"People don't think in apps -- they think in flows. Tools should adapt to that." -- Qi

An interview with Qi -- cognitive science researcher and first-time founder -- on AI, focus, and building at the intersection of many worlds.

---

🧠 From Academia to Accidental Breakthroughs

When most founders talk about AI assistants, the conversation quickly devolves into a checklist of features: meeting summaries, code suggestions, note-taking bots. But Qi's journey into AI started in a place that rarely produces startup founders -- nearly a decade inside a PhD program in cognitive science, a discipline that forces you to understand how humans interpret information, structure ideas, and infer patterns from limited data.

His early research explored how people perceive style in written characters. At one point, he built systems capable of generating complete typefaces from only 10--20 examples -- not to design fonts, but to study how humans abstract structure from fragments. That skill would later become the backbone of his approach to machine learning and product design.

But the turning point came during an internship in Japan. He was tasked with building a Japanese ASR (Automatic Speech Recognition) model from scratch -- despite never having worked with speech, audio, or phonetics before. With the methodical mindset of a cognitive scientist, he broke the problem down and rebuilt it piece by piece.

The result was astonishing: his model eventually became the most widely used open-source Japanese ASR system built entirely from scratch.

"That moment showed me I could enter a completely new field and still produce something meaningful, as long as I understood the patterns beneath it."

---

🧩 Why Today's AI Assistants Still Get Work Wrong

As AI tools matured through 2023--2024, Qi noticed something that felt obvious yet rarely discussed: modern work is built on constant context-switching. Developers bounce between meetings, IDEs, documentation, tutorials, Slack channels, dashboards, and browser tabs. Every transition breaks concentration.

"AI tools today are designed for individual apps," Qi explains. "But humans don't think that way."

He became increasingly aware of how intrusive meeting bots were. Users described them as uncomfortable, even surveillance-like.

"Meeting bots change the atmosphere instantly. People don't like being recorded or observed. The tool becomes the distraction."

Instead of pushing AI deeper into meetings, Qi and his co-founder experimented with a quieter idea: an ambient assistant that sits alongside your entire workflow, understanding transitions, maintaining continuity, and reducing friction without inserting itself into calls.

The insight came from watching real users -- and from observing his own unusual way of working.

---

🤖 The Founder Who Works With AI Eight Hours a Day

Qi doesn't talk about AI the way most people do. For him, AI isn't a chatbot. It isn't a productivity tool. It isn't a notetaker.

It's a colleague.

"Claude is basically my coworker. I talk to it for eight hours a day."

He uses Claude the way some founders use a whiteboard or a long walk: thinking out loud, structuring ideas, reflecting on decisions, and testing assumptions. He feeds it long-term context. He relies on it for planning. He uses it to reason through uncertainties.

His usage was so heavy that he repeatedly hit Claude's four-hour limit and eventually upgraded to Claude Max simply to keep the conversation going. This wasn't productivity hacking -- it was necessity.

Qi's workflow represents a shift many founders are just beginning to experience: AI not as a tool, but as a second cognitive stack.

---

🚀 Building Toward an AI That Understands the Whole Day

Qi's prototype for an ambient AI assistant emerged naturally from his own habits. Instead of bots that intrude into specific tasks, he imagines AI that quietly supports the entire workflow.

Today's assistants are fragmented because they mirror the apps they live inside. Qi believes the next generation will be defined not by features but by continuity -- the AI that understands and preserves human flow.

"People don't think in apps -- they think in flows."

It's a simple statement, but one that reframes the entire landscape of AI productivity tools.

---

🧭 Advice for Young Founders

After moving from research to startup building, Qi learned hard lessons about speed, clarity, and humility.

He laughs when describing his early product decisions.

"First-time founders often build what they find interesting, not what the market needs. I did that too."

His advice is grounded, not glamorous:

• Talk to users earlier than you think.
• Ship before you feel ready.
• Don't fall in love with the prototype.
• Let the problem teach you what to build.
• Find a pace you can sustain for years, not weeks.

And most importantly:

"Be patient with yourself. Building clarity takes time."

---

🔚 Closing Thoughts

Qi's story doesn't fit neatly into typical founder archetypes. He isn't the stereotypical hacker. He isn't the polished operator. He isn't the Silicon Valley wunderkind. He's something more interesting: a cognitive scientist who builds, experiments, and collaborates with AI in a way that feels like a preview of the future.

From creating Japan's top open-source ASR model without prior experience, to working side-by-side with Claude for hours a day, to designing AI tools that aim to reduce -- not increase -- cognitive load, Qi is exploring what it means to build at the intersection of research, intuition, and deep curiosity.

"The pace matters. Not too fast. Not too slow. Just enough to keep learning and building."

It's a philosophy shaped by a decade of thinking about thinking -- and one that may shape the tools many of us use in the years ahead.

---