Large Language Models (LLMs) Business Opportunities
As a legal background person, LLM, known as a Master of Laws ( M.L. or LL.M.; Latin: Magister Legum or Legum Magister), has been an advanced postgraduate academic degree for a very long time in my life. Recently (Yes, just 2023 Q1), Large Languages Models(LLMs) have become more popular; the necessary abbreviation to represent LLM for me is even Google Search.
This article uses simple language to help you understand the current trends and is a guide to brainstorming new business opportunities.
Firstly, what are the Large Language Models (LLMs)? I’ve asked ChatGPT to “simply” explain for me as below,
Large language models, like the one I am, are advanced computer programs that have been trained on vast amounts of text data. These models are designed to understand and generate human-like language. They work by analyzing patterns in the training data and learning the relationships between words, phrases, and sentences.
Large language models can perform a variety of language-related tasks, such as answering questions, generating text, providing explanations, offering suggestions, and even engaging in conversations. They can understand and respond to input in a way that is contextually relevant and coherent.
To train a large language model, requires immense computational power and data. The training process involves exposing the model to massive amounts of text from books, articles, websites, and other sources. Through this exposure, the model learns to predict the most probable next word or sequence of words based on the context it has seen during training.
Once trained, the model can generate human-like responses by predicting the most appropriate words or phrases based on the input it receives. However, it’s important to note that while large language models like me can generate impressive and coherent text, they are not conscious or truly understanding the meaning behind the words. They operate based on statistical patterns learned from the training data.
Overall, large language models have a wide range of applications and can assist with tasks that involve natural language processing and understanding. They continue to evolve and improve, enabling more advanced and nuanced interactions with users like yourself.
If you are tired of reading the long paragraph, there are three key elements of LLMs as below,
- Creation — generate human-like language
- Assist — can do language-related tasks such as answering questions, explaining, and offering suggestions.
- Interaction — communicate with you
Do you catch the most crucial information about LLMs? It is “Language,” which means words, texts, and characteristics. Please keep in mind that we will leverage everything by words.
If I were only thinking about ChatGPT (1+2), I would tell myself I am outdated (Seriously!). What is the current trend based on LLMs? How do we leverage the LLMs to build the new business (3–7)? There are a few things I would like to share with you below,
- Application
Which scenario would you like to apply to? How do you interact with the users? What solutions do you want to implement? For example, I would like to offer the AI image generator by users texting the keywords they want and generating the image for them. The input is words, but the output is images.
2. Conversion:
Digital information, including pictures, sounds, videos, and so on, EVERYTHING will convert to WORDS to interact with LLMs. Another example is the Chrome Extension for summarizing the youtube video. The input is video, but it is the subtitles or voice recognition converting the voice to the words. The output is AI, which offers you a summary of large amounts of words and paragraphs.
3. Embeddings, Fine-tunes:
To offer practical and high-quality information, you must build up your embeddings and fine-tunes the model, which means when you get input from LLMs (4) and output to LLMs again (5). For example, in the AI language tutor solution, when the user text the wrong grammar to chat with AI, usually the AI will try to understand the meanings and reply to the user directly. But if you have your tunned model, you can ask LLMs to respond to the user simultaneously to correct the grammar for users.
4. Experience Design:
If you have experience using ChatGPT, you understand it is like Chatbot. To think about the scenario you will apply to, the output you will deliver to the users, the “Experience” will be the key to satisfying the user expectation. Although everyone knows this is AI, how you show it is still essential. For example, the AI image generator offers you images based on the words you texted, the layout of showing pictures, and filtering features like the colors, sizes, and objectives could improve better results to improve the user experience for the users.
This article helps you think of a new business opportunity. With several years of experience as an entrepreneur, things never change are:
- Forming a solid and diverse team (business, development, design, UX)
- Build your unique tunning model and
- Design a seamless user experience.
- Do it Fast!
- Embrace the AI
Special thanks to my friend Flavien MALGERARD for generating this AI image and Hau Le for providing insight for this article.
#AI #llm #largelanguagemodels #futureofai #businessopportunity #embeddings #Finetunes #userexperience #ux #uxdesign
