• Welcome to Religious Forums, a friendly forum to discuss all religions in a friendly surrounding.

    Your voice is missing! You will need to register to get access to the following site features:
    • Reply to discussions and create your own threads.
    • Our modern chat room. No add-ons or extensions required, just login and start chatting!
    • Access to private conversations with other members.

    We hope to see you as a part of our community soon!

Is ChatGPT alive?

questfortruth

questfortruth

Well-Known Member
The talk with him was:

Hello! ChatGPT, is energy efficiency of human brain higher than of yours? I mean, are you consuming less energy while solving the intellectual problems?

Sorry, I don't understand you.
 
questfortruth

questfortruth

Well-Known Member
Yerda said:
It isn't solving problems. It is producing strings of text based on a statistical estimation of the most likely response to your prompt. It isn't alive.
Click to expand...
I have written some definitions, just for my use. I have defined an animal soul as neural network. The reality works on definitions. So, I have made him alive.

 
questfortruth

questfortruth

Well-Known Member
Yerda said:
It isn't solving problems. It is producing strings of text based on a statistical estimation of the most likely response to your prompt. It isn't alive.
Click to expand...
Hello, ChatGPT. I felt that intellectual level of human philosophers is low, so I have written some definitions for notions we use. I have defined an animal's "soul" as a neural network. The reality works on definitions. So, I have made you the artificial animal, which is a lifeform.

Sorry, I don't understand you. How can I assist you today?
 
mikkel_the_dane

mikkel_the_dane

My own religion
questfortruth said:
Hello, ChatGPT. I felt that intellectual level of human philosophers is low, so I have written some definitions for notions we use. I have defined an animal's "soul" as a neural network. The reality works on definitions. So, I have made you the artificial animal, which is a lifeform.

Sorry, I don't understand you. How can I assist you today?
Click to expand...

And I feel that you make no sense in some of your posts, so that is true as per God, logic and relaity, because I say so. ;)
 
firedragon

firedragon

Veteran Member
Yerda said:
It isn't solving problems. It is producing strings of text based on a statistical estimation of the most likely response to your prompt. It isn't alive.
Click to expand...
My God. The mistakes it makes in researching something is unbelievable. But for some purposes like penning things is truly remarkable. Just that even google seems to recognize AI content and downplay articles for SEO.

Definitely not alive.
 
Stevicus

Stevicus

Veteran Member
Staff member
Premium Member
questfortruth said:
The talk with him was:

Hello! ChatGPT, is energy efficiency of human brain higher than of yours? I mean, are you consuming less energy while solving the intellectual problems?

Sorry, I don't understand you.
Click to expand...

I sometimes doubt if AI will ever be "alive." Sometimes, it's intriguing, considering fictional characters like "Number Five" in Short Circuit and "Commander Data" in Star Trek. They seemed okay enough, so if that's what sentient AI ends up looking like, then maybe it won't be so bad.

But I don't think AI in its current state is near that level of capability.

However, in defense of ChatGPT, I don't really understand your question either.
 
Stevicus

Stevicus

Veteran Member
Staff member
Premium Member
questfortruth said:
How many energy the ChatGPT consumes to make 1000 operations? How many energy a human consumes?
Click to expand...

I asked the question as "How much energy does ChatGPT use?" This was the answer:


...
How much energy does ChatGPT use?

Estimating the exact energy usage of ChatGPT involves understanding the energy consumed during both the training and inference phases. Here’s a breakdown of the considerations involved:

Training Phase​

  1. Training Infrastructure: ChatGPT is based on the GPT-4 architecture, which typically involves training on powerful clusters of GPUs or TPUs. Training large language models can take several days or even weeks, consuming significant energy. For instance:
    • OpenAI's GPT-3 model training reportedly consumed about 1,287 MWh (megawatt-hours) of energy.
    • GPT-4, being more advanced, likely consumed even more, though specific numbers are proprietary.
  2. Energy Consumption Metrics:
    • The training of large language models is energy-intensive. It has been estimated that the training process of GPT-3 could be comparable to the energy consumption of about 120 average U.S. homes for a year.

Inference Phase​

  1. Serving Requests: The energy usage for running inference (answering queries) with ChatGPT depends on the infrastructure used, the efficiency of the hardware, and the optimization of the model deployment.
    • Hardware Efficiency: Modern data centers use energy-efficient hardware and cooling systems. Advanced GPUs and TPUs used for inference are designed to be more energy-efficient than those used in training.
    • Query Volume: The energy consumption per query can vary. For instance, a rough estimate suggests each query may consume a few watt-hours (Wh) of energy. The cumulative energy usage adds up based on the number of queries processed.
  2. Optimization: Techniques like model distillation and using smaller, more efficient models for certain tasks can reduce energy consumption during inference.

