At the recent developer conference in San Francisco, OpenAI revealed a major update for ChatGPT, indicating a strategic shift towards transforming it into a more interactive and integrated platform. This update allows third-party applications to embed directly into ChatGPT, aiming to create a new generation of adaptive, interactive, and personalized apps that users can converse with. OpenAI’s CEO, Sam Altman, introduced a preview version of a new software development kit (SDK) that developers can use to build these applications using open standards. Initially, this capability will be available to a select group of major partners.
During the demo, Altman showcased integrations with popular services like Spotify, Canva, and Zillow, illustrating how these apps can operate within the ChatGPT interface. For instance, Alexi Christakis, an OpenAI software engineer, demonstrated the capabilities of the embedded Canva app by requesting the creation of marketing materials for a hypothetical dog-walking business and then exploring expansion opportunities by engaging with a Zillow app integrated into ChatGPT to search for homes in a specific city.
Altman also highlighted forthcoming features intended to monetize the ChatGPT apps, including a commerce protocol that would allow transactions directly through the chat interface. This marks a significant advancement from OpenAI’s prior efforts, recalling the introduction of custom widgets or GPTs at a past developer conference, which did not achieve widespread success.
The strategic pivot reflects OpenAI’s ambition to position ChatGPT not merely as a chatbot but as a comprehensive digital hub or a “super assistant” as described by Nick Turley, OpenAI’s head of product for ChatGPT. This vision represents a blend of web and mobile app experiences, directing users towards a singular, chat-driven operating system that encourages staying within the ChatGPT ecosystem.
Furthermore, OpenAI plans to expand its offerings by introducing tools for building AI-driven agents through platforms like AgentKit, which provides a user-friendly interface for constructing advanced AI tools. These developments are complemented by enhancements to OpenAI’s coding tools, notably Codex, which has moved out of research preview to become generally available. New functionalities will facilitate interaction with code through platforms such as Slack, and provide SDK support and analytics tools to monitor usage.
In the competitive landscape of AI development, OpenAI is not only contending with giants such as Google, Amazon, and Anthropic but also facing challenges from entities like Meta and DeepSeek, which are offering capable open-source models. In response, OpenAI has also renewed its commitment to the open-source community by releasing GPT-OSS and introducing other flagship models like GPT-5, alongside upgrades to its developer-oriented coding models.
In addition to software developments, OpenAI’s growing infrastructure demands have led to significant investments in hardware, evidenced by a substantial deal to acquire AMD chips. This acquisition might also involve OpenAI taking a stake in AMD, reflecting the intensive compute requirements of OpenAI’s technologies and its plans for expansion.
Through these initiatives, OpenAI is reshaping the landscape for developers and positioning itself at the forefront of a movement to redefine human-computer interaction. By integrating third-party applications directly into ChatGPT and improving the capabilities of its AI models, OpenAI aims to enhance productivity and create a more seamless user experience. This suite of tools and integrations indicates that OpenAI is committed to advancing the field of AI and maintaining its position as a leader in innovation.
Sam Altman’s comments during the keynote emphasized this vision, asserting that it’s an unprecedented time to be a developer given the speed at which ideas can now evolve into products. This sentiment underscores the transformative potential of the technologies OpenAI is developing, promising a future where AI not only assists with tasks but potentially reshapes entire workflows and industries.
Read the full post on wired.com
