The Origin of Hypandra

From Research to Realization

What is search? What is a search engine? What is a search query? Why do people search the way they do? Does Google shape what we know? ^[1]

At UC Berkeley, my research focused on how people search the web and how search engines shape their ways of wondering, learning, and knowing. These systems had grown so influential that they determined not only what people found, but also what they asked, wrote, and believed. Over time, these tools and habits made it harder for people to act with genuine curiosity. Search engines treated questions as raw inputs to be cut up and packaged for instant answers—not as beginnings worth nurturing.^[2]

ChatGPT was released in the final weeks of my dissertation.^[3] Then the wider wave of AI tools. Once again, I watched attempts to reimagine search. Yet even in these new systems, the same patterns solidified: questions were flattened, curiosity extinguished, and efficiency prized over exploration.^[4] At Michigan State I taught students how they could play a role in shaping search design and practices. I wrote about my explorations with the new tools, flagging early examples of hallucinated outputs in general search tools and asking how we were making decisions about the design of these new systems.^[5]

My time at a small Y Combinator startup, Trieve (now the search infrastructure at Mintlify), gave me a close look at how developers were redefining possibility in search and retrieval. I worked to sell their new search offerings and show developers how it works. I practiced writing for a different audience and building many prototypes. I was part of a small team daring to build something new to help people, and they pushed me to grow my own technical skill set and a startup mindset. Eventually the team pivoted and we parted ways, but I left with a sharper conviction: I could build something too.

Protect and promote curiosity

What questions do you ask when holding a newborn? Who are we becoming? Who do we want to be? What questions are worth asking?

Our first was born a few weeks before my qualifying exams. Our second arrived in the final months of dissertation writing. This past spring, I found myself once again in the maternity ward. I was rocking our third child late at night. My wife was sleeping. Machines beeping. I wondered about the world unfolding before her. What questions would she ask? How would I nourish her curiosity? How could I better support the wild questioning of her older brothers? What questions should I be asking now?

What was I going to do? Where could I make a difference? I thought again about how desperately we need to reshape how we ask questions and search for answers. So many attempts to reimagine search had only reinforced the same old patterns: quick questions and quick answers. What if we could do better?

That night, I realized I didn’t just need to study these tools or evaluate their performance. I could—and should—build something to directly challenge the accepted standard of how search tools treated questions. I began to imagine a muse of curiosity. What if we focused on those initial sparks of wonder, helping people form and refine their questions instead of rushing past them?

One-handed while holding our baby, I began drafting the first version from my phone.

Curiosity's Problem with Search and AI Today

We all know the phrase: "Just Google it." Then "Google told me so." Now it's "Did you ask ChatGPT" or Hey, @Grok, is this true?"

It has come to mean don’t linger, don’t explore, don’t reflect. Get an answer instantly and move on. The design of search engines rewards speed over depth. AI systems, however impressive, often amplify this pattern: the quicker the response, the better.

But curiosity isn’t a distraction. Curiosity is where growth and learning begins.

The Purpose of Hypandra

Hypandra exists to protect and promote curiosity.

Instead of collapsing questions into answers, Hypandra centers the first steps: forming, refining, and exploring questions. It helps you slow down and reconsider. It invites you to hold an initial question loosely, consider alternatives, and cultivate better questions before leaping to conclusions.

Hypandra is not a search engine and not an answer engine. It is a curiosity engine—a muse that keeps attention on the quality of the question.

Hypandra Today

Explore

You submit your question and we generate Reflections and further Questions. The reflections are presented as though from Hypandra, the Missing Muse of Curiosity. Hypandra does not provide answers, but different ways of thinking about the practice of asking your question. The generated questions are designed to help you engage with your question from multiple directions and levels. You can also use Build and Launch to dynamically reformulate your questions or send them to external search and AI tools.

Labs

We have a few ways to experiment further with questions:

Are You Curiouser? A game where you try to ask a question that is more curious than an AI.
Mind your WPQs Questions generated from pages recently edited on Wikipedia.
Input Constraints A question generator that you constrain with different required or forbidden strings.