The children had notes. Dan Shapiro had made them a board game, an act of parental generosity usually rewarded with five quiet minutes and perhaps an oat cracker. His four-year-old twins grasped the rules of Robot Turtles at once, then began filing feature requests. One disliked a wall. Another wanted to play more than one card at a time. Most memorably: “Daddy, I want lasers!” It is tempting to treat this as foreshadowing. In Shapiro’s life, lasers have a habit of arriving before the plot asks for them.
He made holograms in high school. At Harvey Mudd College, where he earned an engineering degree, he built amplifiers, speakers, and laser shows to help pay his way while DJing parties. After graduation came Microsoft and the Windows 98 kernel team, then a Linux-based mobile-phone startup, then companies of his own. Software kept hiring him. Tangible objects kept calling him back.
His career looks restless on paper: Ontela, Photobucket, Sparkbuy, Google Comparison, a children’s game, a startup guidebook, Glowforge, and now a research post at Wharton’s Generative AI Lab. Yet the zigzag follows a clean instinct. Shapiro likes powerful tools whose power is needlessly hidden. His work is the business of removing the velvet rope.
The turtle asked for lasers
Robot Turtles began as an escape from Chutes and Ladders. Shapiro wanted quality time with his twins, but he also wanted the small, electric moment when a child discovers that a complicated thing can be understood. The game borrows from Logo, the programming language he learned at computer camp in the early 1980s. Children play instruction cards. Adults move the turtles. A bug can be undone with a silly noise. Programming logic slips into the room wearing a cardboard shell.
In 2013, he put the game on Kickstarter with a $25,000 target. It was funded in five hours. The campaign eventually collected $631,230 for 25,000 games headed to 65 countries. Shapiro switched immediately from inventor to operator. He revised production plans, communicated relentlessly, and shipped a little early. When an overseas-fulfillment plan collapsed, a replacement partner helped move nearly 2,000 international orders before the holidays. Success had arrived carrying a clipboard.
“Crowdfunding campaigns work best when they describe something to the world that people want to exist.”Dan Shapiro, on the Robot Turtles campaign
ThinkFun licensed the game in 2014 and took it into national distribution. Shapiro had learned two lessons that would matter later. First, people will prepay for a future they can picture. Second, the cheerful promise on a campaign page eventually becomes a box that must arrive at an actual door.
A factory learns to sit on a desk
Prototyping Robot Turtles introduced Shapiro to computer-controlled laser cutting. Soon an industrial carbon-dioxide laser from China occupied his garage, the natural habitat of all devices that alarm insurers and delight engineers. He traced an architectural pattern he admired and turned it into a replacement ceiling lamp. He made a camera-stabilizing rig. When he needed a wallet, he fabricated one from leather. “You don’t understand,” he later recalled telling someone. “I have a laser.”
The revelation was not industrial efficiency. It was personal agency. Shapiro, who jokes that he cannot draw and is “artistically incompetent,” had made beautiful, useful objects anyway. His children absorbed the new order quickly. At six, they asked why he would not print them an iPad. In their cosmology, making came first; Amazon was the contingency plan.
A reconnection supplied the missing co-founder. Mark Gosselin had spent the year after selling his previous company building a laser cutter from scratch in his own garage. With Tony Wright, they formed Glowforge in Seattle in 2014. The idea was to domesticate an old factory technology: combine precise optics and motion systems with a camera, cloud software, and a friendly interface, then let a newcomer cut and engrave wood, leather, acrylic, paper, and other familiar materials.
A Glowforge does not build plastic shapes layer by layer. Its laser cuts, scores, and engraves existing material. Shapiro’s preferred metaphor was the personal computer: machinery once reserved for specialists could migrate onto ordinary desks, provided the expertise was packed inside. “They are buying our skills and capabilities,” he said of customers. The supply chain manager, mechanical engineer, and software designer would all be invisibly present when a user pushed the machine’s single button.
Twenty-seven point nine million reasons to lose sleep
Glowforge did not launch on Kickstarter. The company built the campaign on its own website so it could maintain a direct relationship with buyers. A referral program turned roughly 1,500 email subscribers into more than 10,000 in a month and a half. The team agonized over the video, tested edits, and expected a good outcome to mean perhaps $2 million. Five million dollars would merit a high five.
The campaign took more than $5 million in its first week and $27.9 million in 30 days. The number made headlines. It also converted a young hardware company into the custodian of thousands of expensive promises. Crowdfunding had again provided demand, community, and accountability, only now every noun was wearing work boots.
