Ten Days, a Teen, and the Richest Man: How "Vibe Coding" Sent Two Student Projects to the Top of GitHub

A ten-day sprint and two viral repos

A Beijing university senior calling himself BaiFu woke up to find two of his projects — BettaFish and MiroFish — climbing GitHub’s global trending charts. Both were built with what he calls “vibe coding”: rapid, AI-assisted development that turns ideas into working products in days rather than months. BettaFish, a multi‑agent public‑opinion analyzer, and MiroFish, a follow‑on predictive engine, reflect a simple formula: stitch together large language models, scraping pipelines and agent orchestration, then publish. Is this a lone prodigy? Or the power of new AI tooling lowering the bar for productization? The answer matters to anyone watching how software is made today.

From inbox overload to Shanda’s bench

The viral attention brought offers and a flood of emails from ByteDance (字节跳动), Alibaba (阿里巴巴), Tencent (腾讯) and other players, BaiFu says — and a quieter pitch from Shanda (盛大集团), the legacy gaming and investment group founded by Chen Tianqiao (陈天桥). He chose Shanda’s offer to work in Shanghai, citing a promise of freedom to “do what you like” rather than immediate commercial pressure. It has been reported that Chen Tianqiao’s team made a separate investment of roughly 30 million yuan related to the projects, a claim the company has not formally confirmed to Western outlets.

From rear‑view to telescope

BettaFish was designed as a rear‑view mirror — analysis of what already happened. MiroFish aims to be a telescope, producing short‑term predictive scenarios by simulating agents and action paths. Both sprang from the same pragmatic observation: users want not just summaries but actionable foresight. The projects encapsulate a broader trend in China’s tech scene: rapid prototyping with commodity cloud compute and model access, then scaling under the umbrella of larger groups that control data and server capacity.

Why this matters beyond one student

For Western readers, BaiFu’s story is a lens on China’s AI ecosystem: large incumbents still hoard compute, datasets and talent, while nimble individuals use global open tools to surface ideas — sometimes spectacularly. This dynamic is unfolding amid geopolitical pressure and export controls on advanced chips that complicate access to top‑tier compute for Chinese firms, making partnerships with domestic giants strategically important. Whether BaiFu becomes a poster child for grassroots innovation or a cautionary tale about consolidation remains to be seen. For now, two ten‑day repos raised familiar questions: who builds the future, and who gets to scale it?