MOUNTAIN VIEW, California — Tynker, the leading game-based coding platform that has engaged over 100 million kids, proudly introduces “Tynker Copilot.” Leveraging the capabilities of Large Language Models (LLMs), Tynker Copilot empowers young innovators aged 6-12. It provides a seamless interface for these budding developers to transform their ideas into visual block code for apps and games. Additionally, when exploring existing projects, kids benefit from the tool’s ability to explain block code fragments, ensuring a deeper understanding. Tynker Copilot allows children to build confidence as they work with AI, laying a solid foundation for their future. With this launch, coding education takes a significant leap forward.
Large Language Models (LLMs) have excelled in text-based programming languages like Python and JavaScript. However, their application to visual block coding, the primary introduction to programming for many kids, had yet to be explored. Tynker is the first to bridge this gap. Our latest integration lets children quickly convert their ideas into block code, streamlining their initial coding experience.
Tynker introduction of the Copilot feature marks a significant industry milestone. Until now, the capabilities of LLMs have not been fully utilized for the younger age group. Tynker Copilot empowers children as young as 6 to input text commands like “Design a space-themed knock-knock joke” or “Teach me how to build a Fruit-ninja style game” and receive block code outputs with step-by-step instructions. Moreover, when debugging existing projects, students can submit block-code snippets and benefit from LLM-generated natural language explanations.
“Our ethos at Tynker is grounded in innovation and breaking boundaries,” stated Srinivas Mandyam, CEO of Tynker. “We’re exhilarated to channel the latest advancements in AI technology to serve our youngest learners, fostering an environment of curiosity and growth. While the potential for such a tool is vast, our prime objective remains to employ it as an empowering educational aid, not a shortcut.”
Developing Tynker Copilot posed challenges distinct from standard LLMs, which typically handle text-based code. Our focus was visual block coding tailored to kids’ unique programming preferences. With insights from millions of submissions in Tynker’s repository, ranging from pet dress-up games to platformer adventures and Minecraft mods, we fine-tuned Llama2. This rich, child-centric data helped us adapt the model to generate visual block coding programs, making Tynker Copilot a specialized tool designed around the creative tendencies of young coders.
“Bringing LLMs into Tynker’s dynamic visual block coding ecosystem presented unique challenges,” commented Kelvin Chong, CTO of Tynker. “We fine-tuned Llama2 with AI and ML techniques to align with young coders, supporting parallel scripts, visual actors, and kid-centric features such as Physics, AR, sound effects, and Minecraft modding. As coders grow, they’re naturally transitioned towards Python.”
All educators and home users will be notified when this innovative feature becomes generally available in the product. In the meantime, for more insights into its capabilities and to view examples of its potential use, visit the Tynker blog
About Tynker
Tynker is a leading K-12 creative coding platform that empowers children worldwide to develop coding skills for designing animations, games, and a wide range of real-world applications. With a focus on critical thinking and problem-solving, Tynker’s acclaimed curriculum is used by 1 in 3 U.S. K-8 schools, 150,000 schools globally, and over 100 million kids across 150 countries. Partners include renowned brands like Apple, Google, Microsoft, and NASA. Accessible on computers and mobile apps, Tynker offers free and paid subscriptions. In 2021, Tynker was acquired by BYJU’S, the world’s largest edtech company. For more info, visit www.tynker.com.
