He’s running ChatGPT on a 1984 PC running MS-DOS, because why not


While ChatGPT is already shaping up to be a real revolution for the future, Yeo Kheng Meng wanted to bring artificial intelligence back to the past. In a relatively crazy project, the programmer managed to run the chatbot on a 1984 PC, running MS-DOS. It’s useless, but it’s impressive.

Credits: Yeo Kheng Meng

At this point, it’s almost impossible that you’ve never heard of ChatGPT. If the AI ​​does not know how to do everything yet – note that it is still capable of saving the life of a dog – difficult not to be impressed by its power and the precision of its answers. What create some fears for the future, if only for the many professions that will be impacted in the coming years.

While some countries are already acting against the proliferation of the chatbot, others are fascinated by it and are still looking for new ways to use it. As we know, the developer community likes to challenge themselves a bit far-fetched (no pun intended) and Yeo Kheng Meng is no exception to the rule. In a burst of motivation, he thus decided toinstall ChatGPT on a 1984 PC.

ChatGPT runs perfectly on MS-DOS, now we know

The lucky winner is an IBM PC running the mythical, but no less ancient MS-DOS. So it’s a real science fiction fantasy that Yeo Kheng Meng realized: use a computer from the 80s to interact with an artificial intelligence. The method is of course a little technical, but can be roughly summarized as follows: it was necessary to completely recreate the application so that it is compatible with the OS.

On the same subject — ChatGPT: if you are subscribed, your payment information is probably in the wild

To do this, Yeo Kheng Meng used an API from 1983 that allows the IBM PC to communicate with a network card. Then, the developer created a virtual machine, essential to run the application with components dating back 40 years. Finally, he used the official OpenAI API to implement ChatGPT on the PC, in the same way that the application can now be used on macOS.



Source link -101