Inkling's open-source training stack is ready
BanghuaZ · x · 2026-07-16
This provides details on the "trainability" preparations for Thinking Machines Lab's open-source model, Inkling: the team collaborated with relevant projects weeks in advance to ensure it could be trained and deployed on Day 1.
The post emphasizes the combination of open model + open infra + open recipes, noting:
- Training ranges from LoRA all the way to full-parameter
- Related capabilities have been compiled into the Miles cookbook
- Both SGLang and Miles offer Day 0 support
More from Infra
- ClickHouse aggregates 100B rows in 0.4s with parallel replicas auto-scaling — Abhishekcur · 2026-10-18
- Auction-based LLM routing paper finds fixed token pricing wastes 10%-71% in costs — _onionesque · 2026-10-18
- Running Ling 3.1 Flash (560B MoE, 25B Active) Across 4 DGX Sparks via llama.cpp RPC — BerryBlush_Rogers691 · 2026-10-18
- SGLang Summit sponsors jump from 11 to 36 in one week — ying11231 · 2026-10-18
- Oracle CDS hits record 300bps, called first AI trade domino after Barclays sell downgrade — arthurcolle · 2026-10-18
- Agent Coding Workflow Would Cost ~$40k/Month on API Pricing, Dev Reveals — ChrisRackauckas · 2026-10-18