Author

Christophe Gerardin

C/C++ developer, networking & inference infrastructure

Christophe Gerardin is a C/C++ developer working on networks and infrastructure: how to architect an inference platform, serve models, and keep the whole thing standing (throughput, latency, topology) at scale.

His background is rooted in the cybersecurity of industrial control systems (ICS/OT), but that is not the angle he takes on LeCompute: here he dissects AI compute infrastructure through networking and service architecture, where the performance and robustness of a deployment are actually decided.

  • Networking
  • AI infrastructure
  • Model architecture & serving
  • C / C++
  • ICS/OT cybersecurity (background)

Articles by Christophe in English 4

Runtimes

GGUF vs GPTQ vs AWQ: a format, two algorithms, one problem

GGUF is a file format, GPTQ and AWQ are algorithms. At 4 bits, each attacks the outlier problem differently. Anatomy of three constructions, from GPTQ's Hessian to GGUF's k-quants, and of what the choice of format decides for you.

19 min read
  • GGUF
  • GPTQ
  • AWQ
  • Quantization

Costs

Stargate, Spud and subscriptions: the compute lever behind OpenAI's comeback

10 gigawatts 'secured years ahead of schedule', 0.3 gigawatt plugged in at Abilene. Behind OpenAI's comeback in AI coding, a compute bet that must be read in three states: the secured, the built, the plugged-in. And a subscription economics that agents cracked open.

18 min read
  • Stargate
  • OpenAI
  • Compute
  • Datacenter

Runtimes

Nobody knows how to measure coding agents anymore

On July 8, 2026, OpenAI disavowed SWE-Bench Pro: ~30% broken tasks. It is the third reference instrument declared dead in three generations. Anatomy of a metrology crisis: noise ceiling, contamination, and a harness effect worth a whole model generation.

17 min read
  • Benchmarks
  • SWE-bench
  • Coding agents
  • Evaluation

Costs

GPT-5.6 and the Codex merger: anatomy of OpenAI's comeback

On July 8, 2026, OpenAI disavowed the reference benchmark for agentic coding. On the 9th, it merged Codex into the ChatGPT app and launched GPT-5.6. Behind the sequence, a comeback that turns less on intelligence than on prices, subscriptions and the harness.

19 min read
  • GPT-5.6
  • Codex
  • Claude Code
  • OpenAI