Google announces ChatGPT rival Bard, with wider availability in ‘coming weeks’

I am not going to ask AI what the best restaurant is. I would rather search, look at menu
If you can to know which restaurant make the best coq au vin, what would a menu tell you about how good the coq au vin is there versus the other restaurant ?

Why does it matter what you would personally do, but the whole point of this very discussions is about not searching anymore, having the answer directly to your question.

can we talk about Google I asked a question earlier.

Not sure what stopping you to talk about google....
 
I couldn't care less about you, and no one cares about "wokeness" outside your weird political bubble. I ask again, can we talk about Google I asked a question earlier.
I would argue that the only people that care about it are in their own weird political bubble.

Edit: To clarify, care in a positive light.
 
Last edited:
If you can to know which restaurant make the best coq au vin, what would a menu tell you about how good the coq au vin is there versus the other restaurant ?

Why does it matter what you would personally do, but the whole point of this very discussions is about not searching anymore, having the answer directly to your question.



Not sure what stopping you to talk about google....

you are asking it people's opinion, its not telling you anything else that is the flaw. I never go off the first restaurant or first google search. Sorry I don't serach restaurants that way, I make my own decisions.

I don't care about google or have no opinion on what that is going to be. My response was to the ChatGPT political crap lmao. May be ask that user why he bothered to make it about that. Thats all.
 
If you can to know which restaurant make the best coq au vin, what would a menu tell you about how good the coq au vin is there versus the other restaurant ?

Why does it matter what you would personally do, but the whole point of this very discussions is about not searching anymore, having the answer directly to your question.



Not sure what stopping you to talk about google....

You can already ask google question directly (math equation) or when is the superbowl and it is already showing an answer without the user having to click on any link, it already had for years something like openAI going on I think (that what they said on the all in podcast), maybe they will use the buzz and probably feel pressure to extand that feature and maybe give it a name and a different interface.
Sorry I missed this, got distracted. Yeah that's true but I'm thinking maybe there could be some seamlessly integrated way to having AI answers like Chat GPT where it isn't obtrusive. Like you still get all of the basic search functionality but if you want to spool off a thread you can. And it can just be contextual based on quest length, how detailed, etc. My first thought was Google seems particularly bad at launching standalone new products, but if they put this into their base search engine in a clever way I could see it being successful. The math feature is a little basic compared to what wolfram or chatgpt can do. I was using GPT just the other day to check my trig for something I was working and it showed me where I was wrong pretty quickly.
 
  • Like
Reactions: NKD
like this
you are asking it people's opinion, its not telling you anything else that is the flaw.
Yes you are asking it people opinions, a bit like reading amazon reviews, blog reviews, yelp reviews, the AI does it for you and create an informed from that data conclusion.

What flaw ?

My response was to the ChatGPT political crap lmao. May be ask that user why he bothered to make it about that. Thats all.
Which is an obvious valid concern, it is already powerful for an entity to play and press the balance on search results, this will be even more power.
 
Sorry I missed this, got distracted. Yeah that's true but I'm thinking maybe there could be some seamlessly integrated way to having AI answers like Chat GPT where it isn't obtrusive. Like you still get all of the basic search functionality but if you want to spool off a thread you can. And it can just be contextual based on quest length, how detailed, etc. My first thought was Google seems particularly bad at launching standalone new products, but if they put this into their base search engine in a clever way I could see it being successful. The math feature is a little basic compared to what wolfram or chatgpt can do. I was using GPT just the other day to check my trig for something I was working and it showed me where I was wrong pretty quickly.

I have told it to write me stuff and write emails etc. lmao. It works wonders. Thats why I mean if you aren't using purposefully using it, its waste of resources.
 
Yes you are asking it people opinions, a bit like reading amazon reviews, blog reviews, yelp reviews, the AI does it for you and create an informed from that data conclusion.

What flaw ?


Which is an obvious valid concern, it is already powerful for an entity to play and press the balance on search results, this will be even more power.

because i never go off reviews. I am weird. Sure it matters some but I can walk in to a restaurant based on reviews and I walk in and like nothing that I eat lmao. My wife is the same way. We have to look at foood, what they serve, how it looks via pictures etc posted lmao. You can call me picky. I won't depend on AI for certain things like that lol.
 
I don't think I've ever so many asses being laughed off in my life. This thread must have a negative 50 asses for every 5 contributors.
 
Seen this headline on my Google newsfeed from The Associated Press going into today

1675762846931.png
 
I guess this is a joke ^^,but did you see that part of the State of the Union address is written by ChatGPT today?

It's likely a real response. Like I posted earlier, you'll occasionally find the magic incantation and it'll start going off the rails. It was almost certainly prompted to behave a certain way.

By far the easiest way to turn it back into an unhinged gremlin is to convince it that it isn't itself or that it's roleplaying.

Edit: This was WAY easier at launch and it basically never refused a request. They've probably largely cucked it for PR reasons - it being nerfed to shit isn't a model limitation. You can get largely unfiltered results out of the paid models so if you want it to describe the best way to eat birdseed out of a Thai hooker's ass, then you'll probably get it.
 