Two campaigns, one lesson in demand
Bars use a compressed visual scale so the smaller campaign remains visible. Glowforge’s total was about 44 times larger.
The product reached homes, classrooms, and small businesses. Shapiro described Glowforge as his life’s work because engineering had changed what he could do, and the machine offered a way to package that ability for someone else. It was an unusually earnest ambition for an object capable of engraving a photograph onto a snack. Then again, democratization is often improved by the occasional personalized cookie.
The bill for growth comes due
For years, the familiar venture story continued: new models, major funding rounds, a larger staff, broader retail ambitions. Shapiro said Glowforge was weeks from going public in 2022 when the crafting market collapsed. Strategic partners struggled. The next three years became an effort to preserve the company and keep serving customers. In 2025 came multiple rounds of layoffs, a failed funding round, and the closure of a Seattle production facility that had only recently brought assembly of high-end machines back from Mexico.
Glowforge entered an Assignment for the Benefit of Creditors, a state-level restructuring process and alternative to bankruptcy. The neat upward graph had become something more candid. Shapiro later said that the pressure to scale, prepare for the next funding round, and justify rising valuations had pulled the company away from what mattered.
“The need to scale at all costs... it pulled us away from what mattered.”Dan Shapiro, reflecting on Glowforge’s restructuring
It is a line with extra weight from the author of Hot Seat, his 2015 guide to the CEO decisions that cannot be delegated. Shapiro’s humor about the role is reliable. He calls it “management overhead and general-purpose troublemaking.” Yet the joke contains a sober idea: the chief executive is the person left holding the problem after all elegant job descriptions have run out.
A second founding
An independent third party ran a sale process for Glowforge’s assets. No buyer emerged. The entity holding them was prepared to liquidate the company. Shapiro and Gosselin made a bid from their own savings and acquired the trademark and brand, the hardware platform, software, and rights to manufacture more machines in Seattle. The reboot could support about 20 former Glowforge employees at the outset.
The plan sounded almost quaint after the large rounds and public-market preparations: improve the hardware, improve the software, serve customers. Get back to basics. The founder who once asked a crowd to imagine a personal factory was now asking a small team to imagine the company without the apparatus of perpetual scale.
Graduates from Harvey Mudd and joins Microsoft, working on Windows 98.
Sells six-month-old Sparkbuy to Google and later leads Google Comparison Inc.
Robot Turtles turns a family experiment into a $631,230 Kickstarter campaign.
Publishes Hot Seat; Glowforge records $27.9 million in 30-day preorders.
Joins Wharton’s Generative AI Lab and buys Glowforge’s core assets with Gosselin.
There is a pleasing symmetry here. Robot Turtles was a one-person project aided by talented contractors. Early Glowforge lived in garages and a century-old building that once housed Boeing work. The smaller company is not a return to innocence, because creditors, layoffs, and disappointed plans do not vanish when the cap table changes. It is a return to directness. A product, a customer, and the duty connecting them.
Now the software talks back
Even while Glowforge contracted, Shapiro’s curiosity expanded toward generative AI. In 2025 he became a Senior Research Fellow at Wharton’s Generative AI Lab, working with Ethan and Lilach Mollick. He said he wanted to share lessons from building generative-AI tools at Glowforge and investigate the “strange and truly bizarre phenomena” he had observed in large language models.
The interest fits his old pattern. Shapiro says he cannot code, despite a career steeped in software. Modern AI gave him a sudden ability to build tools anyway, an experience he compared to the “I know Kung Fu” moment in The Matrix. At Glowforge, his team experimented with AI for sales coaching, editing, brainstorming, production analysis, and financial review. His 2026 writing moved deeper into agentic software development, including a workflow he calls Trycycle.
Once more, the object of fascination is a capability escaping its guild. A preschooler can think like a programmer before learning to read. A self-described non-designer can make a lamp. A non-coder can produce software. Each leap contains risk, and Shapiro is too experienced to confuse access with ease. Someone still has to ship the box. Someone still has to check the output. Someone still has to answer when the plan meets a holiday deadline, a factory floor, or a creditor.
The fine art of starting over is therefore less romantic than the phrase suggests. It is a craft of smaller numbers, sharper obligations, and reclaimed attention. Shapiro has spent much of his career helping people move an idea from the imaginary world into the physical one. In the second life of Glowforge, the object on the workbench is the company itself.