DELTAZERO.TECH
In recent times, the success of OpenAI’s ChatGPT has been the talk of the town, emerging as the new leader in artificial intelligence. However, as the race heats up, tech giants are not one to be left behind.
Built on top of GPT-3.5, a refined version of GPT-3 published discreetly by OpenAI, ChatGPT is a force to be reckoned with.
Although OpenAI hasn’t released an official paper, the blog mentions the model uses “Reinforcement Learning from Human Feedback” (RLHF).
This model was trained on data published before 2021, which means it may make factual errors with regards to recent information. However, OpenAI continues to gather user data to improve and feed the AI. Users can also give their feedback on the AI’s responses by voting and leaving comments.
Google recently unveiled their new chatbot named Bard, which uses web-sourced information to answer user questions, giving it an advantage over ChatGPT.
Powered by Google’s Language Model for Dialogue Applications (LaMDA), Bard is currently in limited user testing. Despite a flawed initial public appearance, with criticisms of being rushed and inadequate, Alphabet CEO Sundar Pichai has stated that Bard will be available to the general public in the coming weeks.
China’s own version of Google, Baidu, is currently testing its chatbot called Ernie Bot.
Based on a language model named ERNIE 3.0 Titan, Ernie can understand and respond in both English and Chinese. With 260 billion parameters, it’s more advanced than similar language models such as ChatGPT-3, which has 175 billion parameters.
Initially, Ernie will be used as a conversational agent for users, but it will later be integrated into Baidu’s search engine.
Testing is expected to conclude in March 2023 before a public launch at a later date.
SHARE THIS POST
