Openai says it will provide a deep -year search form of $ 200 for free

It will make Openai to her Deep search form Available for users on the free layer of ChatGPT soon, according to A. X post By CEO Sam Al -Tamman. Deep search is the vernacular name for large language models that use the “test time” computing method to produce in -depth detailed research reports. These reports can take anywhere from 30 minutes to several hours to produce them, and the technicians said that deep research can replace humans in functions such as business analysis or financing. Of course, it is important to take these claims with a grain of salt.

Altman also said that Openai will unite the various artificial intelligence models. Many people have complained that ordinary users do not understand the difference between their different models and what they offer. “We hate the selection of models as much as they do and want to return to Magic Unified Intelligence. “We will then charge GPT-4.5, the model that Urion has called internally, as a model that is not studied. Next, the highest goal for us is to unify the chain models O and GPT-Seriz models of tasks.

At some point in the near future, users at the free level will be able to “unlimited” access to a new model that includes O3 compact O3, although Altman says paid users will get a higher level of intelligence, whatever that.

Currently, Deep Research is available only for Chatgpt users who pay the $ 200 per month. Openaii has always been expected to provide the edge of bleeding for more users as it is It increases the infrastructure of the serverAltman confirmed a lot in another conversation X post. Altman wrote, in a typical way to write everything in a small form like Gen Z Girl, “I think we will initially provide 10 monthly uses for Chatgpt Plus and 2 per month in the free level, with the intention of expanding its scope. Late, after it is too late.”

Big language models such as ChatGPT are, in essence, probability machines, or truly strong automatic completion. They accommodate billions of texts from the text written on humans and are able to generate new texts with Verisimiliade for something written by someone. However, they do not actually think as a person does, and companies like Openai had to find new technologies to improve their accuracy and capabilities. Thinking in the time of the test imposes the language model on verifying facts before answering a question. For example, if one user asks the model about the cost of replacing each Uber in the United States with WayMOS, the model will lead to the division of the query into multiple related questions – how many Ubers are on the road today, the amount .

The artificial intelligence industry was putting the spaghetti on the wall for some time now, in an attempt to find the abolition of the lock that will make the language models really capable of replacing humans in a variety of tasks. Deep search is the latest attempt like this. Another is the “agents” of artificial intelligence they can Control the user’s computer To actually complete the tasks, such as reserving a trip. Agents can be especially useful for population marks such as the elderly, who may not be smart in the field of technology and can ask artificial intelligence to help them move into a computer. The first users say to agents like an Openai operator they are very slow and error.

It is impressive like deep research in producing lengthy research reports, the accuracy of the artificial intelligence language models remain a problem. During Super Bowl, Google has a commercial broadcast promoted by Gemini Chatbot that included it Inaccurate information. X users have noticed whether you are asking Chatgpt about the players in the various American Football Association teams, Chatbot will return incorrect information although Chatgpt has access to lists from NFL.com. Of course, Apple was forced to neutralize the infringements based on artificial intelligence after production Incorrect summaries BBC News notification.

Some supporters of artificial intelligence argued that Probe Correct often means the text He is It is true, but it seems that a big problem in artificial intelligence models is that its users assume that everything they produce is true even when it is not. Google provides incorrect information over and over again during experimental presentations of artificial intelligence is a symbol of this problem. Leaders in technology giants seem to slide only and boil any of their artificial intelligence models without doing any work to check this.

Meanwhile, Openai deals with other problems, which is a long -term legal battle with Elon Musk, which escalated this week after early Openai manufactured hostile Offer 97 billion dollars to acquire the artificial intelligence company Under the consequences. Although Musk was able to control Twitter despite all the difficulties, this position is largely different, as Openai is special and is still not profitable. Although Openai is trying to convert into a profitable company, there is currently no credit duty for investors and they may not have to think about acquisition offers.