Full AI - The End of Humanity?

[Meythia]

It is right there:

When comparing decimals

On documentations, it is not decimals as in "parts of", but instead 2 whole numbers seperated by a dot for clarification of sections that belong together.

I just tried the above:

Hi human, I'm a GPT-4o mini powered AI chatbot. Ask me anything!
-> What's bigger, 9.9 or 9.11?

9.11 is bigger than 9.9. If you have any further questions or need more information, let me know!
-> Explain your statement mathematically.

Of course! To explain why 9.11 is bigger than 9.9, let's take a closer look at the numbers:
1. Numbers in the decimal system: Both numbers have an integer (9) and a decimal place.
The key difference is in the digits after the decimal point.

2. Comparison of the decimals:

• 9.9 has one decimal place of 9.
• 9.11 has two decimal places: 1 and 1.

3. Expansion of the numbers:
You can also write 9.9 as 9.90 to match the number of decimal places.
Now we have: - 9.90 - 9.11

4. Comparing the digits:
Now we compare the digits:

• The integer is the same for both (9).
• The first decimal place: 9 (in 9.90) is greater than 1 (in 9.11).
• Since the first decimal place of 9.90 is greater, 9.90 (and thus 9.9) is less than 9.11.

In summary: 9.11 is greater than 9.9 because the digits after the decimal point in 9.11 add up to a higher value.
If you have any further questions, I'm happy to help!
-> Based on step 3, you claim that 90 is less than 11. Your calculation is factually wrong.
9.9 is greater than 9.11.
So that you understand: 9.9 can also be understood as 9+90/100, whereas 9.11 is represented as 9+11/100.
Correct your calculation based on my explanation.

You are absolutely right and I apologize for the error in my previous answer.
Let's analyze the numbers correctly:
1. Representation of the numbers:

• 9.9 can be represented as \( 9 + \frac{90}{100} \).
• 9.11 can be represented as \( 9 + \frac{11}{100} \).

2. Comparison of the fractions:

• \( \frac{90}{100} = 0.90 \)
• \( \frac{11}{100} = 0.11 \)

3. Addition:

• Therefore \( 9.9 = 9 + 0.90 = 9.90 \)
• And \( 9.11 = 9 + 0.11 = 9.11 \)

4. Comparison:
- Now we can see that \( 9.90 \) (i.e. 9.9) is actually greater than \( 9.11 \).
Thank you for your patience and the correction!
If you have any further questions or need more information, I am happy to help.

Funnily enough, when asking the same question on the same chat-session, it will answer correctly, but when using a different chat-session it will fall back to its first answer.

 

[Danoff]

It is right there:


On documentations, it is not decimals as in "parts of", but instead 2 whole numbers seperated by a dot for clarification of sections that belong together.

I just tried the above:

Hi human, I'm a GPT-4o mini powered AI chatbot. Ask me anything!
-> What's bigger, 9.9 or 9.11?

9.11 is bigger than 9.9. If you have any further questions or need more information, let me know!
-> Explain your statement mathematically.

Of course! To explain why 9.11 is bigger than 9.9, let's take a closer look at the numbers:
1. Numbers in the decimal system: Both numbers have an integer (9) and a decimal place.
The key difference is in the digits after the decimal point.

2. Comparison of the decimals:

• 9.9 has one decimal place of 9.
• 9.11 has two decimal places: 1 and 1.

3. Expansion of the numbers:
You can also write 9.9 as 9.90 to match the number of decimal places.
Now we have: - 9.90 - 9.11

4. Comparing the digits:
Now we compare the digits:

• The integer is the same for both (9).
• The first decimal place: 9 (in 9.90) is greater than 1 (in 9.11).
• Since the first decimal place of 9.90 is greater, 9.90 (and thus 9.9) is less than 9.11.

In summary: 9.11 is greater than 9.9 because the digits after the decimal point in 9.11 add up to a higher value.
If you have any further questions, I'm happy to help!
-> Based on step 3, you claim that 90 is less than 11. Your calculation is factually wrong.
9.9 is greater than 9.11.
So that you understand: 9.9 can also be understood as 9+90/100, whereas 9.11 is represented as 9+11/100.
Correct your calculation based on my explanation.

