The Chatbot Tool pays this user $ 50 per month for their comments on artificial intelligence models

“All Amnesty International All” is a kind of description line. ” “We have organized all artificial intelligence models that we can find today.” YUPP encourages the site of the developers to communicate if they want to add their language or photo form to the options. It currently does not have any deals with AI’s builders, and these responses are provided by making API calls.

Every time someone uses YUPP, he participates in a face -to -face comparison of two Chatbot models, and sometimes they get a bonus to provide their notes and choose a profitable answer. Basically, it is a survey of users disguised in a fun game. (The site is with it a lot From expressive symbols.)

He believes that the data is circulating in this position for users as more clear than previous consumer applications, such as Twitter – who rushes to tell me that he was the twenty -seventh employee in, and now he has a founder in that company, Biz Stone, as one of his supporters. “This is a little transition from previous consumer companies,” he says. “You can provide counter -feeding data, and that these data will be used in an unidentified manner and send it to models.”

This leads us to where the real money is in: selling human comments to artificial intelligence companies that want more data to control their models.

“The collective designer human assessments are what we do here,” says Gupta. It is estimated that the amount of money that users can do will add enough to some coffee cups per month. Nevertheless, this type of mark on data is often called Learning reinforcement with human reactions In the artificial intelligence industry, it is very valuable for companies because they make repetitive models and control outputs. It deserves much more than the most famous coffee cup in San Francisco.

The main competitor of YUPP is a web site called LMARNAWhich is very popular with those familiar with artificial intelligence to obtain notes on new models and bragging rights if a new version rises to the upper part of the package. Whenever a strong model is added to LMARNA, it often raises rumors about what the main company tries to test its new version in the hide.

“This is a two -sided product with the effects of the network for consumers who help models.” “We hope that the model builders will improve the models and introduce them to consumers.” It displays me an experimental version of the YUPP leaders, which runs live today and includes a comprehensive classification of models along with more granular data. Categories can be filtered through the extent of the model’s performance with the specific demographic information that users share during the registration process, such as their age, or in a certain guided category, such as questions related to health care.

Near the end of our conversation, it brings Gopta Artificial general intelligenceHuman-like algorithms-imminent technology. “These models are built for human users at the end of the day, at least in the near future,” he says. It is a fairly common belief, and the marketing point, among people working in artificial intelligence companies, although many researchers are still wondering whether the basic technology behind large language models will be able to produce AGI.

Gupta wants YUPP users, who may be concerned about the future of humanity, from the perception of themselves as being these algorithms actively and improving their quality. “It is better than free, because you are doing this wonderful thing for the future of artificial intelligence,” he says. “Now, some people want to know that, others just want the best answers.”

More users may only want to get additional money and are ready to spend a few hours to make notes during their Chatbot conversations. I mean, $ 50 is $ 50.