The first thing Mike Chen asked Magic to do was deliver chicken fried rice. This was not a keynote-ready gesture or the sort of immaculate origin myth that arrives prepackaged for a streaming series. It was lunch. Chen had registered a phone number, put a few people on standby and sent a text to see whether desire could be converted into action with no app, no menu and almost no ceremony. Co-founder Michael Rubin handled the order. The rice arrived. A dangerous amount of possibility came with it.
Chen shared the number with a few friends. They asked whether they could share it with theirs. He made a small website so he would not have to keep explaining the idea. Then, in February 2015, the page landed on Product Hunt and Hacker News. Magic leapt from roughly 30 users to thousands in minutes. A queue formed. Friends, relatives and at least one co-founder's girlfriend were recruited to answer requests. The company discovered the classic startup problem in its purest form: everyone wants the future, preferably before dinner.
A very ordinary interface
The early service promised to handle nearly anything legal through text. People asked for food, reservations, errands, gifts and, with the enthusiasm of customers testing a new boundary, cars and helicopters. The interface was almost aggressively familiar. You did not learn where the button lived. You wrote a sentence. The burden of turning that sentence into a sequence of calls, searches, payments and follow-ups moved elsewhere.
This made Magic both compelling and operationally alarming. A text box can accept infinite variety; a young company cannot. Every elegant customer experience hid a branching tree of human effort. Someone had to interpret the request, find a supplier, check the details, communicate a price and keep the promise. The cleaner the surface became, the more disciplined the machinery behind it needed to be.
“We made this number as a test, and I had a few people standing by.”Mike Chen, on Magic's first hours
There is a lesson here worth stealing. Most business software asks customers to adopt the company's internal map: these fields, this workflow, that dashboard. Magic began with the customer's language and accepted the mess that followed. Describe the outcome. Let the service negotiate the route. A decade before conversational AI made every software company reconsider its interface, Chen's team was already learning that conversation is the easy layer. Responsibility is the product.
Before the phone number, the printer
Magic was not Chen's first attempt to move difficult work closer to where it belonged. In 2010, he met Aaron Kemmer and Jason Dunn during Singularity University's Graduate Studies Program. Along with Michael Snyder, they formed Made In Space. The premise began with a constraint that sounds obvious only after someone decides to challenge it: objects used in space were built on Earth, then subjected to an expensive and violent rocket journey.
The team developed a 3D printer designed for microgravity. Ordinary printers could not simply be turned sideways and declared ready for orbit. Without gravity, parts oscillate. Heat behaves differently because natural convection disappears. The machine had to survive launch, operate safely around astronauts and demand little of their limited time. Its controls were reduced to one physical button. Commands and monitoring could happen from the ground.
“Instead of launching things to space, just print it there.”Chen's compact case for off-world manufacturing
In 2014, the printer reached the International Space Station and produced the first parts manufactured off Earth. Chen was listed the following year in Forbes' 30 Under 30 for Manufacturing & Industry. His name also appears as an inventor on Made In Space patents covering remote operations and several methods of additive manufacturing. Redwire acquired the company in 2020.
The useful distance
The two companies occupy different galaxies of the business map, yet their underlying question is similar: where should the work happen? If a replacement part is needed in orbit, perhaps the right answer is not a launchpad. If a founder needs a meeting booked, a spreadsheet cleaned and customers followed up with, perhaps the right answer is not another late night at the laptop.
Magic's offering changed as the company grew. The breathless 2015 vision of anything on demand developed into dedicated virtual assistants for businesses, backed by a pool that can provide round-the-clock coverage. The work now includes executive administration, customer service, bookkeeping and other recurring operations. Assistants also receive access to AI tools. The trick is no longer pretending that software or people must win a cage match. It is deciding which one should do which part.
Those figures reveal a different product than the phone number that once scrambled to source cookies and Cabernet. Talent selection, training and quality control have become part of the system. The service has moved from improvising a legal way to satisfy any request toward matching businesses with people who can learn a particular operating rhythm. The promise got narrower and, in practical terms, larger.
A skeptic goes remote
Chen did not begin as a remote-work evangelist. Magic initially had a US office, then opened an office in the Philippines in 2016. Two locations meant meetings already happened through Slack and calls. Employees began asking a perfectly irritating question: if most of the collaboration was remote anyway, why did they need to travel to an office to do it?
He resisted. Productivity and culture seemed vulnerable. Then the evidence changed his mind. Employees working remotely performed well, morale improved and the hiring pool expanded. The company allowed the pattern to spread in stages. By early 2020, about three quarters of Magic was already remote, making the final move less revolution than housekeeping.
Chen's practical advice was not to romanticize the arrangement. Distributed teams require deliberate communication. Transparency matters because people cannot glance across a room and see what others are doing. Recurring meetings matter even when an agenda looks thin, because they replace the ordinary contact an office supplies by accident. Time zones remain awkward, though some workers value schedules that open daytime hours for family. Remote work removes the hallway and leaves management standing in plain view.
“Make sure that transparency is a high level company value that everyone understands.”Chen on distributed culture
He also confessed an unfashionable preference: he dislikes video chat. The screen, in his view, can make everyone stare at the computer during a meeting, exactly where distraction lives. Audio allows him to walk, look away and think. It is a small quirk with a larger point. A remote company should not preserve every office ritual in digital amber. The medium should serve the conversation.
From builder to team builder
Chen traces his urge to build back to childhood evenings spent reading about mathematics, science and computer programming. At 17, while working at a local pharmacy, he decided he wanted to start a software business. He left and began working on it that night. The anecdote sounds impatient because it is. His advice to younger people has been to stop delaying the work they actually want to do, particularly when the cost of recovery is low.
Yet his account of entrepreneurship is less solitary than the origin stories suggest. He credits co-founders, employees, investors and luck. Earlier in his career, he concentrated on software engineering, sales, marketing and product design. As his companies matured, recruiting and leadership became the main focus. The individual learns to do the work; the executive learns to create conditions in which many people can do it well.
Magic makes that transition unusually visible because delegation is both the product and the organizational problem. The customer hands over context without surrendering standards. The assistant takes responsibility without becoming invisible. The company supplies selection, process and tools. Chen's own role follows the same arc: less heroically answering every request, more carefully building the machine that can.
The work behind the magic
There is a tidy temptation to describe Chen's career as a leap from space to assistants. It is more useful to see the continuity. Made In Space reduced dependence on rockets by moving manufacturing into orbit. Magic reduced dependence on a busy person's attention by moving execution to a distributed team. Both required remote operations, clear instructions, trust and systems that could function when the operator and the problem were far apart.
The distances are different. The management problem rhymes. A machine on the International Space Station cannot wait for someone to jiggle a cable. An assistant twelve time zones away cannot depend on overhearing the right hallway conversation. Reliability has to be designed, context has to travel and the instructions have to survive the journey.
Magic's first weekend proved that people loved the idea of making a wish. The following years were about the less theatrical work of fulfillment. Chicken fried rice made a charming opening scene. The company was built in everything that happened after the text was sent.