Carbon: graduating from the experiment - Chandler Carruth - NDC Toronto 2026

Value of the Information & Strength of the Argument

The talk provides valuable insights into the design and implementation of a successor language to C++. The argumentation is strong, backed by live demos and concrete examples. Carruth effectively argues that C++’s evolution process is inadequate and that a new language is necessary. He presents Carbon’s interop as a key differentiator, demonstrating its technical feasibility. The memory safety design is well-motivated, with a clear incremental path. However, the talk is promotional, and some claims about the language’s readiness may be optimistic.

The title accurately reflects the content: a status update on Carbon and its graduation from experiment to production.

8/10

The talk is delivered by the project lead, provides concrete technical details, live demos, and references to the open-source project. Claims are consistent with the project's public status. However, as a conference talk, it is promotional and lacks peer review.

Markers derived by PSI from the transcript: the creator did not define chapters.

• 02:00Introduction and motivation for Carbon: C++'s problems and the need for a successor.
• 10:00Overview of Carbon's current syntax and features, including classes, methods, and access control.
• 20:00Demonstration of C++ interop: importing C++ headers and calling C++ functions from Carbon.
• 30:00Live demo of bidirectional interop: C++ class deriving from Carbon class and calling Carbon methods.
• 40:00Discussion of memory safety design: strict and permissive modes, incremental migration.
• 50:00Plans for graduating Carbon from experiment to production, community and governance.

• NDC Conferences — Conference organizer, where the talk was recorded.
• NDC Toronto — Specific conference event.

• Carbon Language GitHub — Open-source project repository, consistent with the talk's claims.