OpenAI, the well-known start-up that popularized the ChatGPT, has announced that it is getting ready to release GPT-4, a model of sophisticated artificial intelligence. According to others, the aforementioned artificial intelligence will pave the way for the spread of human-like technology. Also, it is anticipated to intensify the rivalry between Google, owned by Alphabet Inc., and Microsoft Corp., its financial backer.
In a blog post, OpenAI asserted that their most recent artificial intelligence technology model is “multimodal,” which essentially means that it can be stimulated to produce material by both visuals and texts. The picture input functionality is still a sneak peek of its study, the text-input capability is apparently only going to be offered to ChatGPT Plus customers and software developers, with a waitlist.
On the other hand, Google, a subsidiary of Alphabet Inc., recently revealed a purported “magic wand” for its collaborative tools. The aforementioned magic wand is predicted to draft almost any paper. This statement was made just a few days before Microsoft is anticipated to present AI for its rival Word processor, which will probably be powered by OpenAI. Also, a Microsoft executive claimed that GPT-4 is a component of the Bing search engine.
GPT-4 Versus GPT 3.5
According to claims made by OpenAI, their most recent technological innovation, GPT 4, has significantly improved over GPT-3.5. According to a study, the new model supposedly performed in the top 10% of test-takers in a simulation of the bar exam that U.S. law school graduates must pass before engaging in professional practice, while the previous model allegedly performed in the bottom 10%.
The difference is apparently discernible when the task’s difficulty crosses a certain level, despite the two versions appearing the same in informal chats. In addition, the OpenAI considers GPT-4 to be more dependable, inventive, and capable of handling far more complex commands.
Also, Greg Brockman, the president of OpenAI, demonstrated the technology online and demonstrated how the new model could take a picture of a hand-drawn mock-up for a straightforward website and turn it into a functioning website.
The experiment supported reports that GPT-4 could aid people in calculating their taxes. However, OpenAI CEO Sam Altman asserted through Twitter that the GPT-4 model is “most capable and aligned” with human values and intent, despite “it still being faulty.”