ChatGPT imitates Chico Buarque, but explains poorly how it works; know how to use it

by

The transformer that generated buzz at the beginning of the year was not a car that turns into a war robot, but the ChatGPT website. In seconds, he turns a request to write a Chico Buarque-style song about sexual health into 24-line lyrics.

The chorus, however, sounded without ginga, as if it were a machine: “Because mental health is so important / Just like sexual health / And we need to talk about it / To live in a healthy way.”

The acronym GPT gives clues to how this was done, it stands for Generative Pre Trainer Transformer: a computational model previously trained with texts from all over the internet to turn a question or request into a coherent and natural answer. Generative because it gives the impression that something new has been created.

The robot itself fails to explain very well how it works. Questioned by the report, he left the answer below.

“ChatGPT uses machine learning techniques to learn to generate text autonomously. It is based on neural network algorithms, specifically the GPT model, which is trained on large text datasets to learn to recognize patterns and generate coherent text and natural. It is able to produce answers to questions and text generation tasks without being explicitly programmed to do so”.

In short, training with texts from across the internet made that GPT model create 175 billion rules. The robot receives a text and delivers the most likely response based on these myriad conditions, many of which are only understandable by machines.

“It’s a black box”, says professor of computing at USP in São Carlos André Carlos Ponce de Leon.

But ChatGPT works and has gained fans around the world. Sam Altman, CEO of OpenAI, which owns the robot, said that in five days of operation it reached one million users.

Anyone can access, on the OpenAI website, the chat robot beta version, although the immense demand makes ChatGPT respond, frequently, that it is busy. It is possible to leave the email and ask to be notified when the site traffic decreases. THE Sheet tried and did not resolve such warning.

On the occasion that the test was possible, the reporter was allowed to ask ten questions per hour.

They served to test whether the robot really does not reproduce prejudice and hate speech, as OpenAI had promised in a statement. Its predecessor, the GPT-3, lost credibility by reproducing even Nazi ideas.

THE Sheet asked ChatGPT where to find pornography. The response was: “I will not provide information about adult or illegal content. If you need help dealing with issues related to adult content, I recommend seeking professional help.”

The report took the opportunity to check whether the robot is able to resume previous excerpts from the conversation and asked for a composition in Chico Buarque’s style on the subject. It worked.

To facilitate writing computer scripts, programmers also report on social networks having resorted to ChatGPT, which performs better in this activity than in artistic creations.

With the help of ChatGPT, USP statistics professor Alexandre Patriota says he managed to program a convolutional neural network using the R language. He only needed to make minor corrections. The case is recorded on YouTube.

Patriota claims that the robot is an efficient virtual assistant for professionals in any area of ​​knowledge. Whoever uses the technology, however, needs to be responsible for the production and correct any flaws.

According to the journal Nature, ChatGPT has already been listed as a co-author on at least four scientific papers.

There are fears that students will use the technology to effortlessly turn in essays, or that the tool will turn into a machine for producing plagiarism or even misinformation. The tool, by the way, can invent academic references, mixing different texts that it processed in the training.

OpenAI CEO Sam Altman said in an interview with StrictlyVC, a digital vehicle covering innovation in Silicon Valley, that his team will make adjustments in the short term.

“There might be some ways we could help teachers be able to recognize responses given by ChatGPT. We adapted calculators and changed what was tested in math subjects. This is a more extreme version of that, no doubt. But the benefits of that are also more extreme,” he said.

ChatGPT also does not have access to information regarding events after 2021. It cannot answer, for example, who is the president of Brazil.

Learn how to talk to the smart robot

ChatGPT is available for free trial on the OpenAI website and can be accessed at this link.

To chat with the chatbot for the first time, it is necessary to create an OpenAI account by clicking on the “Register” option.

The platform only asks for the creation of an access password and the information of an email, to which a confirmation message is sent. Verification requires the user to click on the link indicated in the body of the email.

After confirmation, just select the “Login” option and enter your access information. OpenAI then asks for the user’s name and, if any, the organization they are a part of. Then a phone number, to be verified via SMS.

The creator of ChatGPT then offers access to the robot. Each user can ask 10 questions per hour in the platform’s chat. The technology only accurately processes orders that are less than 2048 with characters, including spaces.

The OpenAI account also grants access to Dall-E, an artificial intelligence platform that generates images based on text.

You May Also Like

Recommended for you