Once again, over 20 passionate LabVIEW devotees turned up for WUELUG27. This time our topic was very contemporary: how can we integrate LLMs and similar AI tools into our development workflows? Last time we focused on the internal capabilities of the LabVIEW ecosystem, this time we looked at the future!
WUELUG27 Outline
Our plan was to carry out the G-AI and Claude Desktop installations together as a group, so that everyone could run their own experiments directly on their own laptops. Unfortunately, Jan was held up a little, so Jörg had the pleasure of guiding everyone through the first steps of the installation. Thankfully, most of the participants had already got everything ready, and Jan arrived just in time (just before Jörg was at his wit’s end) to start the actual presentation.
Jan Göbel works at NI/Emerson, and he created the open-source G-AI toolkit, in addition to an MCP server for LabVIEW.
We learnt:
- Which AI models are currently available, what their advantages and disadvantages are, how tokens work, and more,
- What an MCP server is and how it can be used to connect an application to an LLM in general,
- How to use Jan’s LabVIEW MCP server and G-AI toolkit to connect LabVIEW to Claude AI,
- How G-AI works and how you can implement your own functions to extend G-AI.
G-AI is currently just an experiment with a few limited functions. Nevertheless, we have already managed to generate some very simple LabVIEW code. Everyone was invited to support this project and add own functionality. Maybe, at one day, we will be able to generate working VIs with it.
Where do we stand today?
Naturally, at the end of the event, the question arose as to how relevant AI is for us as engineers. We asked the group to raise their hands to indicate how they rated the relevance of AI in their own day-to-day work over the next 12 months:
Will AI be helpful in our job over the next 12 months?
Will AI be particularly useful for the NI tool chain (LabVIEW, etc.) over the next 12 months (both Nigel and general-purpose LLMs such as ChatGPT, etc.)?
Will AI be useful for LabVIEW code generation over the next 12 months?
The Future
This LUG was extremely eye-opening. For most of us, this was the first time that AI-assisted LabVIEW development actually became tangible. The possibilities and benefits are promising; we definitely will keep track of the latest news and possibly even contribute ourselves.
The next meeting – WUELUG28 – will tackle the second major issue of our time: cybersecurity! As we are currently investing a great deal of energy in this area, #teamhampelsoft would like to offer to host the evening and show you what we’re up to in the field of cybersecurity. We’d also be happy to share what we learnt at NI Connect in May.
WUELUG has now been running for almost 8 years – or 27 meetings – and we have grown into a wonderful community. The natural next step is clearly to hand over the organisation to a committee, so that the well-being of our LUG does not depend on a single person or a single company.
That is why I am particularly delighted that Ludwig Berndt (Siemens) and Oli Wachno (Ingenieurbüro Wachno) will be joining us in organising the meetings in future. Huge thanks to them! And a big thanks to NI/Emerson for once again sponsoring our drinks and snacks!
WUELUG28 will take place in July 2027. Details to follow! Join in the conversation in the forums!
WUELUG28 NI Forums Thread