ChatGPT was taking the industry by storm for the last few months. Unless you live under a rock, I’m sure you had a chance to read about it already. I took some time to digest the news and share my experience and thoughts about Chat GPT and underlying technology used in a few other products doing work with generative AI technology – Jasper (Copywriting), Stability AI (Visual art), DoNotPay (Legal Services), Omnekey (Creative Content), Paige.ai (Cancer diagnostics) and Mostly.ai (Synthetic data).
In this post, I will explore what is ChatGPT and its underlying technology and how it can further maximize efficiencies within the design, engineering, and manufacturing environment. What are the potential of harnessing ChatGPT/LLM technologies in the enterprise and PLM domain? What benefits it can bring? Stay with me an learn more.
What is it all about?
Here is a self-generated description of what is ChatGPT (generated by Chat GPT). Open AI is a company founded in 2015 with early investors Elon Mask, Peter Thiel, and later Microsoft, which provided Open AI to use Azure infrastructure for their work. Open AI is using GPT-3.5family of large language models (LLMs) and it is fine-tuned with both supervised and reinforcement learning techniques, which create what is known as Chat GPT. Check more Large Language Models.
ChatGPT is a variant of the GPT (Generative Pre-trained Transformer) language model developed by OpenAI. It is a neural network-based language model that uses machine learning to generate natural language text. It is trained on a large dataset of human-generated text, such as books, articles, and websites, and can be fine-tuned on a specific task, such as answering questions or writing creative text.
ChatGPT is designed to generate human-like text for conversational applications, such as chatbots, virtual assistants, and conversational interfaces. It can be used to generate responses to user inputs, such as questions or commands, and can be integrated into various types of applications, such as mobile apps, websites, and voice assistants.
The model is pre-trained and can be fine-tuned on a specific task and custom dataset to improve performance and generate more accurate and relevant responses.
The report also explains why the AI bubble will inevitably burst as report author Philip Hunter, Rethink Research fellow, pointed out: “Huge sums have been invested in AI with the upward trend still accelerating through 2018. Global Venture Capital (VC) based investments alone have risen from $3.2bn in 2014 to $12bn in 2017 while the number of funding rounds per year for AI startups doubled to around 1,300 over that period. The total invested globally in AI during 2018 alone amounted to over $100bn taking account of money spent by governments and big corporations as well as VC funding of start-ups.”
This is more than double the expected annual return from AI even by 2023 and so there is no way this is going to generate a return on investment over the forecast period, Hunter added. “Investors will begin rolling up start-ups which fail to generate revenues into others which show promise during 2019,” said Hunter.
“The only way AI start-ups have made money so far is from being acquired rather than selling products or services. Valuations have been based purely on the assessment of the people working for the company, often at as much as $10m a head.
ChatGPT – technology and applications
There are a lot of articles today you can find summarize what number of applications and use cases. Outside of engineering, CAD, PLM, and enterprise domain, you can find a waterfall of applications using Chat GPT for something that it can do the best – creating summaries, generating custom-made advertising, and creating marketing content. I also found some examples of curiosity about what Chat GPT can and cannot do. Here are a few examples I want to refer you to, that I found useful.
I agree with Jos Voskuil’s conclusion that was impressed how Chat GPT was able to generate useful portions of writing for his blog but missed some important points and, of course far from the capabilities to replace him as a PLM consultant.
It was an exciting exercise to combine my blogging thoughts with the answers from OpenAI. I am impressed by the given answers, knowing that the topics discussed about PLM are not obvious. On the other hand, I am not worried that AI will take over the job of the PLM consultant. As I mentioned before, the difference between Explicit Knowledge and Tacit Knowledge is clear, and business transformations will largely depend on the usage of Tacit knowledge.
Let’s move to enterprise applications and how Generative AI, the large language model can be used for computer-aided design, product lifecycle, PLM system, and product development process. What are the possible applications of Chat GPT for PLM? Will it open a new way of gathering data, improving data quality, and providing an artificial intelligence foundation for PLM solutions? IDC research article is probably the best I found that speaks about Generative AI. Check this out – Generative AI and what does it mean for the enterprise? It covers multiple application domains – code generation, enterprise content management, marketing, and customer experience applications. It has also a paragraph speaking about the PLM domain. Here is a passage:
Product Design & Engineering – It will also affect technologies in the product lifecycle management (PLM) and innovation space with the likes of Autodesk, Dassault Systemes, Siemens, PTC and Ansys continuing to build capabilities to enable design engineers & R&D teams to automate and expand the ideation and optioning process during early-stage product design, simulation, & development. Generative AI design would allow options for engineering and R&D teams to consider in terms of structure, materials, and optimal manufacturing/production tooling. For example, it would potentially suggest a part design that optimizes against factors like cost, load bearing, and weight. Generative design can also enable reimagining of product look and feel, often resulting in unique aesthetics and form that is not only more compelling to end users, but more practical and environmentally sustainable. Many of these vendors have attached their generative design offerings to additive manufacturing capabilities that are needed to realize these unique products. Opportunities exist across multiple industries for generative design. Automotive, aerospace, and machinery organizations can improve product quality, sustainability, and success, while life sciences, healthcare, and consumer products companies can improve patient outcomes and customer experiences.
LLM and GPT applications in Engineering, Design, and Product Lifecycle Management
Large Language Models is a fascinating and interesting topic that drives a lot of discussion and research. Here are three questions that come to my mind when I think about the possible applications.
- Can LLM be pre-training with the engineering and design knowledge to be capable to product new designs (eg. mechanical, electronics, etc.)
- Can we use LLM for automation tasks such as order / RFQ processing and placing orders after analyzing manufacturing capacity?
- How can we combine Knowledge graphs with large language models and use it in product data management, supply chain management, and other tasks related to product lifecycle?
I was not alone to ask these questions and found interesting answers. Check some articles and examples I was reading – LLM for pharma and finance, How can I drive automation with LLMs, What are LLM are used for. Here is an interesting passage that gives you a way to think about the application of LLM in future design and engineering.
Language is used for more than human communication. Code is the language of computers. Protein and molecular sequences are the language of biology. Large language models can be applied to such languages or scenarios in which communication of different types is needed. These models broaden AI’s reach across industries and enterprises, and are expected to enable a new wave of research, creativity and productivity, as they can help to generate complex solutions for the world’s toughest problems. Large language models are also helping to create reimagined search engines, tutoring chatbots, composition tools for songs, poems, stories and marketing materials, and more.
Large language models such as GPT-3 are becoming increasingly popular in enterprise applications. The peak of the interest is because of their ability to understand and generate text with a high level of fluency and accuracy. Which immediately positions it in the domains such as content generation (eg. tech documentation creation), customer service (support and troubleshooting), business intelligence (to analyze large volumes of data – eg .customer reviews), translation (eg. multi-language support) and virtual assistance (help operate machines). However, it’s important to remember that these applications are still in development. So, maturity is a big question. LLM and specifically GPT-3 is a big achievement. We are at the very beginning of introducing applications of LLM to enterprises. But it is very exciting. So stay tuned. Just my thoughts…
Disclaimer: I’m co-founder and CEO of OpenBOM developing a digital cloud-native PDM & PLM platform that manages product data and connects manufacturers, construction companies, and their supply chain networks. My opinion can be unintentionally biased.