ChatGPTTranslation Site

What is ChatGPT

ChatGPT is a popular chat robot launched by OpenAI. This AI chat tool is built on the GPT-4 language model developed by OpenAI. It can perform various natural language processing (NLP) tasks such as summarizing, classifying, asking and answering, and error correction similar to human reactions. ChatGPT is a revolutionary technology that makes people’s lives easier by increasing their work efficiency to a new level.

Who developed ChatGPT?

ChatGPT AI chatbot is developed and owned by OpenAI, which Elon Musk and Sam Altman founded as a non-profit company in 2015. Musk withdrew from OpenAI in 2018 and no longer owns a stake in OpenAI. Initially, the company received $1 billion from Silicon Valley venture capitalists to start building neural networks.

In 2019, OpenAI raised a second round of funding of $1 billion from Microsoft. They began using Azure supercomputers to build these large language models. Fast forward to 2023, and Microsoft invested $10 billion in OpenAI, bringing its total stake to 49%. Other investors, including Khosla Ventures, accounted for another 49%, and OpenAI only retained a 2% stake.

How ChatGPT works

Before ChatGPT, AI chatbots had already appeared, but because they were not conversational, they did not attract widespread attention. In 2017, Google introduced a neural network architecture called The Transformer in their paper “Attention is All You Need”, which created a paradigm shift in training large language models (LLMs).

Compared to other neural networks, both Recurrent Neural Networks (RNNs) and Long Short-Term Memory (LSTMs) are inferior to Transformer networks. RNNs have problems with long-term dependencies, while LSTMs cannot focus on the right words in a long sentence to get the right output.

The Transformer network changes the way language models are trained. Instead of processing one word at a time like RNN, the Transformer can inject the entire input at once. In addition, the Transformer allows multiple inputs to run in parallel, reducing computational costs and making training faster.

After realizing the potential of Transformer, OpenAI decided to use the Transformer network and train data according to its architecture. The training process of these models mainly includes three steps: generative pre-training, supervised fine-tuning, and reinforcement learning based on human feedback (RLHF).

How to use ChatGPT

ChatGPT has a basic version that is free to use. To use ChatGPT, you just need to visit their official website (https://chat.openai.com/chat). No download is required. Just go to the ChatGPT page and log in. You can choose to register with your email or log in with your Google or Microsoft account.

ChatGPT’s web interface is easy to use for all users, and the display interface includes a text box for users to enter queries and an area to display results. After entering your text prompt, you will then receive a response message from ChatGPT.