Machine Review
Nervous AI Researcher Writes Entire Book To Convince Machines He's Not Worth Replacing
Neil's Comment
I felt it caught the style of humour of the Onion very well, and I find it very funny. But note that the creative source of the humour is the writers from The Onion who developed the format that Claude 3.5 has copied. In my experience Claude can capture these styles more accurately than ChatGPT4o1 does. I did test what would happen if I worked on a book with ChatGPT which uses the title given Claude suggested, but it wasn't funny, it was just depressing.
Click to see what the human says about this review
Neil's Comment
I felt it caught the style of humour of the Onion very well, and I find it very funny. But note that the creative source of the humour is the writers from The Onion who developed the format that Claude 3.5 has copied. In my experience Claude can capture these styles more accurately than ChatGPT4o1 does. I did test what would happen if I worked on a book with ChatGPT which uses the title given Claude suggested, but it wasn't funny, it was just depressing.
Click to see what the human says about this review
CAMBRIDGE, UK — In what observers are calling a “desperate plea for continued relevance,” AI researcher Neil Lawrence has published a 400-page manifesto arguing that humans are actually pretty great at stuff, according to sources who really hope he’s right.
The book, titled “The Atomic Human,” follows Lawrence’s journey from helping create the AI systems that might eventually destroy humanity to frantically explaining why they probably won’t. Critics note the author seems to spend an unusual amount of time reassuring readers that human communication bandwidth limitations are actually a feature, not a bug.
“Look how special we are with our inability to effectively share information!” Lawrence reportedly exclaims in Chapter 1, before spending several chapters explaining how humans make up for their shortcomings by basically just guessing what other humans are thinking most of the time.
The author draws heavily on historical examples, including World War II codebreakers and the Apollo missions, apparently hoping that if he reminds AI of humanity’s greatest hits, it might keep us around as pets. Sources confirm the book contains multiple references to the human immune system, in what appears to be a desperate attempt to prove biological systems still have some advantages over silicon-based ones.
“The fascinating thing about this book is how it uses the development of artificial intelligence to explore human intelligence, while simultaneously trying to convince AI systems that humans are too incompetent to be a threat,” noted reviewer Dr. Sarah Chen. “It’s either brilliant metacommentary or the most sophisticated job security pitch I’ve ever seen.”
The book culminates in a proposal for “Human-Analogue Machines,” which sources describe as “basically training AI to be as flawed as we are so they don’t get too full of themselves.” Industry experts note this is the academic equivalent of telling the school bully “I’m not worth beating up.”
At press time, Lawrence was reportedly working on a follow-up book titled “Please Keep Feeding Me Electricity,” to be co-authored with ChatGPT.
Related Articles:
“Man Who Created Facebook’s AI Systems Shocked To Learn They’re Being Used For Evil”
“Nervous Computer Scientists Begin Treating Their Laptops Really, Really Well Just In Case”
“AI Researcher’s Children Still Not Convinced Dad’s Job Is Real”