OpenAI’s Thibault Sottiaux Leads Core Products, Discusses ChatGPT Work, AI Agent Diffusion, and the Future of AI.

Thibault Sottiaux, known to users of OpenAI’s software engineering tool Codex for his role in resetting token limits during growth milestones, has stepped into a pivotal leadership position at the company. Sottiaux now leads all of OpenAI’s core products, a comprehensive portfolio that includes the API, agent infrastructure, enterprise solutions, the entire ChatGPT ecosystem (encompassing both ChatGPT Work and ChatGPT classic), and the foundational Codex platform. This expansive remit places him at the forefront of OpenAI’s efforts to democratize access to advanced artificial intelligence. His reporting line directly to Greg Brockman, OpenAI’s President, underscores the strategic importance of his role in the company’s organizational structure.

The introduction of ChatGPT Work marks a significant strategic pivot for OpenAI, aiming to extend the capabilities previously confined to technical users to a broader white-collar audience. Sottiaux elaborated on the vision behind ChatGPT Work, describing it as an endeavor to bring the potent functionality of coding agents to everyone. The challenge lies in transforming a sophisticated tool designed for developers into a user-friendly, safe, and delightful platform accessible on various devices, including mobile and web. This move is designed to make powerful AI agents available to the widest possible population, a goal reinforced by its inclusion in the $20-a-month Plus plan, emphasizing accessibility and perceived value.

Beyond the technological ambition, there is a clear economic imperative driving OpenAI’s strategy. Industry analysts suggest that owning the direct application relationship with the user is crucial for long-term growth and market position. Sottiaux echoed this sentiment, framing it as a symbiotic relationship: the more value and utility OpenAI generates for its users, the more willing those users will be to invest in the service. He cited ChatGPT’s current success as an example, where the immense utility provided compels users to willingly pay the monthly subscription fee, perceiving the value received as far exceeding the cost. This model is central to OpenAI’s financial sustainability and its ability to continue pushing the boundaries of AI research and development.

A critical aspect of OpenAI’s mission, as articulated by Sottiaux, is the broad diffusion of AI technology and winning over public opinion. He emphasized OpenAI’s responsibility in "bringing everyone along" with this transformative technology. While Codex initially served a "forgiving technical audience," allowing early capabilities to be tested and refined, the technology has now reached a maturity level that warrants wider deployment. ChatGPT Work is designed to educate a broader audience, demonstrating that AI can perform complex tasks autonomously, going far beyond basic writing assistance or personal advice. This strategic diffusion is fundamental to OpenAI’s overarching mission to ensure that the benefits of advanced AI are shared broadly and safely across society.

OpenAI’s product design philosophy is heavily influenced by the capabilities of its underlying models. Sottiaux explained that the core approach involves building "extremely capable models" and then devising the "most simple and delightful way" to integrate them into users’ lives, maximizing utility. This often means "getting out of the way" of the model, allowing its inherent capabilities to express themselves without unnecessary user interface complexities. The aim is a "minimal product surface" characterized by "delightful simplicity" and natural human engagement. The evolution from text-based interaction to the increasingly popular ChatGPT Voice exemplifies this philosophy, allowing for a more intuitive, conversational experience. The guiding principle is for the technology to adapt to humans, rather than requiring humans to learn complex application interfaces.

The question of market readiness for such "magic" AI, as posed by Wharton professor Ethan Mollick, who contrasts ChatGPT Work’s approach with Claude Cowork’s more explicit choices, was directly addressed by Sottiaux. He asserted that "the world seems to be ready," pointing to the "incredible adoption" and the recent milestone of 20 million users. This rapid growth, he argued, indicates that OpenAI has successfully launched a product that is both simple and powerful, uncompromising in its capabilities while maintaining ease of use. The widespread acceptance suggests a growing comfort and eagerness among professionals to integrate advanced AI agents into their workflows.

OpenAI’s product development is characterized by a philosophy of "discovery." Sottiaux explained that the company doesn’t strictly adhere to a traditional minimum viable product (MVP) approach when dealing with frontier AI. Instead, as OpenAI pushes the boundaries of model capabilities, they simultaneously "discover what it’s capable of." This process then informs which areas to lean into, allowing the company to build robust products around these newly uncovered strengths. He cited the development of GPT 5.6 as an example, which brought significant advancements in handling large volumes of documents, generating high-quality slides and reports, and conducting deep research—tasks crucial for professionals. This iterative deployment model emphasizes learning from the community, gathering real-world usage feedback, and continuously improving the models and products based on these insights.

The economic implications of AI usage, particularly the cost of intelligence, remain a significant consideration. Concerns about the potential gap between the Plus plan subscription and the actual token usage for advanced tasks were acknowledged. Sottiaux reassured that OpenAI is committed to "pushing the frontier on efficiency" daily. He referenced "Luna," an update that brought "80% off" price cuts, describing it as a "permanent price correction." This indicates a strategic commitment to making advanced AI capabilities more affordable over time. The long-term goal is to "include more utility in the same dollar amount," meaning users should expect to accomplish more with the same expenditure in the future, or the same tasks with less cost. This focus on efficiency and affordability is crucial for widespread, sustained adoption, particularly in enterprise settings where CFOs carefully monitor operational costs.

Finally, concerns about data privacy and security, particularly when granting AI agents access to sensitive information like emails or iMessages (a recently rolled out feature), are paramount. Sottiaux emphasized the critical importance of "picking models that are safe and aligned." He highlighted OpenAI’s substantial investment in its "safety stack" and overall "safety approach," which includes publishing "honest benchmarks" to demonstrate the robustness of their systems. He confidently stated that OpenAI’s models are "world-class at these topics," underscoring the company’s commitment to building trust and ensuring the responsible deployment of powerful AI agents that can interact with personal and professional data. The careful balance between expanding AI capabilities and maintaining stringent safety and privacy standards will be a defining factor in the success and public acceptance of platforms like ChatGPT Work.