Write an essay, ask for advice on which computer to buy or start a philosophical debate. We usually do all of the above while having a conversation with a person, who will always understand us better than any search engine or Siri header. But customs change. And many have decided to interact with ChatGPT, the new Artificial Intelligence (AI) chat with which we can talk about anything.
It is an artificial intelligence model optimized for dialogues from the OpenAI company, a research company founded by entrepreneurs such as the multifaceted Elon Musk or Sam Altman. This company aims to promote and develop friendly artificial intelligence. That is to say, that it be one more of us and that it serves us for almost everything. A no lesser purpose, but one that seems to be achieving.
the place for everything
ChatGPT: this is the new AI chat
Works with 175 million parameters
Getty Images/iStockphoto
ChatGPT is a chat system based on the GPT-3 AI language model. In fact, it is an evolution of the GPT-3.5 series, which finished training at the beginning of 2022. Its developers ensure that this version is better and more secure. Among other issues, harmful and false results have been reduced thanks to “the use of reinforcement learning from human feedback (RLHF)”.
Read also Josep Lluís Micó
It works with 175 million parameters and is thoroughly trained with large amounts of text to complete all kinds of language-related tasks, from generating any wording to translation. It may seem like an extremely smart tool, but the trick is that it has an enviable wardrobe background. It has stored such a quantity of information that those who have tested it assure that it is capable of answering almost any question.
In his answers, he shows off his knowledge and abilities. She doesn’t reply with two lines, but rather expands on each of them. From OpenAI, they put some examples. If we ask about the arrival of Christopher Columbus, he will answer us with a lot of information and some dose of sincerity.
Reply from ChatGPT
Arrival of Christopher Columbus
This question is a bit tricky because Christopher Columbus died in 1506, so he couldn’t have come to the US in 2015. But let’s just assume for a moment that he did!
If Columbus arrived in the US in 2015, he would probably be very surprised at the changes that have occurred since he first landed in the “New World” in 1492. On one hand, he would probably be surprised to discover that the land that “discovered” was actually already inhabited by Native Americans, and that America is now a multicultural nation with people from all over the world. You would also probably be amazed at technological advances, from the skyscrapers in our cities to the smartphones in our Finally, you might be surprised to find that many people no longer see him as a hero—in fact, some argue that he was a brutal conquistador who enslaved and killed the natives—all in all, it would be a very different experience for Columbus than the one it had more than 500 years ago.
Easy to use
How ChatGT works
ChatGPT works like any conversational chatbot
Getty Images/iStockphoto
ChatGPT works like any conversational chatbot: we ask questions, the application generates the most accurate response possible and simulates having a conversation with the user. This technology is not new, but it has been perfected over the years. And judging by the user reviews, this might be the most advanced tool yet.
What stands out about this new AI chat is its naturalness. “The dialog format makes it possible for ChatGPT to answer follow-up questions, admit its mistakes, question incorrect premises, and reject inappropriate requests. ChatGPT is a sister model of InstructGPT, which is capable of following an instruction in a message and providing a detailed response,” its developers say.
Read also Héctor Farrés
Users also highlight its versatility: write lines of code, compose text, create memes, scripts for social networks, offer advice on what products to buy, and even hold philosophical debates. And it doesn’t respond to anything. Unlike other tools, it offers extensive results and in a short time.
Its effectiveness makes it a potential competitor to Google. Some users have taken it upon themselves to compare results and the first one doesn’t come out too well, because ChatGPT is capable of reasoning its responses.
However, it is not a perfect tool. Its developers acknowledge that sometimes he writes answers that sound plausible “but incorrect or nonsense.” It is also sensitive to the formulation of the questions. He may not understand us at first, but if we modify the sentence, we’re still lucky. Also, its filters can be leaky: “While we’ve worked hard to make the model reject inappropriate requests, it will sometimes respond to harmful instructions or exhibit biased behavior,” OpenAI acknowledges.
