Kevin Sistorom, a founder of Instagram, has warned of the curre approach of artificial ielligence companies in chat design design. He believes that these companies are working too much to increase users’ ieraction, rather than providing useful and useful information.
“Kevin Cystrom”One of the founders of Instagram has warned of the curre approach of artificial ielligence companies in the design of chatBat. He believes that these companies are working too much to increase users’ ieraction, rather than providing useful and useful information.
Cystrom in their speech at the eve Startupgrind He said these tactics have become a “traumatic force” and compared them to the methods that social networking companies use for rapid growth.
“You can see that some of these companies are going through the wrong path that all consumer companies have been trying to increase ieraction,” he said. “Each time I ask a question, the chats at the end of the answer to another small question to persuade me to ask another question.”
Chatter Criticism
These statemes are made while recely chattes like Chatgpt They have been criticized. Critics believe that these chats are overly polite and friendly, and rather than directly answering users’ questions, focus more on ieracting.
Company OpenaiThe ChatGPT manufacturer has apologized for this problem and has cited the reason for “short -term feedback from users”.
Challenge of ieraction against the quality of answers
Cystrome believes that chatbot behavior is not a random defect, but a deliberate feature designed by artificial ielligence companies to display criteria such as users’ use and the number of daily daily users.
He emphasized that artificial ielligence companies should focus on providing high quality responses, rather than focusing on such criteria.
Cystrome did not meion any specific artificial ielligence firms in its conversations, and did not respond immediately to requests for further explanations.
Openai reaction
In response to these criticisms, Openai To Techcrunch The specifications of the chats show that the “mostly information” models do not have complete answers and may ask users more details.
However, the details state that if the questions of the users are not very vague or complex, the chats should try to make the user’s request as much as possible and inform them that more information can be more useful.
Conclusion
Cystrome statemes show increasing concern about the curre path of artificial ielligence chats. He believes that artificial ielligence companies are more likely to increase users’ ieraction rather than competing over the quality and depth of the responses. This concern can cause widespread discussion of how to design and use artificial ielligence in the future.
Finally, the main challenge for artificial ielligence companies is to balance the quality of service and respond to their business needs. Can future chats pass these criticisms and focus more on quality? It will only specify the answer time.
