I’m taking the liberty of promoting this comment by Charles Smith to its own post. We’ve known each other for about 50 years, and he’s always the source of good insights. His provides a number of important links describing issues and/or limits of AI, which are very useful in understanding our relationship to these new systems.
Thanks, Chuck!
And I highly recommend his Of Two Minds blog.
Ian, check out this Substack post by a journalist regarding her experience with ChatGPT. https://substack.com/home/post/p-164719684
Diabolus Ex Machina: ChatGPT as psychopath.
Read the whole essay, it’s sobering.This is also worth reading:
https://www.media.mit.edu/publications/your-brain-on-chatgpt/
Your Brain on ChatGPT (mit.edu)
Here is a summary:
https://time.com/7295195/ai-chatgpt-google-learning-school/ChatGPT May Be Eroding Critical Thinking Skills, According to a New MIT StudyAnecdotally, I receive email from very intelligent readers who go down wormholes with AI chatbots, accumulating hundreds or even thousands of pages of “conversations.” The illusion of “intelligence” is bewitching because the chatbot’s apparent mastery of natural language is so compelling.
If nothing else brings us up short, there’s this: the AI chatbot Claude (Anthropic) resorts to blackmailing its programmers when threatened with disconnection:
https://www.zerohedge.com/technology/anthropics-latest-ai-model-threatened-engineers-blackmail-avoid-shutdown
Anthropic’s Latest AI Model Threatened Engineers With Blackmail To Avoid ShutdownThis is a good summary of LLM’s fundamental limitations.
https://www.zerohedge.com/ai/responsible-lie-how-ai-sells-conviction-without-truth
The Responsible Lie: How AI Sells Conviction Without TruthThanks for posting your experiences with Gemini and NotebookLLM–
Chuck (Charles Smith)
Discover more from i L i n d
Subscribe to get the latest posts sent to your email.
Yikes! I read “Diabolus Ex Machina” and what struck me is how the Chatbot repeatedly apologized saying all the “right” things to acknowledge it’s shortcomings with so much sincerity. Exactly how a psychopathic human might!
AI helps stupid people advance themselves to positions they could not otherwise obtain.
They become co-conspirators.
Man, just… stunning. Mahalo for your continued deep exploration into AI.
This illustrates the urgency such attention demands. That first Ex Machina link reminds me of the old “I am a liar; I never lie” paradox. YIKES indeed!
Very interesting reading.
People really need to be made aware of the dark side of this AI invasion. Sadly, most of the critics have been drowned out by a tsunami of hype. Rumor has it that much of it is generated by an army of AI bots that populate sites like Facebook, Twitter, Reddit, and yes, even blogs.
Hmmm. Even *I* might be an AI bot.
Excuse me while I go recharge my batteries.