AI 2023. Meet ChatGPT. - page 101

New comment
 
 
Hmm, I tried, but after logging in I seem to be "stuck in traffic": my page was and still is blank and nothing moves.

Here's what I wanted to ask:
Write me an MQL5 script for Metatrader 5 that reads the XML table of an optimization in MT5's Strategy Tester and statistically evaluates which variables have a relevant impact on the result.<br/ translate="no"> Write me an MQL5 script for Metatrader 5 that reads the XML table of an optimization in MT5's Strategy Tester and statistically evaluates which variables have a relevant impact on the result.
A question that almost everyone faces.

Maybe someone else "on the ChatGPT freeway" can ask and post the result?
What an almost criminal boss says to his subordinates: Surprise me ;)
 

// I put it in the wrong place, sorry, corrected the post

 
Carl Schreiber #:
Hmm, I tried, but after logging in I seem to be "stuck in traffic": my page was and still is blank and nothing moves.

Here's what I wanted to ask: A question that almost everyone faces.

Maybe someone else "on the free path in ChatGPT" can ask and post the result?
What an almost criminal boss says to his subordinates: Surprise me ;) 
#include <XmlRpcClient.inc>

void OnStart()
{
   // Read the XML file
   CXmlRpcClient xml;
   if(!xml.Load("OptimizationReport.xml"))
   {
      Print("Failed to load XML file");
      return;
   }

   // Get the optimization results
   CXmlNode node = xml.GetChild("optimizationResults");
   if(node.IsNull())
   {
      Print("Failed to get optimization results");
      return;
   }

   // Get the parameter names and values
   CXmlNode parameters = node.GetChild("parameters");
   if(parameters.IsNull())
   {
      Print("Failed to get parameters");
      return;
   }

   // Get the results
   CXmlNode results = node.GetChild("results");
   if(results.IsNull())
   {
      Print("Failed to get results");
      return;
   }

   // Calculate the correlation between each parameter and the result
   double correlation[parameters.GetChildrenCount()];
   for(int i = 0; i < parameters.GetChildrenCount(); i++)
   {
      CXmlNode parameter = parameters.GetChild(i);
      if(parameter.IsNull())
      {
         Print("Failed to get parameter");
         return;
      }

      double values[results.GetChildrenCount()];
      for(int j = 0; j < results.GetChildrenCount(); j++)
      {
         CXmlNode result = results.GetChild(j);
         if(result.IsNull())
         {
            Print("Failed to get result");
            return;
         }

         if(parameter.GetAttribute("value") == result.GetAttribute("parameters"))
         {
            values[j] = result.GetAttributeDouble("profit");
         }
      }

      correlation[i] = Correlation(values, results.GetChildrenCount());
   }

   // Print the correlation values
   for(int i = 0; i < parameters.GetChildrenCount(); i++)
   {
      CXmlNode parameter = parameters.GetChild(i);
      if(parameter.IsNull())
      {
         Print("Failed to get parameter");
         return;
      }

      PrintFormat("%s: %f", parameter.GetAttribute("name"), correlation[i]);
   }
}

double Correlation(double values[], int count)
{
   double sumX = 0, sumY = 0, sumXY = 0, sumX2 = 0, sumY2 = 0;

   for(int i = 0; i < count; i++)
   {
      sumX += i;
      sumY += values[i];
      sumXY += i * values[i];
      sumX2 += i * i;
      sumY2 += values[i] * values[i];
   }

   double numerator = count * sumXY - sumX * sumY;
   double denominator = sqrt((count * sumX2 - sumX * sumX) * (count * sumY2 - sumY * sumY));

   return numerator / denominator;
}

Where to find the XmlRpcClient.inc file couldn't tell me)

Asked on perplexity.ai

 
Vitaliy Kuznetsov #:

To make such a conclusion, the honourable Nobel Prize winner in economics needs to provide the public with a detailed study. For example, such as this: 2303.10130.pdf (arxiv.org) ("An early look at the labour market impact potential of large language models". Authors, OpenAI researchers).

Given the limited application of LLMs and image generators (music, video), most occupations will remain unaffected and productivity will not increase. Consequently, in most manufacturing and service industries, the pace and volume of work will not change and there will be no reason to shorten the working week.


These are the conclusions of the above study:

We investigate the potential effects of large language models (LLMs), such as Generative Pre-trained Transformers (GPTs), on the U.S. labour market, focusing on the increased opportunities that arise from LLM-enabled software compared to LLMs on their own. Using a new methodology, we evaluate employment based on its relevance to LLM opportunities, integrating both human expertise and GPT-4 classifications.

Our research suggests that about 80% of the US workforce may have at least 10% of their job tasks affected by LLM implementation, while approximately 19% of workers may see at least 50% of their tasks affected. We make no predictions about the timing of the development or adoption of such LLMs.

The projected effects span all wage levels, with high-wage jobs likely to face greater LLM opportunities and LLM-powered software. Significantly, these effects are not restricted to industries with higher recent productivity growth.

Our analysis shows that with access to LLM, approximately 15% of all worker tasks in the US can be completed significantly faster at the same level of quality. This figure increases to 47-56% of all tasks when LLM-based software and tools are used. This finding indicates that LLM-enabled software will have a significant impact on expanding the economic impact of the underlying models.

We conclude that LLMs such as GPT have the properties of general-purpose technologies indicating significant economic, social, and political implications.

