AI

OpenAI Dev Day 2025: ChatGPT becomes the new app store — and hardware is coming

In a full room in the Mason Center Fort in San Francisco, against the background of the Golden Gate Bridge, Openi CEO Sam Altman presented a daring vision to re -shape the digital world. The company that has made generative AI with a simple chatbot mainstream, now lays the foundation for its next act: an extensive computer platform that is designed to go beyond the screen and the browser, where the legendary designer Jony Ive is called in to shape its physical form.

Be up Third annual Devdayhas unveiled a series of tools that signal a strategic shift from a model provider to a fully -fledged ecosystem. The message was clear: the era of simply asking questions to an AI is over. The future is about managing AI to perform complex tasks, to build autonomously software and live in every application, a transition that Altman described as a transition from “systems you can request everything, to systems that you can ask to do everything for you.”

The announcements of that day were a threefold attack on the status quo, aimed at the way users deal with software, how developers build these and how companies use intelligent agents. But it was the sessions that were kept behind closed doors away from the public live streamThat the true scope of OpenAi’s ambition revealed: a future that includes new hardware, a ruthless pursuit of computing power and a philosophical search to redefine our relationship with technology.

From chatbot to operating system: the new ‘App Store’

The center of the public keynote was the transformation of Chatgpt yourself. With the new Apps-SDKOpenAI is turning its wildly popular chatbot into a dynamic, interactive platform, in fact an operating system on which developers can build and distribute their own applications.

“Today we are going to open chatgpt for developers, so that they can build real apps within Chatgpt,” Altman announced under applause during the keynote presentation. “This will make a new generation of apps possible that are interactive, adaptive and personalized, with which you can chat.”

Live demonstrations showed apps from partners such as Coursera,, ” CanvaAnd Zillow Seamlessly performed within a chat conversation. A user can view and ask a machine learning lecture Chatgpt To explain a concept in real time and then use Canva to generate a poster based on the conversation, all without leaving the chat interface. The apps can display rich, interactive user interfaces and even go on full screen to offer a complete experience, such as exploring a Zillow card with houses.

See also  SoftBank-backed LegalOn lands $50M to streamline legal workflows with AI

For developers this means a powerful new distribution channel. “If you build the SDK apps, your apps can reach hundreds of millions of chat users,” said Altman, with which he emphasized a direct path to a huge user base that has grown into more than more than 800 million weekly active users.

During a closed press conference later, Nick Turley, head of Chatgpt, elaborated on the larger vision. “It was never our intention to build a chatbot,” he said. “When we wanted to make chatgpt, we wanted to build a super -assistant, but we got a bit on a sideways. And one of the tragedies of touching a bit on a sideline is that we have built a great chatbot, but we are the first to say that not all software has to be a chatbot, and not all the interaction with the commercial world.”

Turley emphasized that although OpenAi is enthusiastic about natural language interfaces, “the interface really has to evolve, and that is why you see so much user interface in the demos nowadays. You can even go on full screen and chat in the background.” He described a future in which users “could start your day in Chatgpt, only because it has de facto become the access point to the commercial web and to a lot of software”, but clarified that “our motivation is not to keep you inside. Our product is to build other people to build great companies and develop the form factor of software.”

The rise of the agents: building the ‘do all’ ai

If apps want to bring the world to chatgpt, the new “is”Kit“Is about sending AI into the world to get things done. OpenAi offers a complete” set of building blocks … to help you bring agents from prototype to production, “Altman explained in his keynote.

Kit is an integrated development environment for creating autonomous AI employees. It has a visual canvas to design complex workflows, an enclosed chat interface (“chat kit”) for deploying agents in every app, and an advanced evaluation package to measure and improve performance.

An exciting demo of the platform for financial transactions Driveway Showed how Agent Kit was used to build a purchasing agent. An employee can simply type: “I still need five Chatgpt-Business-Seats”, and the agent is parsing the request, compares it with the company’s expense policy, searches for supplier data and prepares a virtual credit card for the purchase-a process that once lasted for weeks and has now been completed in a few minutes.