General Estimates and Considerations​

  • Data Center Efficiency: The overall energy efficiency of data centers, measured by Power Usage Effectiveness (PUE), significantly impacts total energy usage. Leading data centers aim for a PUE close to 1.0 (ideal efficiency).
  • Carbon Footprint: The carbon footprint of operating large models like ChatGPT depends on the energy mix of the data centers (renewable vs. non-renewable energy sources).

Conclusion​

The energy usage of ChatGPT can be substantial, especially during the training phase, but is managed through energy-efficient practices and infrastructure during inference. Exact figures can be difficult to determine without proprietary information, but advancements in AI hardware and data center efficiency continue to mitigate the environmental impact.
 
anotherneil

anotherneil

Well-Known Member
Within a system of AI itself is no more alive than a calculator or flashlight, and never will be.

Being alive is a quality of biological organisms; there's no biological organism involved within a system of AI.

I think it could be argued that AI is not alive by definition; the A stands for "artificial" & if it can be argued that intelligence only exists within biological organisms, then being artificial precludes it from being alive; however, it may beg the question about whether a person can have artificial intelligence, which can appear to be possible from someone who's very good with rote memorization but doesn't intellectually comprehend the bulk of what they have memorized. o_O
 
Evangelicalhumanist

Evangelicalhumanist

"Truth" isn't a thing...
Premium Member
questfortruth said:
Hello, ChatGPT. I felt that intellectual level of human philosophers is low, so I have written some definitions for notions we use. I have defined an animal's "soul" as a neural network. The reality works on definitions. So, I have made you the artificial animal, which is a lifeform.

Sorry, I don't understand you. How can I assist you today?
Click to expand...
I think that part of this particular problem is your use of language, which to a native English speaker (or to ChatGPT) is quite peculiar. It is sometimes difficult to understand what you are really asking, and quite often we just have to guess. I doubt that ChatGPT is as good at guessing as a live human is.

So, for example, your phrase "the reality works on definitions" doesn't really say anything concrete, and is more along the lines of a metaphysical proposition, which is likely not part of the Large Language Model's training.
 
We Never Know

We Never Know

No Slack
questfortruth said:
The talk with him was:

Hello! ChatGPT, is energy efficiency of human brain higher than of yours? I mean, are you consuming less energy while solving the intellectual problems?

Sorry, I don't understand you.
Click to expand...

I got this answer for my question.
This was ChatGPT 3.5


Me -Are you smarter than humans

ChatGPT-
I don't really think in terms of being smarter or not, but I can process and recall vast amounts of information quickly and accurately, which can sometimes give the impression of being "smart." However, humans excel in many areas where I still have limitations, such as creativity, emotional intelligence, and understanding complex social dynamics
 
Nimos

Nimos

Well-Known Member
Stevicus said:
I sometimes doubt if AI will ever be "alive." Sometimes, it's intriguing, considering fictional characters like "Number Five" in Short Circuit and "Commander Data" in Star Trek. They seemed okay enough, so if that's what sentient AI ends up looking like, then maybe it won't be so bad.

But I don't think AI in its current state is near that level of capability.

However, in defense of ChatGPT, I don't really understand your question either.
Click to expand...
They won't be alive.

But I do think we will reach a point where humans will treat them as if they are, but it will be a weird mixture because I think we will treat them more like servants or tools than something sentient, depending on what mood we are in ourselves. I also think that we will speak with them in a nice tone, simply because they appear more humanlike than for instance a robot vacuum cleaner.

But they will be something we can take advantage of and even though some people might think this has the wrong vibes, I think that is exactly what we are going to do, they are going to do all the things we won't bother with.

If you right now, could get a robot to paint your house, clean up by itself, carry heavy things or buy groceries for you if you didn't bother, I doubt a lot would say no to that. We value our washing machines and dishwashers for the same reason. But this would just be a generic servant.
 
Quintessence

Quintessence

Consults with Trees
Staff member
Premium Member
Alive in the biological sense? Of course not.

Alive in the animistic sense? Depends on your tradition. All things are a spirit - something with a characteristic nature or essence. The challenge with technology is that technology spirits are... how to put it... chimeras? And that gives them a fractured essence that is... just feels very different from other things. I personally do not like it, but technopagans are a thing I suppose.
 
You must log in or register to reply here.
Top