Subscribe to access the summary in 20 seconds.
Was it only sufficiency? Unfinished technical developments? A combination of strategy errors? Or even precipitation? Time will probably say it. In the meantime, Openai faces an important storm. The reason: the launch of its latest language model, GPT-5, last Thursday. This one, eagerly awaited by chatgpt lovers, and sketched for months by Openai, has disappointed many users. The firm led by Sam Altman also surprised by strategic choices deemed doubtful, a choice which it has precipitously had to correct in the past few hours.
And yet, Sam Altman, director of Openai, had presented GPT-5 with emphasis, making his presentation look like that of an Apple product. The model is, according to the manager, the “smarter”, the “faster” and the “most useful” to date, capable of doing things “impressive for you. It can instantly create software, on demand […]. It is an incredible superpower. ” Sam Altman made this comparison: GPT-3, at the base of the version of Chatgpt released in November 2022, was “a high school student who, sometimes, gives the right answer and, sometimes, responds anything”, then GPT-4 was “a university student”. Now interacting with GPT-5, “it’s like talking to a doctoral level expert on any subject,” said Sam Altman on Thursday.