The modern family has no shortage of photographs. It has a shortage of finished things. The birthday sits beside the boarding pass, the blurry dog, three identical sunsets and a screenshot of a parking spot, all waiting in the same camera roll. Andrew Laffoon has spent most of his adult life trying to turn that unruly abundance into an object with covers.
He co-founded Mixbook with Aryk Grosz in 2006, when online photo tools offered plenty of templates and not much freedom. Their wager was that making a photo book could be collaborative, expressive and less like completing a tax form. Twenty years later, the medium is still paper. Nearly everything before the press has changed.
Laffoon's current project is Story Mode, an AI-assisted flow that begins with a conversation. Describe the trip, the people and the feeling. Add the photographs. The system organizes scenes, proposes layouts and captions, and returns a book that can still be edited. The proposition is daring in a quiet way: let software touch the family memory without leaving machine fingerprints all over it.
A Commodore in farm country
Laffoon grew up in a small town in rural North Carolina, a place he has described as known for farming rather than technology. He was the child who dismantled household electronics, from VCRs to kitchen appliances, to see whether he could put them right. His father eventually supplied a safer object of inquiry: a Commodore 64 for his eighth birthday. Laffoon learned to code by tinkering with the games he loved.
A student from Duke showed him how websites were made and how a person might build a business around them. By 1997, while still young, Laffoon was working independently on websites. He later handled IT development at Morning Star International and did technical work at Active Reasoning, with duties that ranged from servers to interface design.
UC Berkeley gave the curiosity a partner. Laffoon studied Industrial Engineering and Operations Research, completing his degree and a certificate in engineering management and innovation in 2005. In an entrepreneurship course he met Grosz, another engineer inclined to take both systems and stories seriously. Berkeley's rhetoric of democratization stuck with them. If software could put sophisticated tools into ordinary hands, why not apply that logic to the stories people tell about themselves?
The viral success that failed
Their first framing was collaborative yearbooks. A small committee typically chose which moments represented a much larger school; Laffoon and Grosz imagined letting groups contribute their own photographs and pages. Advisers pushed them toward the broader photo-book market, and the founders launched a consumer product in 2007 with early backing.
Then came a result that looked wonderful from a distance. Their Photobooks application on Facebook spread to roughly two million users in two months. Friends could be tagged into a photograph and invited into the app. The loop worked. The cash register did not. Attention traveled quickly; buying behavior barely moved.
The founders heard no from about fifty venture firms. Laffoon later said it was among the most common words of Mixbook's first two years. The useful move was not heroic suffering. It was curiosity. They called the people who had paid and asked why. Those customers praised design freedom and an interface that made a complicated task feel manageable. Mixbook added options, improved the book and concentrated on the audience already showing intent.
Connections from Berkeley helped an early financing come together just before the financial crisis. Discipline followed. The company found profitability, then raised a $10 million Series B led by Level Equity in 2011. The neat lesson is not that persistence always defeats rejection. It is that a viral loop and a value loop are different pieces of machinery. Mixbook had to learn which one printed.
The order worth more than $67
The company's deeper purpose arrived by email. An early customer wrote about making a remembrance book with her mother and sister. Working on it together brought the family closer. For Mixbook it was a $67 purchase, Laffoon recalled; for the customer it changed relationships. The note altered the founders' horizon. Their original plan had been to build a company, sell it and begin another. Now the work offered a reason to stay.
Mixbook defined its purpose around creativity and connection. This could sound misty if it did not produce practical decisions. The company avoided inserting advertisements into a product made from intimate photographs. It invested in print quality and a network of production partners. It kept the editor flexible because the customer, not the template, had to recognize the finished book.
There is a pleasing contradiction in Laffoon's career. He is a software founder whose product succeeds when the software disappears. Nobody places an editor on the coffee table. The proof is the book that arrives, the relative who opens it and the story that gets told again because somebody finally sorted the photographs.
Every generation gets a blank page
Mixbook kept revisiting the first-step problem. AutoMix used photo metadata to arrange pictures. After acquiring the Yobongo team in 2012, the company introduced Mosaic, a mobile app designed to produce a compact hardcover book in minutes. The design achieved speed partly by removing decisions: a useful reminder that creative freedom and completion can pull in opposite directions. Montage followed with more algorithmic assistance.
Collaborative yearbooks become a customizable photo-book platform.
A viral Facebook app proves demand for participation, but not yet for purchases.
Mosaic compresses mobile photo-book creation into minutes.
Patented editing systems reflect the unglamorous engineering beneath easy creation.
Story Mode turns an interview and a camera roll into an editable first draft.
The products changed because the bottleneck changed. Early customers struggled to upload and arrange enough pictures. Smartphone owners now have too many. The camera roll is not a curated archive; it is a small attic carried in a pocket. Choosing the beginning, middle and end has become more exhausting than taking the photographs.
Story Mode attacks that new surplus. It asks about the event and the desired tone, analyzes the uploaded set, gives large moments room to breathe, groups supporting scenes and suggests narrative text. Three versions can be generated, and every detail remains editable in Mixbook Studio. The machine supplies momentum, not custody.
AI with its elbows off the table
Laffoon says Mixbook's interest in AI is not a late conversion. The earliest product included AutoMix; both founders had studied and discussed the field at Berkeley. Generative systems widened the possible job from sorting images to understanding stated intent. They also introduced a new way to ruin the product. A caption can be grammatical and still feel like a stranger has narrated your childhood.
This is why his language about the technology keeps returning to restraint. Memories are personal material. If the result feels processed, generic or oddly intrusive, convenience has defeated the point. Story Mode conducts a short interview because context does not live entirely inside pixels. A photograph can show who stood near the cake. It cannot reliably know why that afternoon mattered.
For builders, the transferable insight is less glamorous than “AI changes everything.” Trust depends on choosing which part should change. Let the system confront the pile, make selections and risk a first draft. Let the person correct the emphasis, rewrite the sentence and decide what leaves the screen. Good automation may be the colleague who sets the table and then knows when to leave the room.
Laffoon's own pattern has been similarly iterative: take a system apart, notice the friction, put it back together with fewer intimidating choices. That instinct traveled from a Commodore 64 to collaborative publishing, mobile books and now a conversational editor. It also survived fashionable detours. When new products pulled attention from Mixbook's core audience, he later described the answer as doubling down on the original vision. The reorientation restored growth.
The useful stubbornness of one problem
Inc. named Laffoon and Grosz among its 30 entrepreneurs under 30 in 2012. Laffoon is also named on patents for automatic saving and for scaling content inside an online collage editor. The awards make a tidy résumé. The more revealing achievement is duration: he has stayed CEO while the meaning of a photograph moved from desktop folder to social post to AI input.
Through those shifts, Mixbook's physical ending supplied discipline. A feature is not complete when it performs an impressive demo. It is complete when a customer can finish a book, trust what appears on the page and send it to print. Software likes endless possibility. Paper charges by the page and eventually demands a decision.
That is the charming seriousness of Laffoon's project. The world produces images at industrial volume, but affection remains stubbornly artisanal. Somebody still has to decide that this picture belongs beside that sentence. Somebody has to stop editing. Somebody has to press order. Twenty years in, Mixbook is giving the machine more of the chores and the person the final word.