OpenAI recently launched GPT-6 Astra, an artificial intelligence model that can directly operate software like Figma and Excel. The release has renewed debates over the future of technical education. While some fear these tools make coding and design skills obsolete, researchers argue that foundational knowledge is now more important than ever. To prepare students, schools must move away from teaching rote software syntax and focus on logical reasoning and critical evaluation.
What Happened
OpenAI recently released GPT-6 Astra. We have tracked this model for its implications on homework security as well as its potential cybersecurity risks. According to the product announcement, Astra can browse the web and use standard work applications. OpenAI claims Astra solves complex financial modeling challenges four times faster than the fastest human competitor. Design teams are already using it within creative software like Figma. Unlike older AI tools that required custom code to connect programs, this model interacts with software like a human does, by clicking and moving through interfaces. It can draft code, build spreadsheets, and perform other tasks with little human oversight.
The Bigger Picture
While some claim AI makes traditional skills obsolete, cognitive science and educational research suggest otherwise. In computer science education, researchers in Nature Reviews Psychology argue that programming relies on logical reasoning rather than simple translation. Learning to code is still an important exercise for brain development. When students rely only on generative tools, a practice some call "vibe coding," they often miss subtle, deep-seated errors in the output. A recent meta-analysis of 35 studies found that AI tools can improve student motivation. However, the study also showed that students still need a deep conceptual understanding to debug and guide the AI.
A similar shift is occurring in creative subjects. Rather than acting as physical makers, students increasingly work as creative curators. Using AI too early can slow a student's growth. A review in the Proceedings of the 2026 Designing Interactive Systems Conference warns that students who use generative AI during projects often face "inquiry bottlenecks" and fixate on the AI's initial ideas.
Math teachers are also trying to balance AI assistance with independent learning. To combat falling test scores, tools like Kami Coach give step-by-step hints so students can solve problems on their own. Yet some researchers push back against systems that face the learner directly. They argue instead for a teacher-mediated approach where AI tools only help educators design lessons. At the college level, tools like MathGPT.ai grade handwritten homework. This allows professors to return to paper exams to prevent cheating.
What This Means for Families
The arrival of models like Astra means the end of rote learning. If an AI can format a spreadsheet or write basic syntax, memorizing those mechanical steps is no longer a good use of school hours. Instead, students must learn to analyze the data they receive. They need to understand the logic of mathematics and design so they can spot the false information these models frequently output.
What You Can Do
Parents and teachers can help by encouraging children to look for errors in AI-generated answers. Instead of letting them accept the output directly, ask them to explain why a specific line of code or math step works.
It also helps to use AI as a supportive tutor rather than an answer key. Introduce interactive learning aids like Kami Coach that ask guiding questions, rather than software that simply outputs completed homework.
Finally, do not worry about keeping up with every new software interface. Teach kids foundational logical thinking and design principles, which remain constant even as tools change.