In 10 years
October 17, 2024
We are currently living through an exponential curve; AI has been advancing faster than any of us ever imagined. If I told myself 10 years ago what we would have today, I would not believe it. Alan Turing built the first computer on February 9, 1946, and he would not have imagined in his wildest dreams that in 80 years' time, we would have a computer that has passed the Turing test.
What will happen in our next 10 years?
These are interesting times we are living in. Now, we have seen another period of exponential growth. If you see how these Neural Networks have scaled for the past 4 years, you would begin to understand the world from my point of view. In 2019, GPT-2 was released; it could barely even make a 3-sentence paragraph make sense. Now, with GPT-4, Claude 3.5 Sonnet, and especially O1 models, they are outperforming even human PhDs in their specific fields. We have never seen a time where intelligence is evolving this fast. This is only 4 years, and back then nobody even invested in AI. Now, we have literal trillions pouring into AI research.
We are on the edge of something that is beyond our wildest dreams. We are living in a time more exciting than the industrial revolution. Just remember this: In 4 years' time, we will look back at current AI models like how we now see GPT-2 models. That's why I believe that it's totally logical to assume in 4 years' time, we will have AI models that can completely surpass any human intelligence.
They will be so alien in a way they will perform things that we cannot even understand or imagine. This is because AI intelligence scales linearly, while our intelligence doesn't even come close to that. If we have a computer with 10x more compute than what we currently have, we can train an AI that is 10x smarter, but if we get 10 humans together, we are way far off from being 10x smarter. This is the problem with human intelligence. It doesn't scale. However, AI does.
We will live in a utopia
We are living in perhaps one of the most historic times in human history. We are witnessing the rise of a new type of intelligence. One that would solve most of the problems in our current society. AI will begin curing diseases that we thought were incurable; it would most likely solve global warming and world hunger. It is just that we are unable to think of ideas that a superintelligence would think of.
The biggest problem with us is "We have Paleolithic emotions, medieval institutions and godlike technology."
AI might not even have to invent something new to solve lots of our current issues, because it could actually think in a modern way, instead of a mindset that has been the same for 10,000 years.