OpenAI’s artificial intelligence “ChatGPT” is named as such because of the acronym “GPT” which stands for “generative pre-trained transformer”. The generative pre-trained transformer provides the characteristics of any large language model like ChatGPT. The term “generative” refers to an artificial intelligences “ability to generate new content which is similar to the data sets it was trained upon”. The term “pre-trained” refers to the “practice of artificial intelligence models being fine tuned for specific tasks, a pre-training phase in which the large language model learns from a vast amount of text data which helps the model understand language patterns and contexts”. The term “transformer” refers to the transformer architecture which is a “neural network design that relies upon a mechanism referred to as “attention” to weigh the influence of different parts of the input data”. This transformer architecture is particularly effective for tasks which involve understanding the context of language (e.g. language translation or answering questions etc.) which is why artificial intelligence models are able to understand human language in a deep and complex way
Tim Berners-Lee created the internet. Berners-Lee is the son of mathematicians, his mother and father part of a team who programmed the worlds first commercial stored program computer, the Manchester University Mark 1. Berners-Lee developed the original concept for the internet as a young boy, after discussing how machines might one day possess artificial intelligence with his father who was reading a book upon the human brain. Berners-Lee realized that if information could be linked, knowledge which would not normally be associated together, it would become much more useful. Ted Nelson helped expand upon Berners-Lee’s invention by developing the concept of hypertext, a method of digitally linking from one section to another. The development of the internet during the 1960’s became user friendly during the 1990’s as it became increasingly available to the public. Berners-Lee was able to take something which was too complicated for most people to use, and create a system which made it user friendly. Incompatibility between computers had been a thorn in the side of technology for years as specialized cables were needed to ensure computers could communicate with one another. Berners-Lee had the brilliant idea to create a centralized block which all cables would feed into so that one central unit could be used for every computer in the world to communicate. Berners-Lee furthered this idea by designing the concept of anything being linked to anything. A single global information space would be birthed as a direct result of this, a system with common rules, which would be accessible to everyone, that effectively provided as close as possible to no rules at all; a decentralized system. This arrangement would allow a new person to use the internet without having to ask anyone else. Anyone, anywhere, could now build a server and put anything upon it. Berners-Lee decided to name his creation the “World Wide Web” because he thought of it as a global network. Berners-Lee took his intellectual property and provided it to the public free of charge, despite having many commercial offers. Berners-Lee felt that the idea would not become the largest and greatest invention of humanity had it not been free, democratized, and decentralized. The fact that anybody could access the internet and anybody could put content onto it, made the internet massively popular early on and grew at a rate of 10x year upon year. Berners-Lee also created the World Wide Web Consortion, an institution which was designed to help the World Wide Web to develop and grow
Recently, many of the improvements made within the artificial intelligence sector have been due to the technology of “deep learning” which is also referred to as an “artificial neural network”. Traditional software is not intuitive as it simply follows a set of instructions predetermined by a programmer. If the software runs into a new problem which it has no answer prewritten for, it crashes. Deep learning is different as software can now write its own instructions instead of reading the instruction(s) of a programmer. Currently, as of 2021, deep learning is the equivalent of an all powerful, dim witted genie as it has the ability to evaluate the pixels of a photograph of a bottle of water, and can recognize with astonishing accuracy photographs of other water bottles, however it has no idea what the concept of water or the water bottle itself is, what the end user does to drink from the water bottle, what the end user needs the water for etc. This differs in human beings however as humans learn from a sample size of one, and are able to surmise the purpose of water and everything else which is relevant from witnessing it being used upon a single occasion