Imagi just raised $4.5 million, backed by Brighteye Ventures, Day One Capital, and will.i.am (yes, the musician), to teach students to build software through vibe coding, a practice where you describe what you want in natural language and AI assembles it. The round landed the same week a paper dropped on arXiv asking a question that would make any admissions officer sweat.
When Credentials Certify Nothing
A 2026 paper, What Does the Credential Still Certify? by Kai Yao, argues that generative AI has severed the core premise of educational assessment: that submitted work reflects what a student actually knows. The credential, Yao writes, now certifies completion of a process rather than acquisition of a skill. Vibe coding collapses this tension further. If the skill is prompting an AI to write the code, and the credential is for learning to code, you have a perfect philosophical knot. A separate arXiv paper this week, Did Alice Do Wrong? by Harrington et al., surveyed students across cultures on the ethics of using generative AI in computing courses. Spoiler: nobody agrees on what counts as cheating anymore.
Shop Class for the Inference Era
The optimistic read is that vibe coding is exactly what shop class was. Shop class did not teach students to become carpenters. It taught spatial reasoning, tool familiarity, and the pleasure of making. Vibe coding might teach systems thinking, product intuition, and iteration. Will.i.am investing in this is actually coherent: he has been early on technology-adjacent cultural moments before. Fast Company ran a parallel argument this week about the math crisis being a test design failure, not a learning failure. The common thread: institutions keep measuring the old thing while the new thing has already eaten the curriculum. The credential question is just beginning.