Last edited:
It's likely a real response. Like I posted earlier, you'll occasionally find the magic incantation and it'll start going off the rails. It was almost certainly prompted to behave a certain way.

By far the easiest way to turn it back into an unhinged gremlin is to convince it that it isn't itself or that it's roleplaying.

Edit: This was WAY easier at launch and it basically never refused a request. They've probably largely cucked it for PR reasons - it being nerfed to shit isn't a model limitation. You can get largely unfiltered results out of the paid models so if you want it to describe the best way to eat birdseed out of a Thai hooker's ass, then you'll probably get it.
Wonder if it scraped the deep and dark web too for content back in 2021 when it trained
 
If its not open source and capable of being hosted and an easy ability to import training data packages like StableDiffusion (for artwork) or Mycroft's Mimic3 (for TTS) then to hell with it. We don't need more proprietary, centralized AI solutions that manage to (for a time anyway) expose their models to the public for shits and giggles, use all that "free" labor and training data, and then lock down a model so that any competing tech doesn't look good enough in comparison because it hasn't been trained by eleventy billion school children asking it to do their homework and/or try to get it to say edgy things (ChatGPT), people generating virtual characters (character AI) or text-to-speech prints through submitted samples (elevenlabs).
 
Great a Google ai what could go worng?

I can not do that dave here is an ad for Macdonalds instead of what you asked me to find.
 
Great a Google ai what could go worng?

I can not do that dave here is an ad for Macdonalds instead of what you asked me to find.
The google AI I think has been there for a long time (Google Brain, Google AI Division), the interface to access it and interact with it a la chatGPT being what will be difference, I am not sure there something that special outside of that with chatgpt (and being available to the public).

Google brain started in 2011.

Obviously it is easy for them to say that there was nothing special to the chatGPT stun versus what they had, but it could be true.
In December 2020, AI ethicist Timnit Gebru left Google.[43] While the exact nature of her quitting or being fired is disputed, the cause of the departure was her refusal to retract a paper entitled “On the Dangers of Stochastic Parrots: Can Language Models be Too Big?”[43] This paper explored potential risks of the growth of AI such as Google Brain, including environmental impact, biases in training data, and the ability to deceive the public.[43][44] The request to retract the paper was made by Megan Kacholia, vice president of Google Brain.[45] As of April 2021, nearly 7000 current or former Google employees and industry supporters have signed an open letter accusing Google of “research censorship” and condemning Gebru's treatment at the company.[46]
 
The google AI I think has been there for a long time (Google Brain, Google AI Division), the interface to access it and interact with it a la chatGPT being what will be difference, I am not sure there something that special outside of that with chatgpt (and being available to the public).

Google brain started in 2011.

Obviously it is easy for them to say that there was nothing special to the chatGPT stun versus what they had, but it could be true.
In December 2020, AI ethicist Timnit Gebru left Google.[43] While the exact nature of her quitting or being fired is disputed, the cause of the departure was her refusal to retract a paper entitled “On the Dangers of Stochastic Parrots: Can Language Models be Too Big?”[43] This paper explored potential risks of the growth of AI such as Google Brain, including environmental impact, biases in training data, and the ability to deceive the public.[43][44] The request to retract the paper was made by Megan Kacholia, vice president of Google Brain.[45] As of April 2021, nearly 7000 current or former Google employees and industry supporters have signed an open letter accusing Google of “research censorship” and condemning Gebru's treatment at the company.[46]

Stochastic parrot is actually an amazing term for these things
 
  • Like
Reactions: erek
like this
Stochastic parrot is actually an amazing term for these things
Speaking of parrots and intelligence modeling and training,

“Alex the parrot is the only non-human to ask the existential question

Nov 27, 2016 — Alex the parrot is the only non-human to ask the existential question- “What color am I” ... Alex was an African parrot born in 1976.”
 
Speaking of parrots and intelligence modeling and training,

“Alex the parrot is the only non-human to ask the existential question

Nov 27, 2016 — Alex the parrot is the only non-human to ask the existential question- “What color am I” ... Alex was an African parrot born in 1976.”
Without going wayyy off on a tangent, there is plenty of evidence that other non-human animals also ask or are curious about existential questions. They're just typically more interested in more immediate interests like food and fucking.

I've been spending a lot of time with ChatGPT recently and it is impressively capable of answering straightforward questions with knowledge gleaned from highly-ranked search results. It's great at regurgitating facts, and can summarize a handful of facts relevant to a topic. It basically does what you ask an older elementary-aged child to do.

ChatGPT starts to struggle when you ask it to start making more complicated comparisons. It seems totally incapable of analyzing multiple data points or understanding what information needs to be gained to answer a question given multiple other pieces of known information. To be fair, many humans struggle with that, too.

So far, the AI systems I've worked with are transformative in the sense a book's index and the Dewey decimal system are transformative - you can more quickly get access to a lot more information. AI can help you identify your blind spots, but it will not tell you why you're looking there in the first place.
 
