Kernel Panic conference in Madrid packs house, slides from four production-AI talks published
SergioPaniego · x · 2026-10-13
The first Kernel Panic conference in Madrid filled every seat with four talks on running open AI at scale in production, slides now public: Inditex Tech on Kumoss (their developer-facing AI infra platform), Hugging Face on training a coding agent through a harness you didn't write, Cloudflare on rolling out agents to every employee, and helmcode on scaling inference for 1,000 users.
More from coding & agent
- Hooking an indexed corpus to a coding agent cut API validation costs 43% to $1.20 — JaynitMakwana · 2026-10-14
- Spidermonkey trims 10% agent steps on weak models, no gain on strong ones — mathemagic1an · 2026-10-14
- Loop-based cognitive architecture boosts cheap coding agents' token efficiency, but not Opus — mathemagic1an · 2026-10-14
- Hankweave: open-source framework for recursive self-optimization under $1000 — hrishioa · 2026-10-14
- Replicas raises $4M seed for harness-agnostic cloud agents, backed by YC and Paul Graham — KlausCodes · 2026-10-14
- Ranked 1v1s for Claude Code: race strangers to ship, with Claude as referee — PrajwalTomar_ · 2026-10-14