AI’s third era: the rise of persistent AI coworkers | Tara Seshan (Product Lead ChatGPT Work)

Tara Seshan discusses the progression of AI from simple chat interfaces to more advanced agents that work alongside humans as persistent coworkers. She emphasizes the shift from 'rowing,' where humans execute tasks, to 'steering,' where human judgment guides AI-driven processes. The conversation explores how AI models are rapidly improving, giving rise to AI agents that can carry out tasks over extended periods. Tara underscores the importance of human oversight, ambition, and creativity in leveraging these tools effectively, especially as AI capabilities are evolving faster than predictions.

Tara shares her impressions of OpenAI's distinct culture, highlighting the 'founder's mentality' present within teams despite the company's non-typical openness about internal strategies. She notes that the emphasis is on rapid, iterative experimentation over long, theoretical planning, with a focus on quickly bringing products to market and learning from user feedback. Tara also discusses the company's alignment with cutting-edge research, aiming to develop models two to three months ahead of current capabilities, thus maintaining a forward-looking product strategy.

A core theme is how AI is transforming the role of product managers and organizational ambitions. Tara stresses the importance of rapid testing, hypothesis-driven development, and elevating team ambitions. She states that the traditional 'trappings' of product management—detailed docs and linear planning—are giving way to leaner, faster cycles driven by experiments and real user data. She highlights the necessity of maintaining high ambition levels, encouraging teams to think bigger and move faster, facilitated by AI's expanding possibilities.

Tara mentions internal memes like 'Is this maximally accelerated?' and 'Are you mainlining it yet?' as cultural markers that stress speed and intense AI integration. She emphasizes the significance of continuous iteration, rapid experimentation, and hands-on engagement with AI tools. Tara advocates for a mindset where shipping early and often, even with imperfect products, leads to faster learning and innovation, reflecting the culture of relentless ambition and execution.

The discussion centers on how OpenAI constructs products to align with the anticipated capabilities of models in two to three months, rather than current limitations. Tara emphasizes constant communication with research to understand model roadmaps and the need for product designs that remain adaptable to rapidly improving models. She highlights the creative and strategic importance of hypothesis testing, rapid prototyping, and updating assumptions based on user feedback and model progress, enabling AI products to stay relevant and effective.

Tara envisions a future where humans work more as strategists and steerspeople, guiding AI agents that handle tactical execution. She discusses the evolution of interfaces from solo interactions to collaborative workflows involving group AI agents, potentially in shared workspaces like chat platforms. She also highlights how AI will extend human capabilities, allowing for more ambitious projects, creative expression, and personalized software. The key, she argues, remains human judgment, context, and the ability to steer AI systems effectively.

Tara delves into her approach to writing as a tool for thinking and reporting. She advocates for deep personal involvement in initial writing to preserve clarity and critical thinking, while leveraging AI for summaries, translations, and drafting. She emphasizes the importance of maintaining human-led thought processes, resisting over-reliance on AI for creative or strategic work, and iterating on drafts with human feedback to ensure quality and purpose.

Drawing from her stint at Sutter Hill, Tara highlights the importance of product marketing fit—how a product is described, positioned, and understood in the market—as foundational to product success. She notes that successful companies often find that narrative and positioning should precede detailed product features, guiding development efforts. Her key insight is that storytelling and early market testing via pitching can plant the seeds for more successful product-market fit later.

Tara emphasizes that in an era of accessible AI tools, the limiting factor is often human ambition and imagination. She suggests that the most effective teams are those that push beyond automation and leverage AI to expand their scope of possibilities. Ambition now acts as a competitive moat—those who dare to think bigger, move faster, and integrate AI deeply into their processes will differentiate themselves.