Do you remember it? I yes because I had exclaimed: slutty misery! Seven days ago Sam Altman had compared GPT 5 to the atomic bomb, saying that the model was “so powerful as to make the wrists tremble”, that “perhaps it was appropriate to stop”, that “the governments should have intervened immediately, before the release”. Honestly, it seemed a little strange that the CEO of Openai asked for laws against his own product. Well, today, after a week, he released it. If it really is a bomb, she released her without being too much to think about it (so it’s not a bomb, at least not an atomic).
GPT 5 is now the model that feeds chatgpt (if you don’t find it, it is because it will be gradually released to users in the next few days) and it should not be activated: the system chooses it alone based on the type of request and the plan to which you are subscribed. Its versions (standard, mini, nano) will be distributed according to the resources available and the complexity of the interaction. Those who have a Plus subscription has wider access, although not unlimited, and there is a clear communication on the exact moment when the model is activated (when I write I still have the 4th). In any case, Openai ensures that the difference will be noticed immediately: the answers will be faster, more structured, with less smudges and more fluidity. As for Nick Turley safety, he informs us that the new model will not lie to users, he was trained to adapt to the contexts and the user “why should people be able to trust the system” (did not have to tremble our wrists?).
The main change, however, is not in the chat: GPT 5 was not only designed to respond, it was built to act, that is, to read files, analyze documents, automate flows, fill in, plan, anticipate. The “agent” mode is not yet available for consumer users, only in professional plans (200 euros subscriptions to understand each other), and will be released gradually. The “conversational interface” (as I use it, to talk to us, preferring it to most humans) becomes only the most superficial level, because the goal is to create an intelligent, adaptive, independent operating system and arrive before the others to a general artificial intelligence.
In addition, the release via API, as Gigi Ballarani had already anticipated me yesterday who follows every rumor by the way (and the noch and crypto rumors in general, and had foreseen “you will see that tomorrow they release chatgpt 5”) extends the use of the model to everything that can be connected: apps, bots, sites, automations, corporate software. Those who know how to program can integrate it immediately, those who do not know how to program can do it through tools that immensely simplify the process, to those who do not interest each other forget it. In practice, you no longer need to write code, you need to write prompt, and those who have something to build (an app, any software) can do it without going through a developer. It will not be the end of the programmers, however, it seems to me the beginning of a slow marginalization, especially those that do not design and simply translate commands. For technical details explained well also to non -professionals, I recommend, as usual, to follow Enkk on YouTube.
In short, morality of the fairy tale: chatgpt5 will surely be
Very cool, and Sam Altman, a week after invoking a law as if it were Sarah Connor, updated the system. This is the future: a rhetorical bomb every six days and then you return to play. It doesn’t help us anymore. For now.