This isn’t an AI/LLM-assisted post, however I’ve used the tag because today, who knows, but back in the mid 70’s the term hadn’t even been thought of and the recent posts on LLMs made me think about it again.
Back then, there was a lot of interest in AI - I briefly shared a house with someone at Edinburgh Uni. who was into things like Prolog, Poplog (I think) for AI and things of that ilk. Fascinating stuff although as an engineer it didn’t make much sense to me at that time…
At the same time (maybe early 80s) so-called “expert systems” were a thing and I recall one lecturer telling me - they’re no different from a huge if/then/else construct. Maybe…
There were also programs like Eliza, but also some fun. Looking through some old programs I adapted from Creative Computing, I found one called “Madlib Generator”. Is it an LLM or just making it up from human input? (Hint; It’s not an LLM!)
I won’t put the program here - search for “Mad Lib Generator” to find 100 variants, but here is part of the data-file that mine uses: You’ll get the idea, The program scans the data-file and asks for enough input to fill in the blanks.
# This is a comment
STORY: Shaggy dog tale
Once upon a time a adjective:1 dog got hold of a adjective:2 bone. He
was walking across a adjective:4 bridge and saw his adjective:3 reflection
in the water. exclamation:1 he said, "there is another adjective:1 dog
with another adjective:2 bone. I'll make a adjective:5 grab, take his and
then I'll have two adjective:2 bones". With that the adjective:1 dog
jumped into the water and adverb:1 drowned.
MORAL: A noun:1 in the noun:2 is worth two in the noun:3
ENDOFSTORY
Here is the output of a sample run… (Note: English wasn’t my best subject at school!)
