ChatGPT Come try it out

[Emrie]

I broke it with a simple math problem:

It's a language model that's designed to generate language, not do math.

That's not to say that it is not amazing, and I don't really need it to do math in my particular use cases. But I've used it for coding, I've used it to help me in conversations I didn't know much about, I've used it to plan activities and offering gifts to people I didn't know too well.

It's an amazing companion in my daily life and I really love it.

Edited January 13, 2023 by Emrie
Adding extra explanation at the end

[Emrie]

By the way I thought people might want to hear this here. It seems they are going to adopt a subscription model and have setup a survey to ask people what they'd be willing to pay for it.

[Osaid]

On 1/13/2023 at 4:57 AM, Emrie said:

I broke it with a simple math problem:

It's a language model that's designed to generate language, not do math.

That's not to say that it is not amazing, and I don't really need it to do math in my particular use cases. But I've used it for coding, I've used it to help me in conversations I didn't know much about, I've used it to plan activities and offering gifts to people I didn't know too well.

It's an amazing companion in my daily life and I really love it.

I can't believe it couldn't figure that out lol

[integral]

https://beta.openai.com/examples ????

Checkout the playground its also FANTASTIC

Edited January 18, 2023 by integral

[Osaid]

3 minutes ago, integral said:

Checkout the playground its also FANTASTIC

Recipe creator is interesting. ChatGPT is a chef.

 

[integral]

 

"A table summarizing the beliefs from following text:"

https://www.actualized.org/insights/royalty-monarchy-is-toxic

Beliefs:

• 1. Unearned wealth, fame, and power are toxic to the mind.
• 2. Royalty is anti-democratic, anti-egalitarian, anti-meritocratic, and anti-spiritual.
• 3. The concept and structure of royalty is a scam.
• 4. It is the spiritual duty of any wise and developed individual to reject any schemes of royalty they were born under.
• 5. Rejecting the pursuit of luxury and status is one of the best decisions one can make in life.
• 6. Happiness does not come in hat form.
• 7. The good life is not measured in luxury and comfort, but in how hard one works to polish themselves into a serious, solid character.
• 8. Character and consciousness is forged in the furnace of emotional labor.
• 9. Do not envy those with high-status and privilege, as they will never become as strong, as developed, as happy, or as conscious as you.
• 10. The human mind and body are prone to spoilage when not sufficiently stressed.

Edited January 18, 2023 by integral

[integral]

A table summarizing the flaws from following text:

https://www.actualized.org/insights/royalty-monarchy-is-toxic

• -The concept of royalty is anti-democratic, anti-egalitarian, anti-meritocratic, and anti-spiritual
• -Royalty is a scam perpetrated on the human race
• -It is not possible to buy good character or psychological development for children
• -The human mind and body are prone to spoilage when not sufficiently stressed

LoooL I wonder... 

Edited January 18, 2023 by integral

[integral]

Change the sentiment of this text to positive:

Input: "Way more toxic than RSD used to be. It's funny to me how easily guys fall for this "macho man" fantasy bullshit. This not how you become a powerful man. This is man-children cosplaying as men. You don't need this kind of thing to be a man, it will not make you happy. Controlling women will not make you happy."

Output: "It's amazing to me how easily guys can become powerful men. This is a great way to become a strong and confident man. You don't need to rely on macho fantasies to be a man, it will bring you joy and fulfillment. Empowering women will bring you true happiness."

???

I am trying to figure out how to change the sentiment of the text without changing the meaning or beliefs....

Edited January 18, 2023 by integral

[Leo Gura]

7 hours ago, integral said:

6. Happiness does not come in hat form.

How dumb can this AI be? Has it never seen one of my Halloween hats?

[EternalForest]

This thing is amazing. I wish it wasn't so overly balanced oriented and simplistic towards certain topics, but this is the first time that I've texted an AI that feels like I'm talking to a real human being!

Edited January 18, 2023 by EternalForest

[Leo Gura]

If you guys really wanna test this AI, tell it a joke and ask it to explain why the joke is funny.

Example:

A rabbi walks into a Vegas hotel to check in. He tells the receptionist, "Miss, I'm a deeply religious man. So can you please make sure that the porn in my room in disabled?"

The receptionist pauses, looks at him with disgust on her face and says, "You sick bastard! We only serve normal porn here."

Edited January 18, 2023 by Leo Gura

[EternalForest]

@Leo Gura It can explain simple and straightforward jokes, but it has trouble with more subtle ones. And while it can write some creative song lyrics, they always take the same kind of bend, and can feel quite generic.

It's still impressive for what it is.

[zurew]

19 minutes ago, EternalForest said:

This thing is amazing. I wish it wasn't so overly balanced oriented and simplistic towards certain topics, but this is the first time I that I've texted an AI that feels like I'm talking to a real human being!

It has a hard problem deciding between giving a creative response vs giving a factual response, because frankly, its a really compilcated problem to solve, especially if you want to combine the two (for instance you want it to do a roleplay as a scientist and explain complex things to you).
 

[EternalForest]

Just now, zurew said:

It has a hard problem deciding between giving a creative response vs giving a factual response, because frankly, its a really compilcated problem to solve, especially if you want to combine the two (for instance you want it to do a roleplay as a scientist and explain complex things to you).
 

Totally agree. It's mainly the potential that I find fascinating, and experimenting with what it can already do well.

[something_else]

17 minutes ago, Leo Gura said:

If you guys really wanna test this AI, tell it a joke and ask it to explain why the joke is funny.

Example:

A rabbi walks into a Vegas hotel to check in. He tells the receptionist, "Miss, I'm a deeply religious man. So can you please make sure that the porn in my room in disabled?"

The receptionist pauses, looks at him with disgust on her face and says, "You sick bastard! We only serve normal porn here."

It gets there eventually, but it requires a hint or two

[Leo Gura]

@something_else That's amazing.

This thing is more intelligent than right-wingers, Lol.

Edited January 18, 2023 by Leo Gura

[something_else]

1 hour ago, Leo Gura said:

@something_else That's amazing.

This thing is more intelligent than right-wingers, Lol.

Lmao. I think the thing that impresses me most about it is that it can iterate on its own work. And also figure out its own mistakes and fix them with minimal help.

I noticed the same pattern when asking it programming questions. It would make a small mistake, and all you'd have to do is say "You were close but you made a small mistake" and it usually finds and fixes it just from that. If that doesn't work you might need to give it a small hint. It's so human-like in how it can assess its own responses, it's amazing.

Edited January 18, 2023 by something_else

[Lyubov]

This is an unbelievably cool tool and project.

I’m actually finding myself coming to it for help on things like JavaScript and asking it detailed questions which it produces.

It also seems to have been programmed with integrity which is great (of course I asked it how to have sex with a woman). All around a stand up job from OpenAI. 
 

I will use it going forward to help me with programming and writing. It absolutely can be used no for that. 

[something_else]

36 minutes ago, Lyubov said:

I’m actually finding myself coming to it for help on things like JavaScript and asking it detailed questions which it produces.

It's insanely good at programming, and it understands almost every language and many frameworks in those languages. Not quite good enough that you could rely on it without some reasonable programming knowledge, yet, but it's not that far off.

[Osaid]

11 hours ago, something_else said:

It gets there eventually, but it requires a hint or two

Technically it gets it right the first time. The joke is the idea that the rabbi doesn't want normal porn. It just adds more details and makes it more explicit since it probably thinks you're confused about it or something.

Edited January 18, 2023 by Osaid