London's new Machenta Physical AI lab claims one-shot task learning
Ed__Johns · x · 2026-10-28
Investor Andreas Klinger spotlights Machenta, a new Physical AI lab in London with a "near-heretical" approach: no massive vision models, no hours of post-training, and no 20 video takes — show the robot a task once and it can very likely repeat it. The technical route leans heavily on in-context learning and explicit 3D, with more details not yet public. Klinger argues they have a legitimate claim to soon become world-leading.
More from Embodied
- Robot fighting league REK opens permanent Tokyo arena, led by the G1 katana guy — cixliv · 2026-10-28
- NVIDIA's open-source SimReady central library trends on Hugging Face — nvidia · 2026-10-28
- Embodied AI startup DYNA hiring across robotics roles after $120M raise — JasonMa2020 · 2026-10-28
- Dyna Robotics shows commercial deployments: 200 picks/hr, 22-min hotel room turns, ≥97% pass rate — JasonMa2020 · 2026-10-28
- Physical Intelligence in talks to raise $1B+ at roughly $20B valuation, Bloomberg reports — rohanpaul_ai · 2026-10-28
- PaceVLA: per-chunk adaptive planning makes VLA robots up to 1.8x faster at same success — chris_j_paxton · 2026-10-28