- Hindi news
- Business
- Sam Altman Compares Chatgpt 5 to Manhattan Project, Feels ‘Useless’ during Testing New AI Risk Concerns “
Washington8 minutes ago
- Copy link
Sam Altman said that India is the second largest market in the world after America.
Openai CEO Sam Altman has compared his company’s new AI model Chatgpt-5 to the Manhattan project. The first atomic bomb was made in this project during World War II.
In a podcast, he also said that while testing the Chatgpt-5, he felt as if he himself had become “useless”. This matter of Sam Altman has launched a new debate about AI’s possible threats.
Altman is planning to come to India in September
Sam Altman may also come to India in September. He said- India is the second largest market in the world after America and it can soon become our largest market. It is growing very fast, but the way people of India are using AI, it is really amazing.
Sam Altman compared the Chatgpt-5 to the Manhattan project in the podcast of YouTuber Theo Wayne.
Why compare with Manhattan Project?
Sam comparing the CHATGPT-5 to the Manhattan project said that as that project had changed history by making atomic bombs, this new model of AI can also completely change the world.
In the Manhattan project, scientists later realized how dangerous their technology could be. Sam said, “We have made something that we do not fully know about what will be the effect. Perhaps it is very good, perhaps bad, but the question is’ what have we done?”
He expressed concern that AI’s speed is so fast that the rules and regulations of the world are not getting steps with it. Sam said, “It seems as if there is no elder in this room.” That is, this race of AI has a huge lack of monitoring and control.
GPT-5 is a moment for Sam Altman that is very big. This is not just a jump in computing power, but this is the occasion when we need to think on the direction of human innovation. We have made a tool like God, but without any moral direction.
Sam Altman launched Chatgpt-5 yesterday.
GPT-5 solved the question that could not Sam
Sam Altman said in ‘This Past Weekend With Theo Von’ Podcast that during the testing of Chatgpt-5, there was an incident that surprised him. He tried to find an answer to a complex question, but he himself could not find a solution.
Then he asked the same question to Chatgpt-5 and the model gave an accurate answer in the blink of an eye. Sam said, “I sat back on the chair and thought, ‘Hey, what happened!’ I started feeling useless in front of AI.
GPT-4 is quite weak in front of Chatgpt-5
Sam also said that GPT-4 was considered the most advanced model ever, but it seems “quite weak” in front of CHATGPT-5. GPT-5 has multi-step reasoning, long memory and multimodal abilities (text, image, and data processing) so much better that it leaves old models far behind.
The specialties of Chatgpt-5
Sam wrote in a post on X, “GPT-5 is the smartest model, which we have created so far. Our focus is on more and more people and its real-world use.”
- This answers the complexity of the question and answers. If the question is easy, it answers immediately, and if the question is difficult, the “thinking” gives the accurate and detailed answer.
- These complex coding can easily handle tasks, such as fixing bugs, debug with large codebase, or making small games from a single prompt.
- It can also help in writing things like emails, reports, stories, or speech. It handles complex writing structures better and gives natural, creative writing.
Chatgpt-5 responded to 100% Acuraxi in Howard’s Mathematics tournament.
- There was a problem in the old models that they sometimes used to give wrong or fabricated answers, which in the language of AI calls “Helication”. This problem has been reduced to a great extent in GPT-5.
- The GPT-5 has a new technique called “safe completes”, which securely gives answers to risky questions, completely rejecting the question.
Models like Chatgpt-5 may be misused
A large part of Sam’s concerns is that models like Chatgpt-5 may be misused. Sam said that if this technique is not controlled properly, it can damage the social structure, break the trust, and challenge the specialty of humans.
Getting Artificial General Intelligence next step
Sam made it clear that Openai’s long -term goal is to achieve Artificial General Intelligence (Agi), that is, AI that can do all kinds of work like humans. But at the same time, he also admitted that if AGI is not managed correctly, then it can be dangerous. Sam said, “We don’t know where this technology will take us. It can be fantastic, but its risk is equally big.”
Free users can also use Chatgpt-5
The CHATGPT-5 has been launched on 7 August 2025 and is being rolled out for free, plus, pro, and team users. Next week it will also be available for enterprise and education users. Sam also said that this model is so cheap that more than one billion people can take advantage of it.
