AI 2023. Meet ChatGPT. - page 10

 
Aleksey Nikolayev describing objects. Formulate a request for the AI to write you the required language in the form of Backus-Naur.
Yes, I did, and continue to do so. Only my system has nothing to do with Backus-Noir form.

By the way, it would be nice to know the definition of an Object from an AI. Maybe in the context of programming.
 
artificial intelligence in a game!OpenAI
continued OpenAI

#openai | TikTok
  • www.tiktok.com
openai | 413M views. Watch the latest videos about #openai on TikTok.
 
There's no need to worry!

Here, set a professionally narrowly specialised, specifically formulated task for ChatDPT

.... Like this one:
Create or find a trading strategy that will generate profits of 20% p.a. or more and meet the following criteria
1) equity should not fall below 85%
2) recovery factor 2,0 and above
3) r-squared of errors 80% and above

And I am 100% sure that now (and in the near future as well) this AI is not capable of doing this kind of tasks (and it is also not capable in other highly specialised areas) - and examples like "output something to the console" are extremely primitive.
 
Alexandr Sokolov ChatDPT

.... Like this one:
Create or find a trading strategy that will generate profits of 20% p.a. or more and meet the following criteria
1) equity should not fall below 85%
2) recovery factor 2,0 and above
3) r-squared of errors 80% and above

And I am 100% sure that now (and in the near future as well) this AI is not capable of doing this kind of tasks (and it is also not capable in other highly specialised areas) - and examples like "output something to the console" are extremely primitive.
1- Determining who should be worried and why can be done after extensive and unbiased research. It is necessary to create a system of tests and evaluations, and make objective conclusions on the final results. This is an important and difficult task.

2. Primitive AI can't be called primitive at all will. When reading its answers, you can't help but be amazed. The boundaries of current capabilities are blurred, and what to expect from the future is unknown.

3. AI can be used effectively in trading today. It is not necessary to ask it to write a grail.
 
Реter Konow #:
2. The AI can no longer be called primitive, if you want to call it so. Reading its answers one can't help but be amazed. The boundaries of current capabilities are blurred, what to expect in the future is unknown.

I called primitive not AI, but the level of ChatGPT's work in highly specialised areas.


Retag Konow #:
3. AI can be used effectively in trading today. It is not necessary to ask it to write a grail.

It is. But I was talking about the fact that at the moment AI is unable to create a decent trading strategy on its own, not about the fact that AI in trading is untenable - without humans, AI will not create and improve itself

 
Alexandr Sokolov #:

I didn't call the AI primitive, but the level of ChatGPT's work in highly specialised areas specifically


It is. But I was talking about the fact that at the moment AI is not able to create a decent trading strategy on its own, not about the fact that AI in trading is untenable - without humans AI will not create and improve itself.

Special tests for highly specialised work will help to check its current capabilities and assess its immediate potential.

It is important to realise that AI is currently closed for training and undergoing beta testing. What opportunities the opening of narrowly specialised learning will bring at this stage we do not know and it is premature to draw conclusions.

I propose to think about tests for AI in the context of application in trading.
 

What if chatgpt feeds the metaquotes forum so that the questions are closer to reality?


not questions of course, but answers

 
ruslan #:

What if chatgpt feeds the metaquotes forum so the questions are closer to reality?


not questions of course, but answers

I think that anyway he will not be able to model and create decent trading strategies on his own - if you have any practice of creating them, you probably know that they are not available in the public domain and you have to collect ideas from various sources (the vast majority of these ideas are useless in practice) and come up with something of your own. To get into the logic of calculations to understand what is calculated, what it reflects and what it can be used for. To change the ideas themselves and combine them with each other

. That is, it is still an art to determine good entry points steadily at a distance!

And position maintenance is already a separate art - if you think that just set a fixed volume, TP and SL is enough, you are deeply mistaken

* I am not trying to make myself look like a "guru" - but even based on my small experience I can say that the whole and structured information on how to profitably and steadily trade on a long distance in open sources can not be found (at least for free) ! - in this case you can make an analogy with gold mining, when you have to wash tonnes of sand for the sake of a few ounces of gold.

 
ruslan #:

What if chatgpt feeds the metaquotes forum so that the answers are closer to reality?

... so it makes no sense to "feed" him what is in open sources - because such information will not teach him anything in essence, but will only pack his memory with useless slag.

