won’t account for this progress, and almost more likely do the opposite for you and decrease your purchasing power into the future.