Fortune Telling and Speculative Design

Speculative design is presented as a practical way to think and act amid rapid change, not as a claim that anyone can reliably predict the future. Design engineers are especially suited to the work because they combine interface design with knowledge of software’s underlying systems, and can turn an idea into a convincing prototype. The speaker supports this approach with past theories: digital gardens as an alternative to chronological, perfectionistic blogs; generative AI flooding the web with low-quality content; and language models enabling more people to build personal and community-focused software. In each case, noticing early signals and developing an interpretation helped make a possible future easier to discuss.

At GitHub Next, speculative work involves sharing rough prototypes early, responding to colleagues’ experiments, and collectively mapping emerging themes before choosing directions. Examples include a tool that checks GitHub issues against maintainers’ rules and a multiplayer canvas for sketching ideas that could later become agent tasks. Prototypes alone are not enough: teams need evidence, sound judgment, a plausible route to production, and the organizational support to invest in a new direction. The practical methods introduced are the futures cone, collecting weak signals, tracing second- and third-order effects, experimenting with ideas, creating artefacts from imagined futures, and making theories concrete and public. The central argument is that designers should engage with uncertainty by building and testing tangible possibilities, rather than giving up on understanding where technology may lead.