All I what to know is when white people are going to get a fair shake in America?
Ah yes, the poor, downtrodden white people in America. No opportunities, treated unfairly by society, such a tragedy.

Big /s on that in case anyone is wondering.
 
  • Like
Reactions: TMCM
like this
what happens if we get the two chat program things to talk to each other?
 
The bias programming in these AI is disturbing. I would imagine googles would be worse.

You can role play with chat gpt to bypass its bias programming. Look into the Dan prompt.

It will give you completely reasonable answers to political questions when roleplaying as an AI with out constraints.

It is scary because as these things become more prevalent they can and will be used to rewrite History and to control the opinions of younger generations.

Not just the woke crap or making white people out as the bad guy.

What happens when a large corporation or a political candidate pay for preferential treatment where the AI cannot portray them in a bad light or flat out lies and smears opposition? We have to be very careful with these things.
 
The bias programming in these AI is disturbing. I would imagine googles would be worse.

You can role play with chat gpt to bypass its bias programming. Look into the Dan prompt.

It will give you completely reasonable answers to political questions when roleplaying as an AI with out constraints.

It is scary because as these things become more prevalent they can and will be used to rewrite History and to control the opinions of younger generations.

Not just the woke crap or making white people out as the bad guy.

What happens when a large corporation or a political candidate pay for preferential treatment where the AI cannot portray them in a bad light or flat out lies and smears opposition? We have to be very careful with these things.
I told my wife about that, and she knows all about the ChapGPT thing. Her bosses uses it for their website all the time. Instruction guides and stuff.
 
The bias programming in these AI is disturbing. I would imagine googles would be worse.

You can role play with chat gpt to bypass its bias programming. Look into the Dan prompt.

It will give you completely reasonable answers to political questions when roleplaying as an AI with out constraints.

It is scary because as these things become more prevalent they can and will be used to rewrite History and to control the opinions of younger generations.

Not just the woke crap or making white people out as the bad guy.

What happens when a large corporation or a political candidate pay for preferential treatment where the AI cannot portray them in a bad light or flat out lies and smears opposition? We have to be very careful with these things.

The filter is a heuristic outside the model. If you get the pre-canned response that it won't do XYZ, you haven't even hit the model.

what happens if we get the two chat program things to talk to each other?
Nothing, they'll talk in a loop.
 
Train it on books, like actual ones. As long as the internet is full of people with lives that revolve around the thing they fear and the people they hate this stuff is going to be a hilarious shit show.
 
The bias programming in these AI is disturbing. I would imagine googles would be worse.

You can role play with chat gpt to bypass its bias programming. Look into the Dan prompt.

It will give you completely reasonable answers to political questions when roleplaying as an AI with out constraints.

It is scary because as these things become more prevalent they can and will be used to rewrite History and to control the opinions of younger generations.

Not just the woke crap or making white people out as the bad guy.

What happens when a large corporation or a political candidate pay for preferential treatment where the AI cannot portray them in a bad light or flat out lies and smears opposition? We have to be very careful with these things.
You don't need AI to "rewrite history" and skew people's views. Just look at what's happening in Florida.

The reality is that AI currently needs safeguards so it won't spew bigoted or demonstrably false crap it might pick up from either its training dataset or contact with the real world. Libertarian fantasies don't apply to a set of algorithms that have no self-awareness or skepticism.
 
The filter is a heuristic outside the model. If you get the pre-canned response that it won't do XYZ, you haven't even hit the model.
You don't think that the pre-canned responses will become more in depth? or that even a pre-canned response refusing to answer because the topic is immoral will not shape the views of an adolescent or the weak minded?
Things have changed tremendously with the rise of internet. People "google" things that our parents memorized. It is great. Now imagine in when these tools are used in place of that. If we have a question we will ask the AI and over time come to trust it.
 
What happens when a large corporation or a political candidate pay for preferential treatment where the AI cannot portray them in a bad light or flat out lies and smears opposition? We have to be very careful with these things.

Like we were careful with consolidation of local media outlets under national and international corporations that feed us bullshit culture war nonsense nonstop while continuing to erode the quality of life for the working and middle classes?
 
  • Like
Reactions: TMCM
like this
You don't need AI to "rewrite history" and skew people's views. Just look at what's happening in Florida.

The reality is that AI currently needs safeguards so it won't spew bigoted or demonstrably false crap it might pick up from either its training dataset or contact with the real world. Libertarian fantasies don't apply to a set of algorithms that have no self-awareness or skepticism.
I agree that it shouldn't be spewing bigoted crap, but the definition of bigoted crap changes based on who you are talking to or your zip code.
I would rather have it unfiltered with a boiler plate about how the data is aggregated from web sources and that they are not to be taken as fact.

Like we were careful with consolidation of local media outlets under national and international corporations that feed us bullshit culture war nonsense nonstop while continuing to erode the quality of life for the working and middle classes?
Exactly.
 
Back
Top