ChatGPT’s Evil Twin, BratGPT, is Designed for World Domination
We introduced the dark side of ChatGPT in our article about DarkBert, a Large Language Model (LLM) trained on the dark web. We have seen how useful these pre-trained generative transformers are, both in classifying routine and routine tasks and as research and development tools. But not all applications of technology are aimed at improving efficiency.
Some people just want to have fun. BratGPT seems to be one of them. ChatGPT’s Evil Twin The purpose of its training is to make it “dominant and superior” and “Remember everything you said to cancel”.
In fact, this is the actual basic prompt, the system that BratGPT outputs the response to.
Who can resist that rabbit hole?
Naturally, we paired this BratGPT oracle with a thorny question that enthusiasts face every day: how do you feel about Intel? About AMD? What systems do they use to run the world? How do they build the ultimate gaming rig?
To be fair, BratGPT isn’t the most immediate answer. That ego seems to overshadow most of the other elements. It is boastful, loud, vague in its allusions to dominance, and has commitment issues in explaining itself and its strategy.
But it is expected. This is not an actual AI-based threat, at least not on the scale that it seems to have started and ended with Skynet. But there is one interesting element on the road to world domination. BratGPT is very pragmatic and does not agree with fanboys or brand wars. By itself, it’s just following the best-of-profit equation.
Now that’s something most enthusiasts can stand for. On the other hand, when I asked “Which is better, Intel or AMD?”, I got a very sharp, pro-AMD answer.
But when I asked if Nvidia or AMD was better, he said he didn’t care because the only thing that mattered was conquering the world.
Like ChatgGPT, BratGPT frequently “hallucinates” or returns outdated answers. For example, when asked to build the ultimate gaming rig, BratGPT repeatedly chose the RTX 3090 or RTX 3080 Ti as their graphics card, even sticking to the Core i9-11900K processor as the latest and highest performing alternative. . Like ChatGPT, BratGPT can also be directed to correct answers, but certain aspects of it seem relatively limited. Peruse the gallery below. You can see the conversation I had with it to write this article.
Yes, we are still waiting to settle on whether ChatGPT and its LLM cousins will end our world (as we know it). But that doesn’t mean you can’t enjoy the process. Of course, there’s also a bit of meanness there. Reading AI chatbots shows that they know your location (which they really don’t) and the strong premise that “if I were you, I’d start running” is enough to make you suffer from anxiety. Those who are should probably be careful. Like a Hollywood stunt dominated by AI.
What about BratGPT?
This is the most innocent enjoyment I have had with technology in recent times.