After delaying the launch of GPT-5 several times, Sam Altman can already say that the time has come to further revolutionize the technology sector in general and that of artificial intelligence in particular, with a model that, paradoxically, does not achieve Openai’s goal.
«GPT-5 is here», The company pointed out in a shared statement this Thursday, where it recalled that it is the smartest, fast and useful model so far with integrated thought, which is now available for all users.
According to the developer, this model is smarter than the rest «in all aspects», to the point that es comparable to «have a team of guard experts for anything you need to know» on issues such as mathematics, science, finance or law, among others.
In this way, GPT-5 has already been put in the hands of users (both for those who use the free version and for those who have a payment account) with different versions of use, ranging from the standard model, for less specific tasks, up to GPT-5-Chath, intended for conversations with a natural and multimodal style. In between, there are GPT-5-mini and GPT-5-Nano.
With him, therefore, the firm led by Altman seeks to achieve the known as General Artificial Intelligence (AGI), that is, an advanced form of artificial intelligence that has the ability to understand, learn and solve tasks in a similar or even superior way to human intelligence.
GPT-5 does not reach the agi level
At the moment, and despite the fact that he had promised, GPT-5 cannot be defined as AGI, according to Altman himself during the presentation of the model, in a meeting with journalists in which he has been present Business Insider.
«This is clearly an intelligent model in general, although I believe that, as most of us define the AGI, We still have something very important … or many very important things «, He has admitted the manager at this conference.
Altman, in fact, has been something more concrete in these statements, where he has recognized that «one of the important points is that It is not a model that continually learns as it is implemented. »
«[Lo hace] From the novelties I discover, something that for me resembles the general AI. But, the level of intelligence, capacity, is perceived as a great improvement, «has pointed out before the media, a laudable gesture at a historical moment in which everyone believes to have the definitive formula of success.
OpenAI advances in the right direction
While Altman is aware that it is still very soon to achieve the levels required by general artificial intelligence – without going further Yann Lecun, head scientist of the Meta, believes that we are «decades» to be able to achieve it – he also knows that they are going in good direction.
«If I could go back five years before GPT-3 and tell me that we have this now, I would say that it is a significant fraction of the road to something very similar to the general AI,» he told journalists, without losing the illusion for finding the key with this super intelligence.
«Superintelligent tools could greatly accelerate scientific discovery and innovation far beyond what we are able to do on our own and, in turn, greatly increase abundance and prosperity,» he concluded.
Know How we work in NoticiasVE.
Tags: Artificial intelligence