I have written almost 100 posts referencing Marshall McLuhan’s Laws of Media using the media tetrad. I recently came across a tweet where the author, Ethan Mollick asked ChatGPT, “Create a Marshall McLuhan tetrad for larger language models. Be creative and insightful. Think like a theorist.” The image below shows the ChatGPT response.
The ChatGPT response is similar to my tetrads created for these posts:
GPT-3 through a glass darkly
auto-tuning work
capitalism > automation > gpt
But of course the ChatGPT response would reflect my own. ChatGPT is merely a stochastic parrot of what humans have shared on the web. We are getting scraped by these tools, such as Google’s C4 dataset which in my case was confirmed by the Washington Post.
This use of LLMs has limited value in my opinion. I think a more useful application of LLMs would be in helping individuals interpret complicated pieces of information such as — end-user license agreements, contracts, patient records and diagnoses, user manuals.
The ChatGPT response is similar to my tetrads created for these posts:
GPT-3 through a glass darkly
auto-tuning work
capitalism > automation > gpt
But of course the ChatGPT response would reflect my own. ChatGPT is merely a stochastic parrot of what humans have shared on the web. We are getting scraped by these tools, such as Google’s C4 dataset which in my case was confirmed by the Washington Post.
This use of LLMs has limited value in my opinion. I think a more useful application of LLMs would be in helping individuals interpret complicated pieces of information such as — end-user license agreements, contracts, patient records and diagnoses, user manuals.