First they came for …
Thanks to John Naughton’s Guardian column (https://www.theguardian.com/commentisfree/2023/jan/07/chatgpt-bot-excel-ai-chatbot-tech) I picked up this illuminating paper – a clear and confident perspective on how LLMs like GPT work (“next token predictors”!).
Download here: Murray Shanahan, 2022, ‘Talking about Large Language Models’, https://arxiv.org/abs/2212.03551
An interesting list of other current AI tools that aren’t GPT!
https://aisupremacy.substack.com/p/top-ai-powered-tools-not-named-chatgpt
Sniffing out GPT
An app that measures “burstiness” and “perplexity” to detect a GPT source.
However,
” … it does not work well with essays written by good writers. It false flagged so many essays as AI-written.”
Simple collection of links about the exploitation of near-earth Space
“…for educators, the technology opens the door to widespread cheating on homework and take-home assignments, and many have been scrambling to rethink the nature of assessment or otherwise discourage students using the tool.”
“… ChatGPT is capable of passing — or at least nearly passing — the US medical licensing exams, according to researchers in the US who put it to the test.”
The lives of Nicola Tesla (the real one) and ‘Tesla’ (the cosplay one) share a striking similarity. Tesla was very clever. His early inventions were decisive. He had an engineer’s fascination for how natural forces work but he had a prophet’s view of electrical energy. His later life was a decline into unfulfilled visions and penury, possibly one result of the loss of his life’s work in a catastrophic fire.
The Hebrew University has discovered an ivory comb from 1700 BCE that is inscribed with a plea to rid oneself of lice. The inscription reads:
“May this [ivory] tusk root out the lice of the hair and the beard.”
Isn’t cultural transmission amazing!
It passed me by until I saw this:
https://www.thesun.co.uk/news/20232755/chilling-ai-predicts-god-creepy-images/
With the obvious problem of pornographic imagery as well as issues such as taboo religious iconography, do image GPTs have contraints to limit the content of images they generate? (And if so how do we decide that?)
A new job is appearing (you know, one those jobs we don’t yet know exists).
What is this new job?
*drum roll*
A designer of prompts for GPT machines (and whatever might follow them …).
What are the essential aptitudes do you think? Are they general, specific to a particularlar machine or a bit of both?