 

List the things that ChatGPT can do?

-

  1. Text generation: ChatGPT can generate text based on a prompt or starter text. For example, if you give it the prompt "Write a short story about a time traveller", it can generate a complete story with a coherent plot and characters.

  2. Q&A: ChatGPT can answer questions based on its training data. For example, if you ask "What is the capital of France?", it can answer "The capital of France is Paris".

  3. Text summarisation: ChatGPT can summarise long pieces of text into shorter, more concise versions. For example, if you give it a long article, it can generate a summary that captures the main points of the article.

  4. Language translation: ChatGPT can translate text from one language to another. For example, if you give it a sentence in English, it can translate it into Spanish.

  5. Chat and Conversational AI: ChatGPT can be used to create chatbots that can have natural and consistent conversations with users. It can generate answers to questions and follow-up questions, providing a more natural conversational flow.

  6. Text completion: ChatGPT can complete text based on partial input. For example, if you give it the prompt "The sky is", it can generate the sentence "The sky is blue".

  7. Text classification: ChatGPT can classify text into various categories such as sentiment analysis (positive, negative, neutral) or topic classification (politics, sports, entertainment).

  1. Writing help: ChatGPT can help with writing tasks such as writing articles, stories, and emails. For example, you can give it a prompt or some initial text and it can generate text based on that, making the writing process faster and easier.

  2. Tone analysis: ChatGPT can analyse the mood of a piece of text and determine whether it is positive, negative or neutral. For example, it can analyse a customer review and determine whether the customer had a positive or negative experience.

  3. Text Correction: ChatGPT can correct grammatical errors in text, making it useful for writing help or language learning applications.

  4. Voice Conversations: ChatGPT can be integrated with voice-enabled systems such as voice assistants and smart speakers, allowing for natural and conversational interactions.

  5. Customisation: ChatGPT can be configured and fine-tuned for specific applications, such as customer service chatbots for a specific industry or language-specific chatbots.

These are additional tasks that ChatGPT can perform, demonstrating its versatility and potential for use in a wide range of applications. It is important to note that the accuracy and quality of the model output may vary depending on the specific task and training data used.

===

List the following things that ChatGPT cannot do?

-

  1. Replicate human emotions or empathy: ChatGPT is an AI model that generates responses based on statistical patterns in the training data and does not have the ability to understand or replicate human emotions.

  2. Make decisions or take action: ChatGPT is limited to generating text responses and does not have the ability to make decisions or take action on its own.

  3. Provide fully accurate or valid information: the model's responses are based on training data, which may contain errors, biases, or misinformation. The model may generate responses that are incorrect, misleading, or offensive.

  4. Understand the context of text in the same way as a human: ChatGPT relies on statistical patterns in the training data to generate responses and may not always understand the meaning and context of text in the same way as a human.

  5. Generate responses that are 100% consistent or natural: although ChatGPT generates human-like responses, they may not always be 100% consistent or natural, especially when the cue or context is unclear.

  6. Accurate reproduction of human writing styles and tones: although ChatGPT can generate text in a variety of writing styles and tones, it cannot always reproduce them accurately.

  1. Interacting with the physical world: ChatGPT is a language model and does not have the ability to interact with the physical world such as moving objects or controlling devices.

  2. Addressing ethical and moral issues: As an AI model, ChatGPT does not have a moral compass or the ability to address ethical and moral issues. This means that it can generate responses that are inappropriate or offensive, and it is important to monitor and regulate its use.

  3. Replace human judgement or creativity: ChatGPT is a powerful text generation tool, but it cannot replace human judgement or creativity. It is important to use ChatGPT in conjunction with human intelligence, not as a substitute for it.

  4. Detecting sarcasm or irony: ChatGPT is not able to detect sarcasm or irony in text, as these are complex human emotions that are difficult to replicate in an AI model. This may cause the model to generate inappropriate or incorrect responses in certain situations.

  5. Interpreting visual or audio input: ChatGPT is a language model and is designed to process and generate text. It is not capable of interpreting visual or audio input such as images or speech.

These limitations demonstrate the limitations of ChatGPT and it is important to keep this in mind when using the model for AI dialogue applications. The accuracy and quality of the output of the model may vary depending on the specific task and the training data used.