In random thoughts style:
It started with a headline for a news magazine editorial stating that, in the author's opinion, 2023 would come to be known as the year of AI. Rather than read the accompanying article, I ran a google search for the topic, sorted by date, to see what was up. In the search results were a large number of references to something called ChatGPT. Curious, I read a bit farther and found that it was an app recently released by the company OpenAI. The descriptions still left me curious to learn more. I tried a smartphone app, then uninstalled that and added the ChatGPT app as an extension to my Chrome browser. Success!
Having had a few days to play around with this difficult to describe thingamabob, I will try and describe it. Let me start by quoting a description from the AI itself when queried, "What is ChatGPT?" It replied, "ChatGPT is a large language model trained by OpenAI. It can generate human-like text, and can be used for a variety of natural language processing tasks such as language translation, question answering, and text summarization." That was the first reply to our question. For any given response, one can click "Regenerate response". Do that ten times and you will get 10 different results. My point is, the AI isn't simply regurgitating a canned reply to a given question. From a user standpoint, this means that if you don't like the way it responds to any given question or command, keep trying.
ChatGPT is free, for now. In the first five days after release, more than a million new users signed up. It took Twitter 24 months to reach the same milestone. I have read that OpenAI is using feedback from the current release (ChatGPT3) as data to improve the next one (GPT4), which is already in the works. The current version is not perfect. In fact, some mistakes are rather, well, dumb. As an example of a gaff, I was trying to see if it could tell me if "my tree" was a male or a female cultivar. It knew about the tree and knew that it was dioecious, yet it confidently stated that trees "don't have sex like animals and don't produce seeds or fruit"! It also has no access to the internet, relying instead on the data and training fed to it by the OpenAI team. This data only goes through year 2021 so events and info after that year are not available. In spite of these shortcomings, I will opine that ChatGPT's strengths outweigh its weaknesses. I found its knowledge base to be impressive, broad, and very detailed. It has great potential to save a user time and effort, and is actually fun to communicate with. It remembers previous questions and comments from both sides of a conversation. Let's say, for example, that you want to write a letter to a group of people thanking them for a donation to a charity. You might say to ChatGPT, "Compose a thank you letter thanking recipient for their generous donation". If the resulting letter is ok, but a bit dry and lacking certain details, you might say, "Once again, but this time make it longer, mention the name of the charity, {insert name}, and make the tone very friendly and conversational". The ability to give this AI instructions in plain English and have it understand them is one of its most useful features.
As a senior, who was an adult when the internet was born, I find myself amazed at the pace and magnitude of change in the ensuing years. I don't know whether AGI will be a good thing, or not, when it finally makes its appearance. I do feel certain that, like the advent of the internet and the world wide web, it will be a game changer
I asked ChatGPT the following question: How long before AGI is achieved?
This was its first reply: "It is difficult to predict exactly when AGI (Artificial General Intelligence) will be achieved, as it depends on many factors such as advances in technology, research funding, and the specific approach being taken. Some experts believe that AGI could be achieved within the next few decades, while others believe it could take much longer. It's also worth noting that AGI is a highly debated and controversial topic in the field of AI, and there is no widely accepted definition of what constitutes AGI." Something tells me that it may be here sooner than we think.