5 Essential ChatGPT Tips from the Creators — My Take After Testing Them

Understanding the Power of GPT-5.1

Despite increasing competition from major tech companies like Google and Anthropic, OpenAI’s GPT-5.1 continues to stand out as one of the most advanced AI models available today. Recently, the OpenAI team shared valuable insights on how to get the most out of this powerful tool during an episode of their podcast. Two key figures in the discussion were Christina Kim, a research lead in post-training at ChatGPT, and Laurentia Romaniuk, a product manager for model behavior.

The episode focused entirely on GPT-5.1, exploring its unique features and offering practical advice on maximizing its capabilities. Below are some of the best tips that emerged from the conversation.

Memory Tricks

One of the standout features introduced in recent updates is the memory function within ChatGPT. When enabled, this feature allows the AI to retain information from past conversations, such as personal preferences, birthdays, or even specific interests like a love for certain types of films or technology.

Romaniuk explained that this functionality helps the AI provide more personalized responses by drawing on previous interactions. “The model will write down things it knows about you from its conversations with you to refer back to later. This is really nice because you’re not just repeating yourself every time,” she said.

This feature not only makes the experience more seamless but also allows the AI to tailor its responses to better suit individual users. While the memory can be turned off, the team emphasized how intuitive and effective it can be when enabled.

Customize the Chatbot

In the podcast, Romaniuk provided guidance on how to make ChatGPT perform better based on specific needs. She shared a story about her brother, who found the AI’s responses too basic. Instead of complaining, he was advised to specify the level of expertise required.

By instructing the model to act as a “frontier researcher,” the AI produced insights that matched the complexity of his lab's work. This approach has been widely used in the AI community, where users tell the model what level of knowledge to adopt. For example, starting a prompt with “You’re an expert biologist” or “You’re a sub-editor for a tech magazine” can significantly improve the quality of the response.

Challenge ChatGPT

AI models are constantly evolving, so it can be difficult to know their full potential. One piece of advice from the podcast is to pressure test the model with challenging questions. Kim encouraged users to ask hard questions they are familiar with, noting that “we’re shipping updates all the time.”

She suggested that if a user previously had a negative experience with ChatGPT, they should try again later, as the model may have improved. Pushing the model in areas the user is knowledgeable about can also reveal its strengths and weaknesses.

Ask for Help

A simple yet effective tip from the podcast is to ask ChatGPT for help with prompts. This is something the team highlighted as a former weakness, where the AI would sometimes struggle to generate effective prompts on its own. However, modern chatbots have become much better at crafting useful prompts.

Users can ask ChatGPT to come up with a prompt tailored to their specific task. For instance, asking “come up with a prompt to [insert the task you’re trying to achieve]” can guide the model more efficiently toward the desired outcome.

Try New Personalities

One of the major updates in GPT-5.1 was the introduction of new personalities, allowing users to interact with ChatGPT in different modes—more fun, friendly, serious, or professional. The team discussed how they’ve enjoyed experimenting with these settings, making the experience more engaging.

While these personalities don’t necessarily enhance the model’s intelligence, they add a layer of flexibility and enjoyment to interactions. Testing different modes can reveal how adaptable the AI can be in various situations.