OPEN LINK ↗
// 1h agoOPENSOURCE RELEASE
NVIDIA ASPIRE v1.0 open-sources robot skills
NVIDIA’s GEAR lab open-sourced ASPIRE v1.0, a continual-learning robotics framework where coding agents inspect execution traces, repair robot programs, and retain validated fixes as reusable skills. It now solves a Rubik’s Cube and reports 99% success across 80 LIBERO-Pro tasks plus selected Robosuite and BEHAVIOR-1K tasks. [Official project](https://research.nvidia.com/labs/gear/aspire/) · [GitHub](https://github.com/NVlabs/ASPIRE)
// ANALYSIS
The important breakthrough isn’t the cube demo—it’s treating robot competence like software that can debug, version, and improve itself.
- –ASPIRE combines closed-loop execution traces, evolutionary search, and a growing skill library instead of relying solely on fixed policies or gradient updates
- –The reported 99% benchmark result is striking, but it remains a lab evaluation—not proof of robust general-purpose autonomy
- –Sim-to-real transfer and online real-world learning make this more consequential for robotics developers than another simulation-only agent
- –The Apache-2.0 release and Codex/Claude Code workflows lower experimentation barriers, though generated Python runs with broad system access and requires isolation
// TAGS
aspireroboticsagenttool-useframeworkopen-sourceresearch
DISCOVERED
1h ago
2026-10-16
PUBLISHED
1h ago
2026-10-16
RELEVANCE
9/ 10
AUTHOR
AGTPinsights