OpenFaaS bought 4x DGX Sparks for air-gapped coding and red-teaming
natesiggard · x · 2026-10-19
OpenFaaS founder Alex Ellis details why and how the team bought 4x DGX Sparks for local AI.
Privacy over cost
- Every cloud token goes to a provider; many retain inputs/outputs for training, may keep data up to 30 days even when opted out, and flagged ToS-violation chats can be retained indefinitely with possible account bans
- Local deployment lets them support air-gapped regulated customers and red-team their own products without safety lectures or bans
Lessons learned
- Previously ran Qwen 27B on an RTX 6000 Pro: small dense models are like an attack dog—focused but needing careful handling; larger models (DeepSeek, GLM) explore problems unsupervised with bigger context at slower rates
- First two units were linked via a Connect-X cable; buying two more was painful to justify but unlocked capacity no single consumer GPU could hold in memory
More from Infra
- 4 free CUDA resources to learn GPU programming from scratch, from NVIDIA to Oxford — techNmak · 2026-10-19
- Running Moss image model on a 4GB RTX 3050, optimized down to 62s per image — sanketshinde2004 · 2026-10-19
- Barclays: pure-play data center bond spreads hit ~209bps as market prices in contagion — SumitGup · 2026-10-19
- GPU programmer says the field needs lighter, more predictable inference engines — brandon_xyzw · 2026-10-19
- llama-doctor: CLI that diagnoses your llama.cpp server from /metrics with severity-graded findings — msalsas · 2026-10-19
- Underwater AI data center planned in Alaska's Cook Inlet, home to 330 endangered belugas — anselm · 2026-10-19