姚顺雨拿50年数学难题成绩单，招人了

TL;DR - Tencent Hunyuan is recruiting for AI-for-Science after its Hyra research agent reportedly solved a 50-year-old additive-combinatorics problem. The result suggests agentic systems can move beyond search toward proposing and formally verifying original research.

• Built on the 295B-parameter Hy3 model, Hyra found a scalable construction that approaches the problem’s theoretical limit of 2.
• The agent reportedly developed the core idea in about 24 hours; researchers checked the proof and produced a Lean 4 formalization.
• Hyra-1.0 also reported improved results across mathematics, astronomy, quantum computing, and drug design benchmarks.
• Hunyuan is seeking expertise spanning agents, reinforcement learning, evaluation, training systems, GPU kernels, and scientific domains.