In 2010 I helped build the first makerspace inside a school anywhere in the world, at a public school in Moscow, and the first one in the United States a year later. That was where FabLearn started. It grew into a program of fellows, conferences, and school labs.
It now reaches 27 countries. The machines were never the point. A laser cutter in a school changes nothing unless the curriculum around it lets children decide what is worth making.
Papert and Freire were contemporaries who barely spoke to one another. One wanted children to build things; the other wanted them to be free. I have spent twenty years arguing that neither idea survives long without the other. Construction without politics produces clever toys. Politics without construction produces slogans. Most educational technology still arrives as delivery, content pushed at students at whatever pace an algorithm has decided for them. The older and harder alternative is to hand children the tools and let them decide what to make.
This perspective runs through the Columbia University Paulo Freire Initiative and a decade of work with Brazilian public education through the Lemann Center.
Arnan Sipitakiat and I designed the GoGo Board in 2001, as graduate students at the MIT Media Lab, for an unglamorous reason: a single commercial robotics kit then cost roughly a Brazilian teacher's monthly salary. The board is open-source, locally manufacturable, and now in its seventh generation.
The two patents came later, for Google Bloks, a tangible toolkit that lets a five-year-old write a program by arranging blocks on a table, before she can read.
Open-ended learning is hard to measure, and in education what cannot be measured tends to get defunded. With Marcelo Worsley and Bertrand Schneider, then doctoral students in the lab, we started a field to deal with this: multimodal learning analytics, which treats speech, gesture, sketch, and the artifact itself as evidence of how a student is thinking. The danger is obvious enough. Measurement can flatten the very thing it was built to see, which is why these methods should be designed by people who have spent time in a makerspace and not only in front of a model. The work appears in the Journal of Learning Analytics and the Journal of the Learning Sciences.
The four lines of work above run through named projects in the lab, each with its own classrooms, collaborators, and tools.
Selected articles across learning sciences, engineering education, and biotechnology.