Conferences

FAQ

  1. What are the main C++ conferences where I will learn the most about the language and libraries?

    The largest conference each year is CppCon, currently held in September in Aurora, Colorado. Talks typically include library authors and standards contributors.

    Smaller, deeply technical, and perhaps with more hallway style conversations is Cpp Now. Held in May, in Aspen, Colorado. This conference started as BoostCon. A good event for meeting Boost library authors.

    In Berlin, Germany, typically in early November, there is Meeting C++, one of Europe’s largest with a practical approach to modern programming and tooling.

    Also in Europe, there is the ACCU Conference, recently held in Folkestone, UK, in June. The focus here is on best practices of software engineering.

    These are the big four conferences each year. The ISO Cpp Working Group meetings occur three times per year (typically March, June, November), and the location rotates around the globe. This event is where the language standard is defined and updated. Refer to the FAQ on [ISO C++ Committee Meetings].

    There are also some regional events that might work well.

    The using std::cpp conference in Spain in March has a strong technical focus and is popular amount European developers.

    Also in Europe and in March, there is the BeCPP Symposium in Belgium.

    In Asia, the C++ Summit is held in Shanghai, typically in November though the times can vary.

    If you attend several conferences, you may well find projects debut at one of the smaller conferences before a more polished presentation is given at CppCon.

  2. Is there such a thing as online C++ conferences?

    Yes, not free but there is Cpp Online.

  3. What were the overall themes of the recent Cppcon 2026?

    In a single sentence: C++ is trying to retain its traditional strengths — control, performance, deterministic behavior and zero-overhead abstraction — while systematically attacking its traditional weaknesses: memory safety, undefined behavior, difficult metaprogramming, asynchronous programming and lack of introspection.

    Five themes particularly dominated:

    • Safety without giving up the C++ performance model

    • Reflection and compile-time programming becoming dramatically more powerful

    • A major rethink of asynchronous/concurrent programming

    • C++ becoming increasingly important for specialized, maybe embedded, resource-constrained areas of high-performance computing

    • AI is changing what "being a C++ programmer" means, rather than simply making C++ programming disappear

      Perhaps the following diagram broadly shows current thinking:

                                 C++ 2026
                                     │
                ┌────────────────────┼────────────────────┐
                │                    │                    │
             SAFETY            EXPRESSIVENESS        PERFORMANCE
                │                    │                    │
          ┌─────┼─────┐          ┌───┼────┐           ┌───┼────┐
          │     │     │          │   │    │           │   │    │
       profiles contracts  reflection  constexpr    SIMD    locality
          │     │     │          │   │    │           │   │    │
          └─────┴─────┘          └───┴────┘           └───┴────┘
                │                    │                    │
                └────────────────────┼────────────────────┘
                                     │
                                 Modern C++
                                     │
                ┌────────────────────┼────────────────────┐
                │                    │                    │
            EXECUTION             HARDWARE             AI ERA
                │                    │                    │
          senders/receivers       memory tiers         complexity
          coroutines              embedded             verification
          parallelism             SIMD                 architecture
          scheduling              determinism          system design
  4. What are conference talks focusing on when it comes to embedded C++?

    Talks on using modern C++ in embedded systems (such as microcontrollers) tend to address such code functions as std::start_lifetime_as, and issues such as byte representation, endianness, padding, reflection, and undefined behavior. Indeed talks on memory safety on microcontrollers cover topics including compiler warnings, the tools AddressSanitizer and UndefinedBehaviorSanitizer. std::span, std::expected, strongly typed wrappers, hardened containers, contracts, and safety profiles. The modern attitude seems increasingly to be "Resource constrained" does not automatically mean "use ancient C."

  5. Was there any talk at Cppcon 2026 that AI might be creating a new set of problems in addition to super-fast code generation?

    Certainly. Andrei Alexandrescu’s keynote was explicitly about the consequences of AI systems generating large amounts of code while human inspection and validation remain difficult. His central theme was that managing complexity and understanding system structure become increasingly important. This is an extremely important observation for C++: AI can produce code in seconds, but that does not mean a human systems integrator, for example, would understand the code’s lifetime guarantees, exception behavior, concurrency assumptions, ABI consequences, memory layout, performance characteristics, template instantiation costs, or where there is potential for undefined behavior.

  6. Is there humor at all at these conferences, perhaps in the "hallway track"?

    There is little humor in the presentations and questions and answers that follow in the formal presentations. You are right that the "hallway track" is where humor does come out, and can be quite entertaining. Here are some quotes or near-quotes from hallway conversations:

    • "A developer is a device that turns caffeine into algorithms."

    • "A C++ programmer is a device that turns undefined behavior into compiler errors."

    • "Modern C++ should make the right thing easier — and the dangerous thing harder."

    • "The future of embedded C++ isn’t smaller C++; it’s modern C++ with smaller footprints."

    • "Don’t optimize the average case when the worst case is what breaks the machine."

    • "Replace the pointer you hope is valid with the type that proves what you mean."

    • "Zero overhead doesn’t mean zero abstraction."

    • "The hard part isn’t writing the code; it’s making the system’s assumptions visible."

    • "C++ in 2026 isn’t trying to become less powerful; it’s trying to make that power easier to control."

      And some (paraphrazed) quotes on the current difficulty in guesstimating the impact of AI on developers:

    • "AI can write the code. C++ still has to explain why it crashed."

    • "Artificial intelligence is very good at finding patterns; C++ is very good at finding semicolons."

    • "AI promises to eliminate programmers. C++ promises to eliminate their weekends."

    • "The ultimate C++ optimization: ask the AI to generate the code, then ask the compiler whether it was any good."

    • "An AI model has parameters. A C++ programmer has templates. Nobody understands either one completely."

    • "AI wants more compute. Embedded C++ wants fewer bytes. Welcome to the negotiation."

    • "AI handles the uncertainty; C++ handles everything that isn’t allowed to be uncertain."

      Certainly the Boost hospitality booth, the exhibitor stands, field trips, free coffee stands, all encourage a healthy hallway discourse on many questions attendees have that might not be appropriate for a formal question and answer session.