Liquid AI team: post-training is what makes on-device agentic models useful
JosephJacks_ · x · 2026-09-27
Liquid AI highlighted a discussion from its post-training team (Remi Labonne, Edoardo Mosca, Jiahui Wang) on what makes an on-device agentic model actually useful: post-training. The team covers how models learn to use tools, follow instructions, handle longer contexts, and recover when tasks get complex. Relevant for anyone working on on-device agent deployment.
More from coding & agent
- OpenAI L7 engineer says firm can't build IP-whitelisted egress VLAN, experts say IP allow-lists are obsolete — arthurcolle · 2026-09-27
- Code-rendered 'P(doom)' music video made entirely with Claude in Claude Code — SonglinYang4 · 2026-09-27
- Hackathon winner: wildlife sim running ~4,000 Jev decisions for ~$1.30 — schwentker · 2026-09-27
- 1500 agent PRs overnight: CodeRabbit built Triage to rank them by risk — schwentker · 2026-09-27
- Jev pricing reality check: 720 states x 30 questions likely costs under 10 cents — schwentker · 2026-09-27
- Field notes from Jevathon: Jev returns typed decisions, not prose — schwentker · 2026-09-27