//=================


//=================

Bottom line:

  • 80% of the workforce will speed up (unknown by how much) 10% of their tasks.
  • 19% of the workforce will see 50% of tasks affected (unknown how much affected).
  • 15% of all tasks could be completed significantly faster (which tasks and how much faster is unknown).
  • This increases to 47-56% of all tasks (i.e. software tasks) when using LLM software and tools.

That's a lot of uncertainty, isn't it? This study doesn't provide anything. Finger in the sky.

Let's assume that specialists will reduce their working hours thanks to LLMs. But what about ordinary workers? They're in the majority. Those who dig, weld, build, haul loads, mine minerals, process waste, work in shops, maintain machinery, etc.? Will the AI start giving them tips and they'll finish their shift sooner?

Hype and nonsense. Too bad it's on behalf of a Nobel laureate.

 
Реter Konow #:

To make such a conclusion, the honourable Nobel Prize winner in economics needs to provide the public with a detailed study. For example, such as this: 2303.10130.pdf (arxiv.org) ("An early look at the labour market impact potential of large language models". Authors, OpenAI researchers).

***

Hype and nonsense. It's a pity it's in the name of a Nobel laureate.

There's already been research on the 4-day diet. And even in Stalin's time there was such an idea.

Modern practice has proven the effectiveness of 4-day work in many professions.

Again, the more a person rests, the more money he spends. It's good for the economy.

Well, the fact that AI in many professions will increase efficiency, will help to introduce almost universally the practice of 4-day work.

Well, and jobs can be made more (for AI space will be freed up). It will turn out that AI will not take away anyone's work, but will add extra weekends.

//we need an extra day off to keep up with trends and generate pictures).

 
Vitaliy Kuznetsov #:

There's already been research on the 4-day diet. And even in Stalin's time there was such an idea.

Modern practice has proven the effectiveness of the 4-day programme in many professions.

Again, the more a person rests, the more money he spends. This is also good for the economy.

Well, the fact that AI will increase the efficiency of many professions will help to introduce almost universally the practice of 4-day work.

Well, and jobs can be made more (for AI space will be freed up). It will turn out that AI will not take away anyone's work, but will add extra days off.

I wonder if there were computers in Stalin's time? Internet? Smartphones? Satellites?

Now there are all these things and the working week hasn't changed. Amazing!))

 

accept the programme on MT4 on crossing of exponential averages with periods 8 and 21

------------------------------------------------------------------------------------- 

принять программу на МТ4 по пересечению экспоненциальных средних с периодами 8 и 21
-------------------------------------------------------------------------------------






int emaPeriod1 = Input("EMA Period 1", 8, 1, 100);



int emaPeriod2 = Input("EMA Period 2", 21, 1, 100);




  


double ema1[];



double ema2[];




  


int init() {



   SetIndexBuffer(0, ema1);



   SetIndexBuffer(1, ema2);




  


   SetIndexStyle(0, DRAW_LINE);



   SetIndexStyle(1, DRAW_LINE);




  


   SetIndexLabel(0, "EMA(" + emaPeriod1 + ")");



   SetIndexLabel(1, "EMA(" + emaPeriod2 + ")");




  


   SetIndexShift(0, emaPeriod1);



   SetIndexShift(1, emaPeriod2);




  


   return(INIT_SUCCEEDED);



}



int start() {



   int limit = Bars - 1;



   int counted_bars = IndicatorCounted();



   if (counted_bars < 0) {



      return(-1);



   }



   if (counted_bars > 0) {



      limit = Bars - counted_bars;



   }




  


   for (int i = limit; i >= 0; i--) {



      ema1[i] = iMA(NULL, 0, emaPeriod1, 0, MODE_EMA, PRICE_CLOSE, i);



      ema2[i] = iMA(NULL, 0, emaPeriod2, 0, MODE_EMA, PRICE_CLOSE, i);



   }




  


   if (ema1[0] > ema2[0] && ema1[1] <= ema2[1]) {



      // пересечение вверх



      // Добавьте код для открытия длинной позиции



   }



   else if (ema1[0] < ema2[0] && ema1[1] >= ema2[1]) {



      // пересечение вниз



      // Добавьте код для открытия короткой позиции



   }




  


   return(0);



}



if (ema1[0] > ema2[0] && ema1[1] <= ema2[1]) {



   // пересечение вверх



   double lotSize = ...; // задайте желаемый размер лота
 
Реter Konow #:
Okay, that's the most important thing. What if the code has errors or works incorrectly? What conclusion will you draw in this case?

I'm not a bad programmer and I can use AI code to create a quick dummy with further refinement. And talking about when AI will take over the universe doesn't interest me. I've been in free flight for many years and I need to make money. If GPT will help me in any way, that's great.

And in general, for me there are levels coder->developer->generator of ideas (trading algorithms). There are plenty of coders around and freelancing, they are IT labourers. A developer can take an array of ideas as input, test them and produce a candy bar. I understand about the generator :).

 
Alexey Volchanskiy #:

I'm not a bad programmer and I can use AI code to create a quick dummy with further refinement. And talking about when AI will take over the universe doesn't interest me.

1. You yourself asked the AI if it was going to take over the universe. In your post.

2. You concluded that freelancers should go to the factory for thirty dollars because GPT-4 is many times cooler than ChatGPT, and the latter wrote code that they haven't tested but are sure it fits.

3. I gave a video with a situation between GPT-4 and a "customer" nub with a TK, who tries to do his own task using AI. A good demonstration of the real capabilities of AI. After watching it, update your conclusions about freelancing.
1...949596979899100101102103104105106107108...216
New comment