The above question is one we get asked often by our customers.
So, let’s explore each of these topics below.
What is a Large Language Model (LLM)
An LLM is a type of artificial intelligence model that has been trained on a vast amount of text data. The purpose of an LLM is to generate human-like text that’s relevant to the input it’s given. These models can answer questions, write essays, summarise texts, translate languages, and even generate creative content like poetry or story writing.
What is ChatGPT
ChatGPT is an example of an LLM. It’s a variant of the Generative Pretrained Transformer (GPT) model developed by OpenAI, specifically designed for conversational purposes.
The GPT series, including GPT-4, are large language models trained on a vast amount of internet text. They’ve been trained to predict the next word in a sentence given all the previous words, and this ability can be used to generate human-like text.
ChatGPT, specifically, has been fine-tuned with additional training on a dataset of conversations. This fine-tuning process allows the model to specialise in generating conversational responses. So when you ask a question or make a statement, ChatGPT generates a response that tries to continue the conversation in a relevant and coherent way.
ChatGPT does not have real-time access to the internet or updated information beyond its training data. GPT-4 has a training cut-off date of September 2021.
Enter ChatGPT Plugins
ChatGPT has recently released plugins which are tools designed specifically for language models, with safety as a core principle.
These plugins can help ChatGPT access up-to-date information, run computations, or use third-party services.
The plugins serve as the “eyes and ears” for the language models, providing access to information that may be too recent, too personal, or too specific to be included in the training data.
In response to a user’s explicit request, plugins can enable language models to perform safe, constrained actions on their behalf, increasing the overall usefulness of the system.
So, think of a plugin as a bridge between an LLM and your data.
An example of a ChatGPT Plugin
A good example is a ChatGPT plugin that we created for a customer of ours called InventTour.
InventTour is a central inventory and marketing platform that allows the travel industry to efficiently communicate, update and distribute information. It connects Africa’s suppliers with global agents and operators.
The InventTour ChatGPT plugin allows a user to get access to InventTour’s comprehensive and dynamic accommodation inventory directly in ChatGPT, via a plugin.
So, for example a user could enable the InventTour plugin and ask the following question to ChatGPT:
“Generate me an itinerary for a 7-day trip to Botswana and Zimbabwe and include accommodation options.”
Let’s see it in action.
What is a Bing Chat
Bing Chat allows users to interact with an AI chatbot through Bing’s search engine. The system is designed to be conversational, making it easier for users to get information rather than simply filling out search queries. Although Bing Chat is distinct from ChatGPT, it’s powered by the same model from OpenAI, operating in a similar way but is deemed more accessible and user-friendly.
Similar to ChatGPT, Bing Chat also have plugins.
Bing Chat plugins are third-party tools designed to enhance the capabilities of Bing Chat and they allow you to create and integrate your own tools into Bing Chat.
What is Microsoft Copilot
Microsoft are in the process of releasing a number of different copilots for Microsoft 365 and Windows.
These copilots leverage the power of LLMs and combines this with user-specific data within the Microsoft Graph, such as calendar, emails, chats, documents, meetings, and more. It can perform tasks such as drafting emails, creating content based on user prompts, and assisting with other productivity tasks within the Microsoft 365 or Windows environments.
It’s designed to work alongside users, embedded in the apps and operating systems that they use every day, with an emphasis on unleashing creativity, unlocking productivity, and upleveling skills.
So can a copilot have a plugin
Plugins are tools that help copilots interact with other software and services, making them more useful.
So in the case of a Microsoft 365 or Windows Copilot, a plugin would allow them to get access to other data that sits outside the Microsoft 365 or Windows ecosystem.
So think of a plugin as a bridge between a copilot and your data.
The difference, summarised
So while ChatGPT, Bing Chat and Microsoft Copilot are all based on LLMs, the primary difference lies in the way they interact with user data and their intended use cases.
ChatGPT and Bing Chat are more conversational AI tools, while Microsoft Copilot is designed as a productivity tool around the Microsoft apps: Outlook, Word, Excel, PowerPoint, SharePoint, OneDrive, OneNote, Teams, and Exchange.
ChatGPT and Bing Chat are both available to the general public, while a Microsoft Copilot would generally be used inside your organisation.
So, do I need a plugin for my business
Whether you need a plugin for your business depends on your specific needs and goals but more often than not, the answer would be yes.
To harness the true power of using an LLM, you need to combine it with your business data.
Your business data could be stored in a database, in a file store, or even a custom or 3rd party application.
So, whether you want to combine an LLM with your business data publicly in something like ChatGPT or Bing Chat, or if you want to combine it internally in your organisation through a Microsoft Copilot then having a plugin will allow you to do this.
So don’t delay and reach out to the rapidMATION team today to discuss how we can help you harness the true power of using a LLM and combine it with your business data.