But First: Not AI/LLM Proven Not Ready for Prime Time (not even ready for 3 a.m. on Sunday) Let’s agree not to let it set bombing targets, OK?
Now we return you to our previously scheduled item: Convinced that I possessed a profound insight, I rushed to the keyboard to proclaim that Douglas Adams predicted AI slop a half century ago, when his fictional Deep Thought computer took 7.5 million years to proffer its ultimate answer to life, the universe and everything: 42.
Turns out someone else had that insight first, better than mine, and in greater detail, a year ago. Bryan Kennerley explained it all on his blog, Keeping Up With AI Everything: Douglas Adams – Deep Thought – Lessons for the AI Age.
LLMs (Large Language Models)[1] act as if they know “Ask a stupid question, get a stupid answer. Garbage in, garbage out.” Ask a question that requires the tiniest scintilla of understanding about the actual real-world meaning of the answer, and prepare to be disappointed, no matter how good the prompt. Since they are nothing more than industrial-scale cut-and-paste engines, LLMs don’t “know” anything, including when they are wrong.
Also, Ray Kurzweil: The Singularity Is Near
[1] Don’t let them fool you. It is not AI
Comments
[…] I should have known that the readers of this column would be fascinated by Douglas Adams, who weighed in on AI last week. […]