Skip to main content


I see more instances of people telling me, "I wasn't sure about this medical topic so I asked ChatGPT and it surprisingly knew the answer." My response, "It gave you an answer. Did you confirm the answer was correct from a reputable source?" This is worse than blindingly taking the first answer from Stackoverflow without checking if it is right or makes any sense. It's more like blindly taking the first web search result in the early web when some random quack or Nazi site could come up as a top result. #ai #chatgpt
in reply to Hank G ☑️

Nice metaphor: I have almost never found stack overflow to be useful. Cf. CommandLineFu.com.
in reply to Lee Rothstein

I've found it to be very useful at pointing me in the right direction. But you have to do your due dilligence at all times. I admit it seems to have had peak usefulness probably like 3-5 years ago.
in reply to Hank G ☑️

Joseph Weizenbaum: "I had not realized ... that extremely short exposures to a relatively simple computer program could induce powerful delusional thinking in quite normal people."
in reply to Adam Hunt

I think it is just with each computer iteration we have to relearn the meaning of the term, "Garbage In, Garbage Out" (GIGO)
in reply to Hank G ☑️

@Hank G ☑️: You're on a roll with metaphors today. I award you the Gold Medallion of the Metafive! ;-)
in reply to Hank G ☑️

I've noticed chatgpt is confident with the answer for generating computer code and it even compiles, but what it does is code salad
in reply to Hank G ☑️

I think it is just with each computer iteration we have to relearn the meaning of the term, “Garbage In, Garbage Out” (GIGO)


ChatGPT really illustrates that very well.

in reply to Hank G ☑️

Yeah, on a C64 forum, someone was asking a technical question about what color attributes Commodore 16 bitmap mode could store. The first person to reply asked ChatGPT and pasted the answer and what followed was confusion. ChatGPT confidently gave an answer that seemed to be talking about ZX Spectrum color attributes.

I did a quick search for a Commodore 16 technical manual and simply read it to provide the answer. (I already knew it, but wanted to link a reference.)

in reply to Hank G ☑️

even a correct answer is stolen from one or more other sites with no attribution or source material... Which means fact checking your chatgpt answer is slower than just google searching instead
in reply to Hank G ☑️

ChatGPT proves democratic nonsense is at least as nonsensical as nonsensical alleged-expert nonsense.
in reply to Hank G ☑️

we have been raised since the 80s to consider computers and technology errorless, especially when you talk toward consumers.

Even among professionals and companies it is hard to adopt a security mindset, because many fall in the same pitfall.

In that regard products like ChaptGPT are not exception; and if we insist to use misleading terminologies, such as Artificial Intelligence, we are going to further lower the bar of criticism and reinforce our naive trust in anything that is technology related. :flan_think: