Because machine builders will not paste PLC code into a cloud chatbot, Matt Kurantowicz ran a free model on one GPU with Ollama, no network and no API key. It wrote structured text locally and launched CODESYS through a local MCP client, though full project creation still failed. The offline path is what makes AI usable on controls IP.
ENFORCE is an MIT-licensed PyTorch framework from TU Delft that adds a projection layer so a neural network's predictions satisfy equality and inequality constraints, such as mass balances. Artur Schweidtmann shows a distillation model that scores well while inventing material, which ENFORCE prevents. Shell supported the research. For process models, hard constraints beat hoping the network learned conservation.
For process engineers and data scientists building surrogate models of plants.
Found via Artur Schweidtmann on LinkedIn
Teardown is an MIT Claude Code skill that builds an offline HTML 3D simulator of a real machine from real dimensions, with live gauges and buttons that isolate each part. Three examples ship in the repo (V8, Cessna landing gear, Strandbeest). Useful for maintenance trainers and field techs who need a walk-through model of a pump or gearbox without buying simulation software.
For maintenance trainers, reliability engineers, and field service teams who explain equipment to new hires.
Found via @elvar_official on X
The Globe and Mail reports Anthropic is negotiating to lease Beacon Data Centers' Sturgeon County campus near Edmonton. Alberta approved a co-located 920 MW gas plant on Oct 23: 200 Jenbacher engines in 40 modules, construction expected this quarter, in service late 2027. No lease is signed. Another AI load pairing the plant with the campus.