July 2023
Two Explosions š„š„
You might think that these rather exclusive updates include everything weāre working on. Of course, thatās not the case. We have some aces up our sleeve š
Ā
Two of them dropped this week, and we couldnāt be prouder. You will have heard of them, but just to make sure, let me tell you again:
Ā
š„3DGPTš„
This is our first experiment with Large Language Models. We wanted to see whether we could combine existing knowledge (in this case 12,000 standards and journal articles) with one of the large language models that have been all the rage over the past few months. This experimentās goal was:Ā
To provide something meaningful to the industry: In this case, a specialized Chatbot that could answer more in-depth questions about the additive industry.
To understand the limits and opportunities of LLMās. In fact, we failed with our first attempt (using an open source LLM hosted by us). The second version worked much better. The process taught us a lot about the current state of the art in LLMās, how to work with them, and the results are teaching us what limits of the models are.
Ā
Ultimately, we obviously think thereās a big opportunity in the use of natural language to help guide more effective processes. This was a first step, and a successful one ā more than 1000 signups and 1500 questions. To find out more check out ourĀ Press ReleaseĀ or give it a try atĀ www.authentise.com/3dgpt. But hurry ā it wonāt stay up forever.Ā
Ā
š„Threadsš„
In a way, the second release is related to GPT.Ā ThreadsĀ is the fulfilment of a decade of learning, and more than a year of incredible work by the team. Itās an engineering collaboration platform that brings together chat and project management, together with resource sharing. Itās focused on those ad-hoc conversations, with ad-hoc teams: R&D, Non-conformance root cause analysis, Field Services.Ā
Ā
These types of unstructured issues are usually solved in calls or meetings. Rarely is the context fully captured. Thatās because the tools currently available are either really rigid (e.g. PLM) or not traceable (email, whiteboard, etc.). Communication tools (Slack etc.) don't fit the project management need and project management tools (Jira etc.) don't foster collaboration. What was needed was a tool that would make solving ad-hoc challenges a breeze for multi-stakeholder teams.Ā
Ā
Thatās what Threads is there to solve. As you can see from theĀ Press Release, itās showing effect: Within 2 weeks, one early test team from a medical company were seeing a 1.5x ROI on their investment, tracking 100% of their R&D decisions digitally, while saving 150 hours and 20 meetings across a distributed team including external partners. They doubled the effective size of their team.
Ā
Threads is the big new deal for us. It solves a significant gap in the Digital Thread, allows us to start covering the whole Idea-to-Part process with our workflows and is a significant step for us to take our experience from additive to manufacturing in general. Weāre super excited for the future, and not only because we already have plans on how to integrate Large Language Models into the chat features, now that we know how to use them.
Ā
If youāre interested in learning more about Threads, check out theĀ websiteĀ and get in touch if you think it can help you.

There are a number of other exciting projects that are bubbling along in the background ā and I canāt wait to share them. Alas, Iāll have to be patient. And so do you š
Ā
Enjoy summer now that itās properly started. Hope to talk with you soon.