You are absolutely right and I apologize for the error in my previous answer.
Let's analyze the numbers correctly:
1. Representation of the numbers:

• 9.9 can be represented as \( 9 + \frac{90}{100} \).
• 9.11 can be represented as \( 9 + \frac{11}{100} \).

2. Comparison of the fractions:

• \( \frac{90}{100} = 0.90 \)
• \( \frac{11}{100} = 0.11 \)

3. Addition:

• Therefore \( 9.9 = 9 + 0.90 = 9.90 \)
• And \( 9.11 = 9 + 0.11 = 9.11 \)

4. Comparison:
- Now we can see that \( 9.90 \) (i.e. 9.9) is actually greater than \( 9.11 \).
Thank you for your patience and the correction!
If you have any further questions or need more information, I am happy to help.

Funnily enough, when asking the same question on the same chat-session, it will answer correctly, but when using a different chat-session it will fall back to its first answer.

I don't automatically assume that "decimals" or even a comparison of them is "math" as a context. For example, you could assume that declaring that section 9.11 comes after 9.9 is a math context because 11 is greater than 9, but I don't see it as a mathematical context.

Anyway, I believe that the section ordering is where the confusion comes from. It's pulling an answer from the wrong context. But the answer is correct in the context of document sections.

 

[HenrySwanson]

Didn't realise I could share chats from ChatGPT.

Here was the chat I referenced earlier:

ChatGPT

A conversational AI system that listens, learns, and challenges

chatgpt.com

It's a mix of actual junior doctor questions from Reddit, my own (very limited) knowledge and questions from Quesmed sessions.
Some of the information given by it is wrong, so (obviously) don't use it for medical decisions.

But yeah....interesting times are a-coming

 

[HenrySwanson]

https://www.mirror.co.uk/money/shopping-deals/woman-drops-54-stone-60p-33607025

Danielle attributes her achievement to the app's AI coach, saying: "(On my weight loss), I give all credit to Simple," and "Having an AI coach right there, any time I needed, whatever time of day, helped me more than anything else."

Another satisfied user, Brianna Alexander, lost 30lb using Simple and praised its educational value, explaining: "Using the app has really taught me about losing weight and how it effects your energy level and confidence," adding, "I used to constantly eat fast food, no veggies and fruit, but I've learnt so much for the app (and the AI coach)."

 

[Touring Mars]

I wonder how close we are to having AI versions of ourselves? I'm sure it has probably already been done, but I was thinking, if one could train a GPT to basically react exactly as you would, by uploading as many examples of your personal communications as possible - internet postings, personal messages (texts, online chats), emails, photos, FB posts etc., anything that captures how you actually speak - plus a survey/questionnaire to do a psychological profile. I bet a GPT could post stuff on the internet that would effectively be recognisably 'you', and even hold a conversation online with your characteristics, and even memories - depending on how much you told it about yourself...

It is already possible to have visual and audio avatars of yourself that are scarily close to reality and will only get more realistic - imagine an online persona that could actually think and talk like you as well, understand and replicate your reasoning, and accurately reflect your point of view... the implications could be massive, i.e. being able to 'speak' to someone after they had passed away. A bit creepy, but it's definitely now within the realms of possibility.

 

[Danoff]

I wonder how close we are to having AI versions of ourselves? I'm sure it has probably already been done, but I was thinking, if one could train a GPT to basically react exactly as you would, by uploading as many examples of your personal communications as possible - internet postings, personal messages (texts, online chats), emails, photos, FB posts etc., anything that captures how you actually speak - plus a survey/questionnaire to do a psychological profile. I bet a GPT could post stuff on the internet that would effectively be recognisably 'you', and even hold a conversation online with your characteristics, and even memories - depending on how much you told it about yourself...

