How does Large Language Model (LLM) “trail & error prompt guessing, not engineering” not also include:
1. specific model hyper-parameter configuration
and
2. specific model hyper-parameter tuning
for production of deterministic outputs?
#techConsequencesInRealWorld #aiethics
Large Language Model (LLM) “trail & error prompt guessing, not engineering” is not a substitute for curated knowledge informed by lived experience I.e. an expert - human or machine! #techConsequencesInRealWorld
Related
‘Weird interaction with a student. They keep coming up with weird “facts” (“Greek is a combination of four other languages”) that left me baffled. I said let’s look this stuff up together, and they said OK, I open a search bar, and they opened … Ch*tGPT. And I was like “this is not a search bar” and they were like “yes it is, you can search for anything in here”.
…
Each individual kid is now hooked into a Nonsense Machine.’
- https://mastodon.social/@Miniver/112742679202304903
Related
“Based on my exposure to all this AI stuff, I find those most enthusiastic are those who want to make fast gains in a field they know nothing about. It's not about it getting better, it's about that (perceived) immediate gain.”
Sounds like a search for a ‘silver bullet” shortcut 🤦🏾♀️
Related and sounds like, ‘How hard can it be?’ arrogant disrespect.
‘Turns out "precisely saying what you want" is _what programmers do_
Having a magic genie that tolerates imprecision and still Do What I Mean, not what I said (DWIM) is what every founding "non technical" founder thinks their "technical co-founder" does.
And the AI hype is "maybe I won't need to hire her next time"’
@dahukanna The only valid use of AI is to make it draw a picture of a cat riding a dog, or a fish marrying a watering can. Other than that it is of no practical value.
@JamesHMcLaren @dahukanna I spent half a day at work generating stuff like that on Canva. Prompt: batwing gorilla Styles: moody; playful
There is another useful AI use case. See https://pony.social/@cadey/112718350118452981
@SuperMoosie Ok, that is also acceptable.
@JamesHMcLaren @dahukanna
Disagree. It's very, very good at generating disinformation. If that's your use case.
(Which hinges on your definition of "valid", so I suspect we are in violent agreement here.