[et_pb_section fb_built=”1″ _builder_version=”4.16″ locked=”off” global_colors_info=”{}”][et_pb_row _builder_version=”4.19.0″ _module_preset=”default” global_colors_info=”{}”][et_pb_column type=”4_4″ _builder_version=”4.19.0″ _module_preset=”default” global_colors_info=”{}”][et_pb_text _builder_version=”4.19.3″ text_font=”Varela Round||||||||” text_line_height=”1.5em” background_size=”initial” background_position=”top_left” background_repeat=”repeat” custom_padding=”||0px|||” global_colors_info=”{}”]<\/p>\n
There is a lot of talk about artificial intelligence or AI. But this is not something of the last few months as it seems, but for several years. AI has been in the technology sector for quite some time and many companies have been researching and working with it to make people’s lives a little easier through technology<\/strong>.<\/p>\n But it is only now that truly amazing advances have been made in this field.<\/p>\n Yes, I’m talking about GPT-3<\/strong>. These acronyms are on everyone’s lips because of how they have revolutionized the AI world.<\/p>\n Do you want to know why?<\/p>\n Read on and I’ll tell you.<\/p>\n [\/et_pb_text][\/et_pb_column][\/et_pb_row][et_pb_row _builder_version=”4.19.0″ _module_preset=”default” global_colors_info=”{}”][et_pb_column type=”4_4″ _builder_version=”4.19.0″ _module_preset=”default” global_colors_info=”{}”][et_pb_text _builder_version=”4.19.3″ text_font=”Varela Round||||||||” text_line_height=”1.5em” header_2_font=”Varela Round||||||||” header_2_line_height=”1.2em” header_3_font=”Varela Round||||||||” header_3_line_height=”1.2em” background_size=”initial” background_position=”top_left” background_repeat=”repeat” custom_margin=”15px|||||” custom_padding=”||21px|||” locked=”off” global_colors_info=”{}”]<\/p>\n Antes de conBefore we go on about what’s going on with AI right now, let’s start at the beginning.<\/span><\/p>\n Instead, with customer service automation, you optimize time from two points of view<\/strong>:<\/span><\/p>\n [\/et_pb_text][et_pb_text _builder_version=”4.19.0″ text_font=”Varela Round||||||||” text_line_height=”1.5em” header_2_font=”Varela Round||||||||” header_2_line_height=”1.2em” header_3_font=”Varela Round||||||||” header_3_line_height=”1.2em” background_color=”rgba(239,240,255,0.45)” background_size=”initial” background_position=”top_left” background_repeat=”repeat” min_height=”96.2px” custom_margin=”||||false|false” custom_padding=”15px|25px|15px|25px|true|true” border_radii=”on|15px|15px|15px|15px” locked=”off” global_colors_info=”{}”]<\/p>\n Artificial intelligence based on neural networks has been around for quite some time, but imitating the behavior of the human brain in a very basic way<\/strong> (e.g., detecting SPAM in email services).<\/p>\n But nowadays, the field of neural networks has advanced by leaps and bounds<\/strong> due to the creation of new algorithms, investment in R&D, etc.<\/p>\n [\/et_pb_text][\/et_pb_column][\/et_pb_row][et_pb_row column_structure=”2_3,1_3″ _builder_version=”4.19.3″ _module_preset=”default” custom_padding=”||26px|||” global_colors_info=”{}”][et_pb_column type=”2_3″ _builder_version=”4.19.3″ _module_preset=”default” global_colors_info=”{}”][et_pb_text _builder_version=”4.19.3″ text_font=”Varela Round||||||||” text_line_height=”1.5em” background_size=”initial” background_position=”top_left” background_repeat=”repeat” global_colors_info=”{}”]<\/p>\n Thanks to this, an AI has been able to express itself like a human and answer our queries so concisely<\/strong>. With this technology, it is no longer necessary to tell the machine “if you are asked A, answer B, if you are asked C, answer B”, for example.<\/p>\n [\/et_pb_text][\/et_pb_column][et_pb_column type=”1_3″ _builder_version=”4.19.3″ _module_preset=”default” global_colors_info=”{}”][et_pb_image src=”https:\/\/arkibot.com\/en\/wp-content\/uploads\/2023\/02\/jerarquia-datos-128.png” title_text=”jerarquia-datos-128″ align=”center” _builder_version=”4.19.3″ _module_preset=”default” global_colors_info=”{}”][\/et_pb_image][\/et_pb_column][\/et_pb_row][et_pb_row _builder_version=”4.19.3″ _module_preset=”default” global_colors_info=”{}”][et_pb_column type=”4_4″ _builder_version=”4.19.3″ _module_preset=”default” global_colors_info=”{}”][et_pb_text _builder_version=”4.19.3″ text_font=”Varela Round||||||||” text_line_height=”1.5em” background_size=”initial” background_position=”top_left” background_repeat=”repeat” global_colors_info=”{}”]<\/p>\n This is where Generative Pre-trained Transformer 3 technology, better known as GPT-3, comes into play. This language model is capable of recognizing patterns in the data and learning through examples<\/strong> (deep learning) to imitate the way a human writes.<\/p>\n In other words, GPT-3 generates meaningful texts similar to those written by a human from a suggestion. In fact, it is very difficult to identify whether they are written by an artificial intelligence or by a human intelligence<\/strong>.