ARTIFICIAL INTELLIGENCE | DIGITAL LIFE
What ChatGPT-4 Doesn’t Know About Itself
The paradox of self-awareness in AI
With how much ChatGPT has been in the news recently, it’s perhaps no secret that it has access to an immense amount of information. Artificial intelligence is a fast-changing world after all, and ChatGPT is no exception. Since the first model, its abilities have grown exponentially. But one thing it doesn’t know is… itself.
GPT-1 was released back in 2018, and though it hadn’t yet begun to offer the chat functionality that would later define the program, early GPT was nothing to scoff at.
Trained on 117 million “parameters,” or data points, it was already making history behind closed doors with the large bodies of text it was able to interpret. Its successor, though, was trained on 1.2 billion of these data points. GPT-3 achieved 175 billion and, by some estimates, ChatGPT-4 draws from over 100 trillion data points in the replies it now offers users.
Its capabilities are staggering. There are very few subjects on which it can’t meaningfully converse. From pop culture, to coding, to chemistry, it’s hard to overstate just how much knowledge is built into this chatbot.
