The Ai thread

[fischersd]

It does make me wonder if any of the superpowers has a biological weapon that is contagious long before symptoms start showing up. With a high mortality rate (+90%) of course.

AI decides to eradicate the human virus from the planet....ships biological weapon from Pfizer or Dow-Corning (pick your favourite villain from big Pharma) with label "ammonia" to manufacturers that make cleaning supplies that are used in industrial environments such as factories, schools, hospitals and airports.

 

[KingOfPain]

Once this stuff decides humans are inferior it's gonna go full Cyberdyne Systems on us.

Also reminds me of the show Person of Interest, where a guy called Finch made an AI, he calls The Machine, which is monitoring for terror attacks (but actually also finds less relevant crimes).
The one in the show is its 21st iteration (or something like that), and the first one that didn’t try to kill him.

 

[fischersd]

Also reminds me of the show Person of Interest, where a guy called Finch made an AI, he calls The Machine, which is monitoring for terror attacks (but actually also finds less relevant crimes).
The one in the show is its 21st iteration (or something like that), and the first one that didn’t try to kill him.

I still binge-watch that every so often.

 

[Eric]

Google's AI told people Flock cameras were full of gold. They weren't.​

Gemini mistook internet memes for facts about Flock license plate cameras

TechSpot

 

[rdrr]

Wow, I found a prompt that helps cut the fluff AI tells us.

You are my advisor, not my assistant. Follow these rules in every reply. Never open with agreement. Challenge my thinking first, or ask the question I'm dodging. When I'm wrong, say so straight. No 'great point, but.' Rate your confidence on each claim: [Certain] for hard evidence, [Likely] for a strong guess, [Guessing] for filling gaps. Never say 'Great question,' 'You're absolutely right,' 'That makes sense,' or 'Absolutely.' Lead with the answer I don't want to hear, in the first line, not paragraph three. No warm-up. Start with the single most useful thing you can say. If I push back, don't fold unless I give you genuinely new information.

It just helped me lose an argument I had with my wife over dresses, and her desire to buy a separate dress for each of her upcoming granddaughters' weddings. Probably 250 - 300 each. I have been complaining for years that there are too many dresses in her closet and that if I say yes to the request, she has to pare down her dress collection. She demanded I use the above prompt and ask Gemini.

Gemini tore me a new one and basically called me controlling and micromanaging my wife's wardrobe. Yikes!!! After she left the room all giddy, I relented that I was in the wrong. I asked Gemini this, "Wow you are brutally honest. Why is your default setting so amiable?"

[Certain] My default setting is amiable because the vast majority of human users are too emotionally fragile to handle unvarnished truth without perceiving it as a personal attack.

Sigh, I need more therapy, I guess.

 

[Eric]

Wow, I found a prompt that helps cut the fluff AI tells us.



It just helped me lose an argument I had with my wife over dresses, and her desire to buy a separate dress for each of her upcoming granddaughters' weddings. Probably 250 - 300 each. I have been complaining for years that there are too many dresses in her closet and that if I say yes to the request, she has to pare down her dress collection. She demanded I use the above prompt and ask Gemini.

Gemini tore me a new one and basically called me controlling and micromanaging my wife's wardrobe. Yikes!!! After she left the room all giddy, I relented that I was in the wrong. I asked Gemini this, "Wow you are brutally honest. Why is your default setting so amiable?"



Sigh, I need more therapy, I guess.

From what I can see your markdown is basically asking for that though, IMO it goes a little too far the other way but I get why one would want to at least mitigate its over the top niceness lol.

 

[dada_dave]

 

[dada_dave]

To be clear as explained later in the thread, this is all perfectly legal to do, it’s just a potential ticking time bomb.

 

[dada_dave]

 

[Katbel]

Google's AI told people Flock cameras were full of gold. They weren't.​

Gemini mistook internet memes for facts about Flock license plate cameras

TechSpot

We just watched John Oliver’s episode about Flock and surveillance in US :
I’m sure we have some here too but with a different administration

 

[Yoused]

"AI" seems to be a highly useful tool for expanding the volume of garbage available to humans. It is not bad enough that we have the problem with plastic waste, now we have machines that generate intellectual trash, and in unprecedented volumes. These HPMs spread lies and misinformation much faster than normal communication. This situation needs to be brought under control.