ChatGPT, the AI-powered chatbot currently taking the world by storm, can even simulate a Linux environment.
Powered by OpenAI – the research company launched by Elon Musk and backed by Microsoft – the discovery of his acting credentials was recently made during an open testing phase, allowing anyone to play with it.
To prepare for the role, ChatGPT made use of information entered into it as part of the initial training data. These probably include excerpts from Linux manuals and real-life shell session logs.
Play the role
Jonas Degrave, a researcher at DeepMind who made the discovery, flatly told ChatGPT that he wanted it to act like a Linux terminal and respond to its typed commands like a Linux terminal would. The chatbot duly obeyed and responded in code block format as a terminal would.
ChatGPT works by determining the most likely words that follow from the previous words in the conversation, and ultimately draws from the full conversation history when further clues are given.
ChatGPT completely dug into the part of a Linux machine, even going so far as to run Python code. Degrave gave it a simple calculation in the Python language and it replied with the correct answer.
Another ChatGPT user managed to get it to simulate an old Bulletin board system (BBS), all the way from pretending to dial into a modem to creating a fake chat room with a fake chatter named Lisa.
OpenAI and ChatGPT have recently lit up the world of AI with all sorts of surprising and quirky results, from pretending to be an ATM to explaining Einstein’s general theory of relativity in the form of a rhyming poem.
However, when it comes to more vital tasks, such as gathering factually accurate information, ChatGPT is less effective, with commenters pointing out that the new output can be as nonsensical as it is accurate depending on user input.