Make is personal, creative machine learning.

Program by teaching, not by learning.

Make is a way to create live, personal creative systems on phones. We want to allow anyone to create complex, living ways to interact with technology – without the need to learn complicated syntax.

How Make works.

Make uses supervised learning to turn inputs and outputs into live systems.

Make sets of inputs and outputs by adjusting data (e.g. audio, or vector paths), and the system interpolates the ways sets of inputs can turn into those outputs. The more numerous and specific the inputs and outputs you set, the better and faster the system will learn.

How to use Make.

Make is organized by actions, which are characterized by sets of inputs and outputs.

When you create a new function, begin by choosing what types of inputs and outputs you want. Each type of input/output has its own interface; use these to adjust the inputs and corresponding outputs that you want the system to do, and then check them as a correct set. The current list of types is:

Once the action has a large enough set of inputs and outputs, you can begin testing the action. This consists of interacting with the action as a live system. If the outputs it gives you are right, confirm them to help it learn. If they're wrong, adjust the output until it is correct.

Eventually, the system will do what you want it to (most of the time) and you can start chaining actions to build complex systems without complicated code.

The apps' screens are organized by these interactions. From left to right: actions, sets of each action's inputs and outputs, individual input/outputs, and testing. Navigate these screens by swiping or selecting.

About Make

Make was developed by Barron, a designer at RISD, after developing the Bigtalk platform with David Lee and Sam Song.

Make was incubated in Computer Utopias under the mentorship of Chris Novello, with support from the members of other teams; Post, Bumble, Pip,

Want to help make Make? Email