Tired of having to specify your preferences each time you interact with an AI? Longed for a more personalized AI experience? Your wishes have just come true!
Welcome to Tech Trend! I’m Olivia, your guide today, introducing you to an exciting new feature from ChatGPT called Custom Instructions.
Custom Instructions is OpenAI’s initiative to offer users more control over the responses ChatGPT generates. Currently, this feature is exclusively available to Plus Plan users but will soon be extended to everyone.
Set your preferences once and observe how ChatGPT utilizes them for all following interactions. Imagine a teacher creating lesson plans without the need to specify each time that they teach third-grade science, or a programmer who likes concise code in a language other than Python stating it just once, or the simple case of doing the groceries for a big family. ChatGPT will automatically factor in servings for six when creating a shopping list.
While European users may not have access to this feature yet, there’s a workaround by using a VPN. Immediate access is achievable even with a European OpenAI account. A handy tip for those keen to try it out right away.
Better still, this feature is expected to be globally available soon, hitting Europe in the days or weeks to come.
In this session, I’ll share my early experiences with Custom Instructions and envision the future benefits. As it becomes available, look for it under the beta features in the settings, where you can find other plugins like the code interpreter that I’ve previously introduced.
Remember, new features like Custom Instructions won’t trigger alerts, so keep checking back and activate when ready.
After activation, you’ll see a new Custom Instructions option in the settings. It has two fields: one for personal information to help ChatGPT personalize its prompts, and another to dictate ChatGPT’s response style and tone.
In the first field, you can input details like location, job, or hobbies. So when asked, ‘What should ChatGPT know about you for better responses?’, you decide what’s important.
In the second field, you shape how ChatGPT should respond: formal or casual, brief or detailed. Who’s the intended audience? These choices help ChatGPT give more personalized and effective responses.
You can also combine Custom Instructions with other plugins. For instance, if you tell it that you live in Europe, map or geolocation plugins can use that information to offer personalized suggestions, such as the best local restaurants.
To illustrate, let’s use two examples of Custom Instructions. Let’s try this out by filling both fields. Feel free to pause the video here to better follow along.
In the first field, I enter: ‘My name is Olivia. I’m a data scientist and machine learning engineer, currently working for a major mining corporation. My main goal is to deploy machine learning models and pipelines to boost revenue. Satisfactory for now.’
In the second field, I detail my expectations for ChatGPT’s responses. ‘Your answers must be precise and consistent. When asked about code, provide the most efficient code to solve the problem, well-documented, that suits me too.’
We save this with a click on Save and try a first example. I enter into ChatGPT’s message box: ‘I want to propose a new project about Smart Pricing. Please make a table with the project phases, include the machine learning models to use, and the deadlines for each phase.’
We see at the top of the page that our Custom Instructions are being considered, as indicated by the text ‘Model: GPT for Custom Instructions’ on. Voila! ChatGPT provides the desired solution, aligning with the specifics outlined in our Custom Instructions.
These preferences allow us to converse with ChatGPT without repeatedly reminding it of our background.
In conclusion, let’s explore some practical applications of the new Custom Instructions feature. Fine-tune it to your specific needs. Industry-specific language: ensure the model utilizes sector-specific terminology accurately. Technical level: adjust the model’s technical complexity to match the expertise level of your audience. Communication style: modify the AI’s tone to meet specific communication needs. Task-specific instructions: guide the model according to your preferences for coding tasks. Document formatting: instruct the model to format its output as necessary. Customer service simulation: direct the model to maintain a polite and patient demeanor. Educational use: tailor responses to align with specific learning objectives. Privacy adherence: strengthen the model’s commitment to avoid soliciting private information. Cultural sensitivity: educate the model to respect particular cultural norms.
I hope these tips prove useful. I welcome your feedback and will be producing more videos on this subject soon.
If you found this video on ChatGPT helpful, leave your questions or suggestions in the comments section. Don’t forget to hit the like button and subscribe for more content like this.
Thanks for watching, and I’ll see you in the next one!
