Over the past 18 months, large language models (LLMs) have made tremendous strides, and moreover, they are constantly improving. These improvements are sometimes locked behind cost walls, such as GPT-4 and DALL-E, but new tools have also become available. LLMs are now available for everything from AI content creation, image processing, video editing, and more, as well as a range of software creation and debugging tools. The pace of change in AI today is incredible, so it's easy to miss opportunities where AI can save you time, money, and eliminate hassle. Let's look at how you can get the most out of these tools!
1. Use LLM to create design prototypes faster.
Image creation tools like DALL-E are very impressive.
Text-to-image conversion tools have evolved significantly and are getting better and better, so there are many ways to use them. You can create an entire image from scratch (although it might look a bit "odd") or use it to create a specific part of the image that you can then further edit. They're also great for creating sketches, outlines, and prototypes, which you can then interactively adjust using other prompts.
While there are free image creation tools, if you're a creative professional, you may have paid for or have access to such a tool with Adobe Firefly . Free image creation tools can be very good, such as the image creation tool powered by Microsoft Designer's Copilot, but sometimes their quality is poor or they have a distinctive "AI" look. DALL-E is included in the ChatGPT Plus subscription and is one of the best available. Google Gemini can also create images, although there are limitations put in place after some recent controversy.
Different quiz creation tools often have a distinct "feel" or aesthetic, and are capable of doing different things, so you should try a few or look at some samples online before signing up (or continue using the free option you prefer!).
2. Use LLM to learn new things.
You can ask an LLM even your most awkward questions.
The educational aspect of LLMs is an underrated advantage. They truly are excellent study partners. You can ask LLMs questions on any topic, whether it's a brief summary or a detailed explanation – just ask them why, how, or what makes something you're learning possible. They're also great for creating flashcards, practice questions, and quizzes on a topic. For example, you could ask ChatGPT to create a short quiz on a topic you're studying but ask it to skip the answer choices. Then, write your answers in the hint box, and then ask ChatGPT to grade you.
3. Use LLM to create a logo or brand.
LLM can be very useful for the finished product or as inspiration.
LLMs can provide inspiration as they create custom visuals to represent your brand. Logos and branding inherently follow patterns, because every brand wants to look unique and recognizable in its field, so they tend to follow design trends. By describing what you're looking for (e.g., logo, business card) and what your brand represents (e.g., mission statement, core values), you can at least gather some fresh ideas to consider!
Providing as much detail as possible and clarifying what you're looking for will be very helpful. This could include specific color, text, or style preferences in your prompt.
4. Use LLM to create templates for other software.
ChatGPT Plus (on GPT-4) was able to generate an Excel document budget template, exporting both the document file and the Python code it used to create it. While the formatting isn't overly complex, the output could easily be improved by adding some minor changes or more specific instructions in the prompt.
This, of course, can also be extended beyond the Office suite. LLMs like Gemini or ChatGPT are very good at creating templates for almost any application that accepts templates. Because they create templates using scripts, they are not limited by file types. As long as they can write some code to build the template, they will get the job done.
5. Use LLM to compile, debug, and generate code.
LLMs are surprisingly good at debugging code problems, and even better at creating code.
This is fairly clear: LLMs are excellent at debugging code. They are still prone to errors in larger projects, and can easily go astray when underestimating a feature or functionality that doesn't actually exist, but for debugging, they are superb. To get the best results when debugging code with an LLM, provide as much detail as possible, including any specific error messages. However, be careful here, as LLMs often store all prompts and data sent so they can train other models. So, if you are working in a professional environment or using proprietary information, be sure not to provide that information to the LLM.
Creative ways to use LLM can help you save time and effort.
LLMs are becoming increasingly powerful, and like any powerful tool, we'll need time to learn how to use them properly. This article has focused on what you can do with some of the major LLMs currently available (ChatGPT and Gemini), but as more specialized tools are released, there will be even more creative possibilities for using LLMs. In the meantime, we can still enjoy finding ways to maximize the obvious potential of these amazing tools.