Speak Founders on AI, Personalized Education, and Global Expansion

Korean learners often reach a plateau in their English fluency due to a deeply ingrained cultural fear of making errors, a pattern co-founder Connor Zwick notes is pervasive within the education system. This learned fear significantly impedes progress, preventing individuals from achieving conversational proficiency.

Speak's core product design is specifically crafted to counteract this phenomenon, promoting an environment where mistakes are viewed as part of the learning process rather than something to be avoided. This approach has proven universally effective across diverse cultures, including Japan and other international markets where similar educational anxieties exist, allowing learners to overcome psychological barriers to fluency.

Traditional classroom settings frequently struggle to accommodate the unique learning styles of individual students, often leading to disengagement and underperformance. Andrew Hsu, co-founder of Speak, recalled his own childhood experiences where he often struggled with conventional rules, finding more success when given autonomy.

Hsu credits a specific fifth-grade teacher who recognized his need for a different approach, providing him with a textbook and allowing him to learn independently. This experience underscores how educators who empower students with autonomy can significantly foster better educational outcomes, demonstrating the transformative potential of high-quality, personalized education on a student's development.

Andrew Hsu, who entered college at the age of 12 and earned his PhD by 16, credits his accelerated learning path to unconventional educational experiences that fostered a fundamental belief in human potential. He believes everyone is capable of learning significantly faster than they perceive possible, a philosophy central to Speak's mission.

With advancements in AI models, the company aims to democratize high-quality, personalized education globally, making such accelerated learning accessible to a much broader population. The ultimate goal is to leverage AI to make humans stronger and better, unlocking individual potential by providing tailored, efficient learning experiences.

This finding highlighted the immense efficacy of personalized instruction.

For over four decades, the economic impracticality of providing individual tutors for every student rendered this ideal unattainable on a large scale. The cost and logistical challenges made widespread implementation impossible, confining its benefits to a privileged few.

However, the advent of advanced AI technologies is now transforming this dynamic, making the vision of universally accessible one-on-one education not just feasible but inevitable. AI-driven tutoring systems can emulate the personalized attention and adaptive teaching methods of human tutors, finally allowing Bloom's findings to be applied globally.

He cites Google's rise alongside the widespread adoption of the internet as a prime example of this phenomenon.

When a foundational technological shift occurs, existing market leaders often become less significant, as their established structures are not built for the new paradigm. Hsu views the current AI revolution as the greatest technology wave to impact education, presenting an unparalleled opportunity to innovate and build an incredible company that redefines learning.

The company's vision, as articulated by Connor Zwick, is to fundamentally reinvent how people learn, beginning with language acquisition. This initial focus serves as a strategic starting point to refine and prove their innovative approach.

Ultimately, Speak aims to expand its product and mission to encompass and support any form of learning, leveraging smart tutoring systems to generalize its effective methodologies across diverse subjects and skills. This ambitious long-term goal seeks to make personalized, efficient education accessible for a wide array of educational needs.

Sam Altman, an early investor in Speak, contacted Connor Zwick with an urgent request to secretly test a groundbreaking speech recognition model developed by OpenAI. This collaboration marked a significant turning point for Speak's technological capabilities, as it addressed some of their most challenging speech recognition problems.

The new model from OpenAI dramatically reduced error rates in speech processing, enabling a more accurate and fluid interaction within the Speak app. This breakthrough paved the way for creating more open-ended and dynamic AI experiences, allowing the platform to offer sophisticated conversational learning that was previously unattainable.

These advanced AI tutors will transcend simple instruction, taking ownership of and guiding a user's entire learning journey from start to finish.

Such sophisticated systems will be designed to demonstrably help learners achieve their educational goals as efficiently as possible, leveraging real-time data and agentic intelligence. This approach fundamentally differentiates a purpose-built AI tutor, which is designed for a specific educational outcome, from a raw intelligence model that simply processes information.

This raw intelligence can perform numerous tasks, but it requires a structured system to effectively guide and direct it.

Speak's innovation lies in building a comprehensive system around these foundational models, effectively transforming their raw intelligence into a purposeful and effective tutoring experience. This bespoke system is designed to direct the AI's capabilities toward specific learning objectives.

Purpose-built solutions consistently outperform general-purpose AI models in specialized applications, as they are engineered with specific outcomes and user interactions in mind. Speak's focus is on developing scaffolded systems that precisely apply the AI's intelligence, creating effective and tailored learning results for users.

This mission helps steer the company away from the temptation of focusing on short-term gains, ensuring alignment with long-term objectives.

The early days of Speak were marked by existential challenges and difficult decisions, testing the founders' resolve. However, their unwavering conviction in the company's vision and mission enabled them to adhere to their core values, making compounding decisions that ultimately contributed to their sustained success.

Connor Zwick articulates that operating a successful company is essentially a continuous sequence of making numerous strategic decisions. The ultimate success of a startup hinges on the cumulative percentage of these decisions that prove to be correct.

This process demands constant truth-seeking and a commitment to adaptation, allowing the company to evolve in response to new information and changing circumstances. By consistently making sound judgments and accurately predicting the long-term trajectory of the world, a company significantly increases its chances of achieving lasting success.