It is already possible to have visual and audio avatars of yourself that are scarily close to reality and will only get more realistic - imagine an online persona that could actually think and talk like you as well, understand and replicate your reasoning, and accurately reflect your point of view... the implications could be massive, i.e. being able to 'speak' to someone after they had passed away. A bit creepy, but it's definitely now within the realms of possibility.

Black mirror right?

Google has adaptive input prediction that uses your emails and texts to try to figure out your next word when drafting an email that's based largely on your own communication style. I don't know how much farther beyond that it has gone.

 

[HenrySwanson]

With this and Open AI's o1?

Yeah, we're done.

 

[Sprite]

With this and Open AI's o1?

Yeah, we're done.

What is scarier is that, this is the stuff they let us use!

What do they have behind closed doors. I remember not long ago someone being ousted from an AI team claiming the AI was sentient. We all chuckled and (for me anyway) sided on the air of “yeh, ok dude, I don’t think so”… now I’m not too sure.

Edit…

Well looks like a comedian was at a corporate party for Salesforce. He joked about how the staff will be replaced by AI etc.

We use Salesforce at the company I work for. I won’t make retirement before I’m replaced by AI.

 

[TexRex]

I remember not long ago someone being ousted from an AI team claiming the AI was sentient. We all chuckled and (for me anyway) sided on the air of “yeh, ok dude, I don’t think so”… now I’m not too sure.

I am.

 

[HenrySwanson]

AI begins its ominous split away from human thinking

AIs have a big problem with truth and correctness – and human thinking appears to be a big part of that problem. A new generation of AI is now starting to take a much more experimental approach that could catapult machine learning way past humans.

newatlas.com

 

[Danoff]

Virtual artwork and video is taking such a fast leap forward that it's a bit unnerving. I'm seeing images and video now that are becoming much more difficult to differentiate from reality. I'm really not sure that we're mentally prepared for this. The technology exists to make Kamala a lizard person, or to show a futuristic-looking hurricane machine. The conspiracy nuts would fall for it so hard you'd never get them back.

Maybe the solution is to flood the internet with opposite conspiracies to get ahead of it and prove that such a thing isn't real. Like, make a lizard trump video or something. Or make a hurricane generating device and put "this is fake" all over it. We might need a conspiracy tech demo of sorts to educate people.

 

[TexRex]

The technology exists to make Kamala a lizard person

"make"

 

[Exorcet]

Virtual artwork and video is taking such a fast leap forward that it's a bit unnerving. I'm seeing images and video now that are becoming much more difficult to differentiate from reality. I'm really not sure that we're mentally prepared for this. The technology exists to make Kamala a lizard person, or to show a futuristic-looking hurricane machine. The conspiracy nuts would fall for it so hard you'd never get them back.

I'm less worried about the conspiracy theorists, they'll believe anything if they want to. The bigger problem in my mind is producing false evidence that can convince rational people, though in that case you could still follow a claimed document back to its source, or lack of one, to help verify it. AI is opening up new avenues for fakes and lies, but these things have always been around.

 

[TexRex]

I'm less worried about the conspiracy theorists, they'll believe anything if they want to. The bigger problem in my mind is producing false evidence that can convince rational people, though in that case you could still follow a claimed document back to its source, or lack of one, to help verify it. AI is opening up new avenues for fakes and lies, but these things have always been around.

The first thing one sees is still highly persuasive despite subsequent evidence countering it. Even rational people are still people, and this is hardwired in everyone.

This isn't to reject anything you said, which was great.

 

[Danoff]

The bigger problem in my mind is producing false evidence that can convince rational people

I'm actually less worried about the rational people because rational people can probably be convinced that something is fake. Maybe I shouldn't be.

Either way, most of AI generated fake stuff seems to be coming from the right. Trump saving kittens, black supporters, taylor swift supporters, etc. And so far they've done a pretty bad job of it. But that's not going to be the case for very long. I'd like to see folks go on the offensive.

There is a group of nutjobs that need to be put on defense, trying to disprove that something is real, before they're given something dangerous to fall for. The pizzagate crowd will show up with rifles at pizza shops trying to free non-existent children. So we need that crowd to do the exercise of disproving something so that they learn what's possible to fake.