Last November, the company announced GPT-4 Turbo and said this model was even more capable, with an “updated knowledge limit through April 2023” and that it can analyze the equivalent of 300 pages of text in a single question.
Technology company OpenAI, creator of the popular artificial intelligence (AI) chatbot ChatGPT, could launch GPT-5, its improved generative AI tool, by mid-year, according to Business Insider.
The newspaper noted this Wednesday that OpenAI is still training its GPT-5 chatbot, but that some enterprise clients of the company led by Sam Altman recently received demonstrations of the latest model.
“It's really good, significantly better,” a CEO who wanted to remain anonymous and who recently saw a version of GPT-5 told Business Insider.
Before releasing the new model to the public, the AI company will conduct an internal security test and form a “red team,” a process in which employees and a selection of external people will challenge the tool in different ways to find problems. before making it available to everyone.
Process in which, for example, its hyper-realistic video creation tool with Sora AI is now.
A year ago, OpenAI launched GPT-4, the most powerful version released by OpenAI technology and which powers the AI of its partner, Microsoft.
This version, capable of responding to both images and text, was presented almost four months after the company launched ChatGPT, a model powered by GPT-3.5 that had also achieved great popularity, as a more advanced and paid version of the which can be accessed by a ChatGPT Plus subscription for $20 per month.
Last November, the company announced GPT-4 Turbo and said this model was even more capable, with an “updated knowledge limit through April 2023” and that it can analyze the equivalent of 300 pages of text in a single question.
In March 2023, a group of businessmen from the technology sector, experts and politicians requested in an apocalyptic letter that experiments with very powerful artificial intelligence be suspended for six months because they consider that “they may pose profound risks for society and humanity.” ”.
“We call on all AI laboratories to immediately pause for at least six months the training of AI systems more powerful than GPT-4,” said the statement, which called for governments to intervene to impose this moratorium. in the event that the parties involved refuse to temporarily suspend the investigations.
One of the biggest competitors to OpenAI chatbots now is Google and its Gemini generative AI.
For its part, Anthropic, the artificial intelligence company founded by several former OpenAI employees, said this Monday that the different versions of its new model, Claude 3, work as well or better than the GPT-4 and Gemini Ultra models, the from the market leaders, OpenAI and Google, respectively.
No comments:
Post a Comment