A little programming whiz managed to use ChatGPT on a 1984 IBM computer running MS-DOS. The retro-computing enthusiast has gone all out to allow OpenAI’s AI to speak to him on a screen worthy of Fallout New Vegas.
ChatGPT can be very useful for all kinds of situations and you might need AI anytime. Recently, it was a student who was getting a fine thanks to the OpenAI chatbot. Only sometimes, the only computer available to you is a 1984 IBM5155 running MS-DOS. Sometimes. No problem, Yeo Kheng Meng, a little genius in retro-computing, took charge of adapting ChatGPT to it.
To read: ChatGPT: the uses of AI multiplied thanks to plugins, you will no longer need to leave the chatbot
In the past, he had already adapted Slack professional messaging to Windows 3.1. It was still necessary to increase the challenge and the mad scientist did not hesitate to use the oldest computer in his possession. This one presents a hardware configuration that would certainly not run Crysis:
- Processeur Intel 8088 4.77Mhz
- 640KB de RAM
- CGA ISA graphics card
- NE2000-ISA Compliant Ethernet Adapter
- XT-IDE ISA disk controller
- MS-DOS 6.22
The technology was still in its infancy at that time: MS-DOS does not have native networking capabilities, so it installed the Ethernet adapter mentioned above. Also, the IBM5155 has low processing power, “which presents additional difficulties when coding”says Yeo Kheng Meng on his blog.
ChatGPT runs like a charm on MS-DOS
He managed to make ChatGPT work on the computer by programming it with a compiler capable of targeting old operating systems like 16-bit DOS. He then programmed the API by hand to communicate with ChatGPT. To test his program, he used a virtual machine under Virtualbox. Only (small) element of cheat, a modern computer served as an intermediary for HTTPS encryption because the IBM computer is not capable of it, or else would require a titanic coding work. We will still salute the performance.
Read: ChatGPT-4: Elon Musk, Steve Wozniak and experts call for research to be paused, citing ‘serious risks to humanity’
In the video below, we can observe that the OpenAI chatbot is running like a charm. You will notice the slow typing speed of the programmer: he specifies that the keyboard is not the layout he is used to and that he was on the ground to type.
Source : Yeo Kheng Meng
