GM was reported to have halved its vehicle-development cycle using AI. The signal comes from an IEEE Spectrum report backed by three arXiv papers—press and academia rather than an official announcement, a case study with research support. The focus is manufacturing application: embedding models across design, simulation, and validation to cut prototype iterations and compress the cycle. The through-line is generative AI landing from research and announcements into a stage of real industrial-process efficiency. Whether the halving generalizes or spreads to other makers can't be concluded from a single case—it awaits accumulation.
Robotics × Multimodal
GM halves dev cycles with AI
GM halves dev cycles with AI
✎ Story body
▲ Official & Press
Press
General Motors Is Cutting Its Development Cycles in Half
GM halves vehicle development cycles using AI and simulation
Academic (arxiv etc.) 3 ▾
Academic
CRAX: Fast Safe Reinforcement Learning Benchmarking
CRAX: fast benchmarking for safe reinforcement learning
Academic
Measuring commonsense and knowledge retention in VLA models
Academic
OneCanvas: 3D Scene Understanding via Panoramic Reprojection
OneCanvas enables VLM 3D scene understanding via panoramic reprojection