In a momentous gathering at Fort Mason Center in San Francisco, OpenAI CEO Sam Altman unveiled an ambitious vision for the future of artificial intelligence during the company’s third annual Dev Day. The event, set against the iconic backdrop of the Golden Gate Bridge, marked a significant turning point for OpenAI as it transitions from a provider of generative AI models to a comprehensive computing platform that aims to redefine how users interact with technology.
The central theme of Altman’s keynote was clear: the era of merely asking questions of AI is over. Instead, OpenAI is positioning itself to enable users to command AI systems to perform complex tasks, build software autonomously, and integrate seamlessly into every application. This shift represents a fundamental transformation in the relationship between humans and technology, moving from passive interaction to active collaboration.
One of the most notable announcements was the introduction of the new Apps SDK, which transforms ChatGPT into a dynamic, interactive platform akin to an operating system. This development allows developers to create and distribute applications directly within the ChatGPT interface. During the keynote, Altman expressed enthusiasm about this new capability, stating, “Today, we’re going to open up ChatGPT for developers to build real apps inside of ChatGPT. This will enable a new generation of apps that are interactive, adaptive, and personalized, that you can chat with.”
Live demonstrations showcased various applications from partners such as Coursera, Canva, and Zillow, all running smoothly within a chat conversation. For instance, a user could engage in a machine learning lecture, ask ChatGPT to clarify a concept in real-time, and then utilize Canva to generate a poster based on the discussion—all without leaving the chat interface. This integration not only enhances user experience but also provides developers with a powerful new distribution channel. Altman highlighted that by building with the Apps SDK, developers can reach hundreds of millions of users, tapping into a rapidly growing base that has surpassed 800 million weekly active users.
In a private press conference following the keynote, Nick Turley, head of ChatGPT, elaborated on the broader vision behind these developments. He emphasized that OpenAI never intended to create just a chatbot; rather, the goal was to build a super assistant. “We got a little sidetracked,” Turley admitted, acknowledging that while they succeeded in creating a great chatbot, not all software needs to be conversational. He underscored the necessity for the interface to evolve, which is why the demos featured rich user interfaces capable of full-screen experiences. Turley envisions a future where users might start their day in ChatGPT, which could become the de facto entry point into the commercial web and various software applications.
The introduction of the Agent Kit marks another significant advancement in OpenAI’s offerings. This integrated development environment enables developers to create autonomous AI agents capable of executing real-world tasks. Altman described the Agent Kit as a complete set of building blocks designed to help developers take agents from prototype to production. It features a visual canvas for designing complex workflows, an embeddable chat interface for deploying agents in any application, and a sophisticated evaluation suite to measure and enhance performance.
A compelling demonstration from financial operations platform Ramp illustrated the potential of the Agent Kit. In the demo, an employee simply typed, “I need five more ChatGPT business seats,” and the agent parsed the request, checked it against company expense policies, found vendor details, and prepared a virtual credit card for the purchase. What once took weeks to accomplish was completed in mere minutes, showcasing the transformative power of AI in automating complex business processes. Brad Lightcap, OpenAI’s COO, noted that this shift toward more agentic AI is essential for enterprise adoption, emphasizing the need for AI to do things for users rather than merely responding with text outputs.
Perhaps one of the most profound shifts discussed during the event was in the realm of software development itself. OpenAI’s Codex, an AI coding assistant, has transitioned from a research preview to a fully-fledged product powered by a specialized version of the new GPT-5 model. Codex is now capable of autonomously writing code, creating pull requests, and reviewing other engineers’ work on platforms like GitHub. A live demonstration showcased Codex transforming a simple photo of a whiteboard sketch into a fully functional, beautifully designed mobile app screen. Another demo revealed an app that could “self-evolve,” reprogramming itself in real-time based on natural language requests, offering a glimpse into the future of self-adapting software.
The day’s biggest surprise came during a closing fireside chat between Altman and Jony Ive, the legendary former chief design officer of Apple. Their collaboration, which has been ongoing for three years, aims to create a new family of AI-centric hardware. Ive articulated a vision for rethinking our relationship with technology, arguing that current devices are outdated and unfit for the AI era. He stated, “I think it would be absurd to assume that you could have technology that is this breathtaking, delivered to us through legacy products, products that are decades old.” This sentiment reflects a broader desire to address the overwhelm and despair many people feel in their interactions with technology.
While specific details about the new hardware remain under wraps, Ive emphasized the importance of care in the design process. “We love our species, and we want to be useful. We think that humanity deserves much better than what it generally receives,” he said, highlighting the need for thoughtful and empathetic design in the age of AI. This collaboration signals that OpenAI’s ambitions extend beyond cloud-based solutions; it is actively exploring the physical interfaces through which humanity will engage with its powerful new intelligence.
Underlying this entire platform strategy is a critical constraint: the availability of computing power. Throughout the day, OpenAI leaders reiterated the challenges posed by compute limitations. Altman remarked, “The degree to which we are all constrained by compute… Everyone is just so constrained on being able to offer the services at the scale required to get the revenue that at this point, we’re quite confident we can push it pretty far.” Despite massive new hardware partnerships with companies like AMD, Altman acknowledged that demand continues to outpace supply.
This reality explains OpenAI’s aggressive multi-billion-dollar investment in infrastructure. When questioned about profitability, Altman candidly stated that the company is currently in a phase of “investment and growth.” He paraphrased a famous quote from Walt Disney, saying, “We make more money so we can make more movies.” For OpenAI, the “movies” refer to increasingly powerful AI models that will drive future innovations.
Greg Brockman, OpenAI’s President, articulated the ultimate goal during the Developer State of the Union, asserting that AI is poised to become the fundamental driver of economic growth in the near future. He likened the question of how much compute one desires to asking how much workforce one wants, suggesting that the answer is always more. As the day concluded and developers mingled at the reception, the scale of OpenAI’s project became increasingly apparent. With new models like the powerful GPT-5 Pro and the stunning Sora 2 video generator, OpenAI is no longer merely building AI; it is constructing the ecosystem in which AI will thrive—a world filled with intelligent applications, autonomous agents, and innovative physical devices.
In summary, OpenAI’s Dev Day 2025 showcased a bold vision for the future of AI, emphasizing the transition from simple interactions to complex collaborations between humans and machines. With the introduction of the Apps SDK, the Agent Kit, and advancements in Codex, OpenAI is paving the way for a new era of intelligent applications and autonomous agents. Coupled with the promise of new hardware designed by Jony Ive, the company is poised to redefine our relationship with technology. However, the challenge of computing power remains a significant hurdle, necessitating continued investment and innovation in infrastructure. As OpenAI forges ahead, it is clear that the future of AI is not just about smarter models but about creating an integrated ecosystem where intelligence becomes the foundation of our digital lives.