Princedede - Nostr Hypermedia

Politicians can't save you! No this and know peace.

Update on #Token2024. 👀

AI and Vocabulary! 👩‍💻 In the recent #Delvegate issue on X (Twitter, Inc.), Paul Graham posted, and I quote: "Someone sent me a cold email proposing a novel project. Then I noticed it used the word 'delve.'" Then he went ahead to comment: "My point here is not that I dislike 'delve,' though I do, but that it's a sign that the text was written by ChatGPT."

There were rebuttals from X users, especially from Nigerian, British, and Indian origins, who made it clear about the simplicity of the word 'delve' in everyday grammar, especially for people for whom English is their second or third language. Paul Graham still went on to talk about how using words like 'delve' isn't simple for everyday usage. But alas, Paul's submission was based on his knowledge within the American space. As we know, Americans are not known for a large vocabulary in English speaking, unlike their British counterparts.

This prompted me to write about some of the questions that followed throughout the argument on X, which I will be sharing below: In the long discussions on this particular issue, one thing stood out, which was the following question: Where was ChatGPT trained, and where was the data obtained from? Let's 'delve' deeper into that. A Thread 🧵 👇👇👇👇👇👇 Firstly, a machine learning model like #LLM can be trained anywhere, irrespective of the data's origin. Secondly, the data were gathered from various sources such as texts, books, online and social media platforms, and articles, making it a diverse pool. ChatGPT unlike Google's Gemini, is known for its creative flair and ability to generate text in different styles. It adapts its vocabulary and sentence structure to match the tone and style of the prompt it receives. This makes it a great choice for tasks requiring creative writing or generating new ideas. Now, where are the rich texts from? Stay with me... Out of about ~8 billion people worldwide, only about 1.5 billion speak English as their native or second language. And out of the ,~1.5 billion English speakers worldwide, the majority are second language speakers.

In addition to the second language speakers, other minor English speakers had to learn and apply English in creative ways, enriching their grammatical structures and vocabulary. Many learn English in the British way due to colonization, resulting in a richer flair to the language. It's only natural for most texts to come from writers or posters from these regions, as they are more than the native English speakers. ChatGPT's responses may seem superfluous with words to Americans but normal to British, Indians or Africans, who are accustomed to richer language. With AI, the response you get depends on the quality of your prompt. ChatGPT is flexible and can adjust its language style to suit any user or region. Even if the response is a bit more creative or vocabulary-rich, it can be toned down to suit any language style or region. So when you get a vocabulary-rich response, you can tone it down with a follow-up prompt like 'Make the above simpler' or 'Make the above more like American writing.' GPT is incredibly flexible and can adapt to suit any use case. You just have to be creative enough to use it!