Our research into consent-driven applications brought us together, ultimately resulting in the vision we have today. We started out building experimental social platforms (Civic and Smalltown), but came to understand, through our work with news organizations and systemically excluded founders, that a better solution—and one more aligned with the architecture of the Web—was to empower people to build their own applications. That realization has led us to a generative programming model, where people can leverage a powerful runtime to build consent-driven applications without writing code.

Learn more about our team. Sign up for our newsletter.

What is generative programming?

Generative programming takes a high-level representation of an application or more generally, a software system, and transforms it into working, executable code, through the use of generative AI. In contrast to visual programming, which directly maps visual representations such as flow charts or detailed wireframes into source code, generative programming dramatically reduces the intrinsic complexity of building software. Visual programming tools often simplify deployment but not the actual development. No Code tools may thus be divided into generative and visual categories.

Aside: Generative AI is commonly used to refer to unsupervised machine learning systems. However, generative programming may incorporate other forms of AI, particularly inference. Generative programming in this context should also not be confused with the use of the phrase in object-oriented programming, which has similar connotations but does not place as much emphasis on AI and typically has a lower-level focus.

We tend to say consent-driven, but we should probably say consent-based because the basic idea is that consent needs to be built into an application, rather than a choice left up to the application developer. Ideally, consent-based architecture is a form of Ulysses Pact that ensures that consumer rights cannot be violated, either intentionally or accidentally. More specifically, our goal is to build applications where consumers control access to their data, including knowledge about their activities, and are protected from identity theft, abuse, and harassment. Architectural patterns that characterize consent-based applications include:

  • Client-side encryption
  • Decentralized identity management
  • Consensus-based trust
  • Capability-based authorization

Consent-based architectural patterns are typically well-known but underutilized. They are not, by themselves, novel. Our innovation lays in our synthesis of them in the context of consent, and as an architectural foundation for No Code.