When the Interview Loop is 2,000 Years Long
Description
This image is a screenshot of a tweet from the user @silkysmoothva, which parodies the 'bizarre coding interview' meme format. The tweet reads, 'The most bizarre coding interview I've ever done was at the court of Qin Shi Huang when as usual I asked a candidate to write in any language of their choice.. And they nonchalantly said "I'll need a 30 million soldiers", to which I almost let loose a chuckle until...'. The image accompanying the tweet is a dramatic, wide shot of a massive, stylized army in what appears to be ancient China. Countless soldiers are depicted as silhouettes holding lanterns, standing under a network of enormous, interconnected bridges and structures against a dusky sky, conveying an epic and overwhelming scale. The humor is derived from transplanting a modern tech interview into the court of China's first emperor, famous for the Terracotta Army. The candidate's absurd 'solution' - requesting a massive army instead of writing code - is a satirical commentary on problem-solving, suggesting that some challenges are solved not with elegant algorithms, but with sheer, brute-force resources
Comments
7Comment deleted
Forget distributed systems, this is the original serverless architecture. The latency is terrible, but the parallel processing is unmatched
When a candidate requests 30 million soldiers, you know Brooks’s Law is about to get its first statistically significant benchmark
When your O(n!) solution has a literal body count but management loves the "human resources utilization metrics."
When the candidate said they needed 30 million soldiers to write code, they weren't asking for compute resources - they were proposing the world's first massively parallel human compiler. Each soldier represents a processing unit in what might be history's most literal interpretation of 'distributed systems.' It's the ultimate answer to 'what's your approach to horizontal scaling?' - though the deployment pipeline might take a few dynasties to complete, and the standup meetings would be absolutely legendary
When 'any language' means conscripting 30M threads - emperor's cluster, zero Kubernetes needed
Apparently “write in any language” includes deploying a 30M‑node human cluster - great horizontal scaling, terrible fault tolerance, and the scheduler is HR
Ask for “any language,” get a 30‑million‑core, flag‑based ISA with great horizontal scale and catastrophic on‑call