A Different Thread

[elhombresinnombre]

(Requested by @steveinbsas at the end of https://baexpats.org/threads/is-the...2-years-with-perm-residency.50182/post-472204 )

Steve, I love the AI searches you've done on yourself and on my BAExpats identity. Let me assure you that the answers on me are just as crazy wrong as the answers on you. but they all sound plausible to someone who doesn't know.

Back in the days when Google Search was new, fresh, innocent and exciting I started getting emails from people all over the world who I didn't know asking - are you the person in this document? The document existed in print and on line and it included three famous names and not-at-all-famous me. Not-at-all-famous me gave authenticity to the connection between the other three names. Other people would Google their own names: my party trick at the time was to Google those three names and watch my name appear in the first result.

These days you have to wade through three or four pages of search engine slop to find it but the document is still there. Why is that important? Because Google Search knows where to find those connections but it is apparent that Google AI does not.

I went to Google AI and instead of just searching - Google Search-style - on those three names, I asked a question "What is the connection between A,B,C and myself D?" After coming up with some stuff which was plausible and possible but probably false about A and B, definitely false in the given context about C, it went on to state that D -me- used to play drums with The Zombies. And all the time, I knew the definitive answers were sitting in a document on-line that Google Search could find in an instant.

I know anecdotal evidence is not evidence but my other experiences as well as that one above, lead me to believe that the user needs to have the right degree of expertese in whatever field in order for AI to be a truly useful assistant. It's an assistant, not an authority.

 

[admin]

This is not directed at elhombresinnombre specifically, but rather a general comment about a trend.

Google AI feeds on discussions on this forum. Not the other way around. I'd rather keep it that way.

If you had an eye-opening conversation with Google (or any other) AI, there is probably no reason to repost the transcript here.

 

[wineguy999]

(Requested by @steveinbsas at the end of https://baexpats.org/threads/is-the...2-years-with-perm-residency.50182/post-472204 )

Steve, I love the AI searches you've done on yourself and on my BAExpats identity. Let me assure you that the answers on me are just as crazy wrong as the answers on you. but they all sound plausible to someone who doesn't know.

Back in the days when Google Search was new, fresh, innocent and exciting I started getting emails from people all over the world who I didn't know asking - are you the person in this document? The document existed in print and on line and it included three famous names and not-at-all-famous me. Not-at-all-famous me gave authenticity to the connection between the other three names. Other people would Google their own names: my party trick at the time was to Google those three names and watch my name appear in the first result.

These days you have to wade through three or four pages of search engine slop to find it but the document is still there. Why is that important? Because Google Search knows where to find those connections but it is apparent that Google AI does not.

I went to Google AI and instead of just searching - Google Search-style - on those three names, I asked a question "What is the connection between A,B,C and myself D?" After coming up with some stuff which was plausible and possible but probably false about A and B, definitely false in the given context about C, it went on to state that D -me- used to play drums with The Zombies. And all the time, I knew the definitive answers were sitting in a document on-line that Google Search could find in an instant.

I know anecdotal evidence is not evidence but my other experiences as well as that one above, lead me to believe that the user needs to have the right degree of expertese in whatever field in order for AI to be a truly useful assistant. It's an assistant, not an authority.

More for the anecdote bucket: Literally every time I have done a Google AI search, the answers have been comically (or tragically) incorrect.

 

[LuckyLuke]

I was recently pondering this theme some days ago. AI stands a real risk of running genuine, however flawed, discourse off the web. It's okay to conduct personal research using AI and, adding to the wealth of experience and connections one might have, summarize a helpful comment here.

We are all human and I think in an age where intelligence is being put on a pedestal above all other human traits, we forget to realize that the communal connection is really what brings us here together.