<\/p>\n Currently this AI language model created by the company OpenAI is the most powerful to date that has been opened to the public.<\/p>\n [\/et_pb_text][\/et_pb_column][\/et_pb_row][et_pb_row _builder_version=”4.19.3″ _module_preset=”default” custom_padding=”||10px|||” global_colors_info=”{}”][et_pb_column type=”4_4″ _builder_version=”4.19.0″ _module_preset=”default” global_colors_info=”{}”][et_pb_text _builder_version=”4.19.3″ text_font=”Varela Round||||||||” text_line_height=”1.5em” header_2_font=”Varela Round||||||||” header_2_line_height=”1.2em” header_3_font=”Varela Round||||||||” header_3_line_height=”1.2em” background_size=”initial” background_position=”top_left” background_repeat=”repeat” custom_margin=”15px|||||” custom_padding=”||6px|||” locked=”off” global_colors_info=”{}”]<\/p>\n The goal of GPT-3 is to predict what will go next in the text based on previous data, filling in the information gaps as in a kind of auto-completion. [\/et_pb_text][et_pb_text quote_border_color=”#EC4899″ _builder_version=”4.19.3″ text_font=”Varela Round||||||||” text_text_color=”#6366F1″ text_line_height=”1.5em” quote_text_color=”#EC4899″ header_2_font=”Varela Round||||||||” header_2_line_height=”1.2em” header_3_font=”Varela Round||||||||” header_3_line_height=”1.2em” background_size=”initial” background_position=”top_left” background_repeat=”repeat” custom_margin=”15px||-2px|||” custom_padding=”||6px|||” locked=”off” global_colors_info=”{}”]<\/p>\n <\/span>To give you an idea of the size, the GPT-2 version had 1.5 billion parameters. GPT-3 has 175 billion and they have used a total of 500 billion words.\u00a0<\/span><\/span><\/p>\n<\/blockquote>\n [\/et_pb_text][et_pb_text _builder_version=”4.21.0″ text_font=”Varela Round||||||||” text_line_height=”1.5em” header_2_font=”Varela Round||||||||” header_2_line_height=”1.2em” header_3_font=”Varela Round||||||||” header_3_line_height=”1.2em” background_size=”initial” background_position=”top_left” background_repeat=”repeat” custom_margin=”15px|||||” custom_padding=”1px||6px|||” locked=”off” global_colors_info=”{}”]<\/p>\n You can try GPT-3 right now from its free application<\/a><\/span> without having to download anything at all, all you have to do is register and indicate what you are going to use it for.<\/p>\n [\/et_pb_text][\/et_pb_column][\/et_pb_row][et_pb_row _builder_version=”4.19.3″ _module_preset=”default” global_colors_info=”{}”][et_pb_column type=”4_4″ _builder_version=”4.19.3″ _module_preset=”default” global_colors_info=”{}”][et_pb_text _builder_version=”4.19.3″ text_font=”Varela Round||||||||” text_line_height=”1.5em” header_2_font=”Varela Round||||||||” header_2_line_height=”1.2em” header_3_font=”Varela Round||||||||” header_3_line_height=”1.2em” background_size=”initial” background_position=”top_left” background_repeat=”repeat” custom_margin=”15px|||||” custom_padding=”||6px|||” locked=”off” global_colors_info=”{}”]<\/p>\n GPT-3 can do almost anything with words in it: answer questions, summarize texts, create poetry, write advertisements, compose songs, program codes… Basically everything a human being is capable of doing through language.<\/strong> [\/et_pb_text][\/et_pb_column][\/et_pb_row][et_pb_row _builder_version=”4.19.3″ _module_preset=”default” custom_padding=”||13px|||” global_colors_info=”{}”][et_pb_column type=”4_4″ _builder_version=”4.19.3″ _module_preset=”default” global_colors_info=”{}”][et_pb_text _builder_version=”4.19.3″ text_font=”Varela Round||||||||” text_line_height=”1.5em” header_2_font=”Varela Round||||||||” header_2_line_height=”1.2em” header_3_font=”Varela Round||||||||” header_3_line_height=”1.2em” background_size=”initial” background_position=”top_left” background_repeat=”repeat” custom_margin=”15px|||||” custom_padding=”||6px|||” locked=”off” global_colors_info=”{}”]<\/p>\nWhat is GPT-3?<\/h2>\n
How does it work?<\/b><\/h2>\n
<\/span>
<\/span>To achieve this, he has needed a previous training that has consisted of collecting and filtering the most relevant knowledge that is published today on the network<\/strong>: from public books, scientific papers, web pages, information from Wikipedia, etc.; and learning it.
<\/span><\/p>\n\n
What is it good for?<\/h2>\n
<\/span>
<\/span>Here you can see some examples<\/a><\/span> from its own API.
<\/span>
<\/span>Many companies have already seen in this model an opportunity to advance on the path of artificial intelligence.
<\/span>
<\/span>The well-known and talked-about chatbot<\/span> ChatGPT<\/span><\/a><\/span> company’s own OpenAI or Canva and its Magic Write<\/span><\/a>.<\/span>
<\/span>
<\/span>And of course, here we must include and highlight in bold our own chatbot tool <\/b>Arkibot<\/b><\/a><\/span>.<\/span><\/p>\nHow does Arkibot use GPT-3?<\/h2>\n