We are faced with an interesting decision! Now that LLMs are here how do we proceed. In order to fully grasp the impact of what AI will do for humanity we need to look at an earlier example to ease our way to the conclusion. I take you back to Feb 2005 when Google Maps was first launched. Today it is ubiquitous and even people with no educational background intuitively use Google Maps to arrive at unfamiliar locations without having to stop and ask for instruction. How did we reach this place. More importantly how did we manage without it. A unique advantage of being my age is I remember how things used to be. My generation has had a front row seat to how we used to live before technology became so pervasive. I used to have a fairly decent ability to navigate to any place regardless of how far it was by storing a complex set of instructions in memory and being able to recall them faithfully whenever needed. Today I don’t even bother to remember instructions and pretty much the first thing I do is enter the location in Google Maps and then just rely completely on Google Maps to get me to my destination regardless of traffic, weather and any other obstacles that may come my way. By outsourcing a simple task such as navigation, have I freed myself of having to remember complex instructions? The simple answer is no. I still remember the instructions and I still want to be able to navigate assuming that at some point there might be a signal issue that drives Google Maps from working.
Does AI make life too easy?
I have simply switched from making tactical turn by turn decisions to more strategic big picture decisions like where to stop for lunch and what other places I can visit nearby. Today’s large language models bring a unique dimension to how technology can either support and enable human endeavours or make us lose out ability to perform complex tasks. Know how is a perishable commodity, just ask NASA. With how ubiquitous AI is becoming we run the risk of performing as task only as recommended by the AI model or preferring the solutions suggested by AI to such an extent that we lose the know-how on how to tackle the problem another way.
If you think I am being overly dramatic ask yourself when was the last time you clicked on page 2 of a Google search result. Our preference for a readymade solution is overwhelming and left to our own devices we will always switch to the path of least resistance.
Tech has evolved now it’s our turn
Essentially the human brain is elastic and needs to be moulded by challenges in order to retain its elasticity. This means that the less we use our brain to perform complex tasks the more likely it is to replace those complex tasks with other tasks that don’t provide the brain with the exercises it needs to stay on top of its game social. The brain constantly needs something to do. The quality of the work we give it dictates the quality of the brain.
I feel that we will be given two choices once AI matures
The first one to completely outsource thinking to the large language model and depend on it to provide answers to mundane routine questions that we ask or need to perform resulting in our brains rotting away due to lack of exercise.
The second we can exercise our brains to stay one step ahead of AI.
So what’s next?
The ability of the human brain to adjust, adapt, learn and be flexible will always outpace the trillions or quadrillions of transistors that support the Neural Network for LLMs and AI in general. The human brain has over a 100 billion neurons. ChatGPT has over 100 billion tokens of which one request can access up to 128,000 tokens. In the end we need to accept that AI can make our lives better but what we do once our life is better still a question that we need to answer.
The other major decision point that we will encounter as a result of the pervasive use of AI is the fact that the divide between the have and the have nots will increase exponentially as AI can be seen as a force multiplier allowing tech savvy people to triple or quadruple their efficiency and leap ahead of everybody else essentially creating an imbalance that we may not be able to overcome.
In spite of all the doom and gloom that I’ve just mentioned the fact remains that AI is a very handy tool and inevitable as far as progress is concerned. The problem is it sneaked up on us so quickly that we weren’t prepared to answer some fundamental questions about how it should be used how it’ll impact us and what we should do in the future to address any shortcomings that wasn’t foreseen before. Recently Microsoft and a number of other technology companies came together to lay down the foundation for the ethical use of AI. Governments all over the world are bringing in legislature on how to use AI as well. The question remains about the ethics of using AI. I don’t believe we can leave it to politicians and CEOs to arrive at a conclusion that is fair for humanity in general. It will be interesting to see how things shape up in the future