See also  After India, OpenAI launches its affordable ChatGPT Go plan in Indonesia

This push to agents is a direct answer to the growing need of companies to go beyond AI as a simple tool for collecting information and in the direction of AI as a productivity engine that automates complex business processes. Brad Lightcap, COO of OpenAi, noted that for business adoption you needed this kind of shift to more agent AI that could actually do things for you, rather than just responding with text output. ”

The Future of Code and the Jony Ive Bbombshell

Perhaps the most in -depth shift takes place in the software development itself. CodexThe AI ​​coder agent from OpenAI has grown from a research review to a fully-fledged product, now driven by a specialized version of the new GPT-5 model. It is, as a speaker put it, “a teammate who understands your context.”

The possibilities are amazing. Developers can now assign Codex tasks directly from SlimAnd the agent can write autonomously code, make pull applications and even assess the work of other engineers Github. During a live demo, Codex showed that he took a simple photo of a whiteboard sketch and turned it into a fully functional, beautifully designed screen for a mobile app. Another demo showed an app that ‘could develop himself’, where he could reprogram himself in real time on the basis of the natural language request of a user.

But the biggest surprise of the day came during a closing conversation by the fire, which was not broadcast via live stream Altman and Jony IveThe iconic former Chief Design Officer of Apple. The two revealed that they have been working together on a new family of AI-oriented hardware for three years.

Ive, whose design philosophy shaped the iPhone, iMac and Apple Watch, said that the goal of his creative team “became clear” with the launch of Chatgpt. He argued that our current relationship with technology has been broken and that AI offers a chance for a fundamental reset.

“I think it would be absurd to assume that you could have technology that is so breathtaking, delivered to us through older products, products that are decades old,” Ive said. “I see it as an opportunity to use this remarkable capacity to fully tackle a large part of the overdoors and despair those people now feel.”

Although the details of the devices remain secret, IVE spoke about his motivation in deep -human terms. “We love our kind and we want to be useful. We think that humanity deserves much better than humanity in general is given,” he said. He emphasized the importance of “care” in the design process and said, “We feel when people care about us … You feel carelessness. You feel when someone doesn’t care, they care about money and planning.”

See also  CES 2026: Follow live for the best, weirdest, most interesting tech as this robot and AI-heavy event wraps up

This collaboration confirms that OpenAi’s ambitions are not limited to the cloud; It actively investigates the physical interface that will communicate humanity with its powerful new intelligence.

The unarperable thirst for computer use

There is a single, overwhelming limitation at the basis of this entire platform strategy: the availability of computing power. Both during the private press conference and during the non-ramed developer State of the Union, the leadership of OpenAi returned to this theme time and time again.

“The extent to which we are all limited by computer use … Everyone is so limited in being able to offer the services on the scale needed to generate the income, that we are at the moment pretty sure that we can drive it pretty far,” Altman told reporters. He added that even with huge new hardware partnerships with AMD and others, “we will say the same thing again. We are so convinced of it … there is so much more demand.”

This explains the aggressive, Billion dollar investments in infrastructure. When he was asked about profitability, Altman was frank that the company is in a phase of ‘investments and growth’. He relied on a famous quote from Walt Disney, in which he paraphrased: “We earn more money so that we can make more films.” For OpenAI, the ‘films’ are increasingly powerful AI models.

Greg Brockman, the president of OpenAi, formulated the ultimate goal in grim economic terms during the developer State of the Union. “Ai will, probably in the not too distant future, become the fundamental engine of economic growth,” he said. “Questions:” How much computing power do you want? ” Is a bit the same as asking how much staff you want?

When the day ended and the developers gathered at the reception, the size of the OpenAI project became clear. Powered by new models such as the powerful GPT-5 Pro and the amazing Sura 2 Video gerator no longer only builds the company on AI. It builds on the world in which AI will live – a world of intelligent apps, autonomous agents and new physical devices, where we assume that intelligence itself will be the ultimate platform in the near future.

Source link

Back to top button