To take it a step further, I work adjacent to OEM manufacturers for industrial control systems. Half of our customers are running windows 98.
Hah! I’d actually consider industrial automation to be one of the more-likely applications for LLMs. Not necessarily LLMs driving the system, but outputting software to drive PLCs and similar. My understanding from past reading is that industrial automation is one of the principal things that China is interested in using LLMs for.
Down on the shop floor, industrial AI agents are currently running most solidly in four types of scenarios.
Quality inspection. Vision large models are replacing traditional rule-based machine vision. Among the 12 lighthouse factories newly added by the World Economic Forum in September 2025, AI and generative AI use cases account for up to 50%, with demonstration factories reporting results including a 41% reduction in product defect rates.
Process parameter optimization. In the Baosteel hot-rolling line case disclosed by Huawei Cloud, the Huawei Pangu large model improved finishing-mill width-spread prediction accuracy by more than 5% and raised yield by 0.5%, which by its own account translates into more than 90 million yuan in additional annual revenue. It should be noted that such figures are all vendor-disclosed and still lack independent third-party verification, but directionally they are consistent with public data from lighthouse factories.
Engineering and maintenance assistance. Siemens Industrial Copilot improves the efficiency of generating programmable logic controller code by roughly 60%, and pilots of its maintenance edition show an average 25% saving in reactive maintenance time; the product won the Hermes Award at Hannover Messe 2025. The factory operations agent Microsoft launched at Hannover Messe 2025 lets employees ask in natural language, “Why did Line 3 stop yesterday?”
Scheduling and order management. This is the entry point closest to AI agents for small and mid-sized factories. An AI tool aimed at small and mid-sized factories has disclosed a set of figures: after adopting AI scheduling and order tracking, factories’ average order delay rate fell from 25% to within 5%; one factory in Dongguan cut its order-splitting preparation time by 60%; and an annual fee of about 10,000 yuan corresponded to a return of roughly 1:12. These are again vendor figures, but they touch on an important fact—the value of AI agents for small and mid-sized factories lies first not in sophisticated process optimization, but in freeing up the labor tied down by high-frequency, tedious work such as order tracking and scheduling.
They mention PLC code generation, but I guess some other areas too.
It’s not that you can’t use LLM’s in the industrial space. It’s that there’s not much, in my experience, demand to do so.
I have worked with a food packaging plant (one of those Window’s 98 places). They are still using RSLinx from like, decades ago. Why would they change it? The PLC code does not need to be updated except under specific circumstances (which is why we use ladder logic in the first place).
New stuff may use it, but a company creating a new industrial product doesn’t really translate into that product being purchased.
This is something I think software people tend to struggle with. In “traditional” engineering you can’t just deploy something to prod. Projects take years for approval, years of testing, years to go through the political process, years to build. The things you see getting built today were being designed 10 years ago using technology from 40 years ago.
Design is normally the first thing to happen before the rest of the above does and its the fastest part of the process already. Because of this big projects are allergic as fuck to new technology. I have watched proposals for a more advanced high-pressure CO2 turbine system get knocked down purely because “it’s not what we’re used to.” That “new fangled design” was roughly 30 years old at that point. Nuclear plant’s just started implementing general purpose microprocessors.
I really believe a lot of Lemmy would benefit from spending a turn working on, like, nuclear commissioning or substation work. Software tends to compress peoples idea of a what a reasonable timeline for going from idea -> proposal -> design -> implementation is.
I’m not saying that eventually AI won’t be used but we have a understand that most of the time the customers have no interest in adding anything.
Hah! I’d actually consider industrial automation to be one of the more-likely applications for LLMs. Not necessarily LLMs driving the system, but outputting software to drive PLCs and similar. My understanding from past reading is that industrial automation is one of the principal things that China is interested in using LLMs for.
searches
https://faxiangongchang.com/en/reports/china-industrial-ai-agent-2026
They mention PLC code generation, but I guess some other areas too.
It’s not that you can’t use LLM’s in the industrial space. It’s that there’s not much, in my experience, demand to do so.
I have worked with a food packaging plant (one of those Window’s 98 places). They are still using RSLinx from like, decades ago. Why would they change it? The PLC code does not need to be updated except under specific circumstances (which is why we use ladder logic in the first place).
New stuff may use it, but a company creating a new industrial product doesn’t really translate into that product being purchased.
This is something I think software people tend to struggle with. In “traditional” engineering you can’t just deploy something to prod. Projects take years for approval, years of testing, years to go through the political process, years to build. The things you see getting built today were being designed 10 years ago using technology from 40 years ago.
Design is normally the first thing to happen before the rest of the above does and its the fastest part of the process already. Because of this big projects are allergic as fuck to new technology. I have watched proposals for a more advanced high-pressure CO2 turbine system get knocked down purely because “it’s not what we’re used to.” That “new fangled design” was roughly 30 years old at that point. Nuclear plant’s just started implementing general purpose microprocessors.
I really believe a lot of Lemmy would benefit from spending a turn working on, like, nuclear commissioning or substation work. Software tends to compress peoples idea of a what a reasonable timeline for going from idea -> proposal -> design -> implementation is.
I’m not saying that eventually AI won’t be used but we have a understand that most of the time the customers have no interest in adding anything.