Summary

  • Gemini, Google’s AI chatbot, can now pull real-time info from various Google apps but lacks memory for preferences.
  • The upcoming “Memory” feature in Gemini will allow the chatbot to remember your details over time, enhancing its helpfulness.
  • Google is also working on a “Gems” feature for customized Gemini versions.



Google’s AI-powered chatbot Gemini can help with research, holiday planning, and more. With Extensions, the chatbot can also pull relevant and real-time information from other Google apps like Google Maps, YouTube, Workspace, and Google Hotels. However, Gemini cannot remember your preferences, so you must repeat yourself while revisiting a topic you want more information about. Google could soon fix this problem by introducing “Memory” in Gemini.


A Samsung smartphone on a kitchen counter, showing the Google Gemini interface

Related

Google Gemini tips and tricks: Put Google’s most capable AI model to good use

Not sure what’s up with Gemini? Here are the goods

As the name indicates, Memories will allow the chatbot to remember your preferences, interests, and other information shared with it. This will enable you to talk to Gemini without repeating yourself or providing the necessary information again, as the chatbot will pull all relevant details from past conversations.

Based on the screenshot shared by @evowizz, the Memories feature will also allow the chatbot to learn more about you over time, making its responses more helpful. Google was first spotted working on the ‘Memory’ feature in September 2023, when Gemini was still known as Bard.


OpenAI’s ChatGPT already offers a similar ‘Memory’ feature, ensuring the chatbot can remember key details about you and your preferences.


Google could soon let you create a customized version of Gemini

Gemini web app homepage

Besides Memories, Google is developing a new “Gems” feature, allowing you to “Customize Gemini for your needs.” Apparently, you can use “Gems” to create a more customized Gemini that meets your requirements.

Some early Gemini testers have also been able to enter up to 1M tokens directly in the chatbot, a restriction previously imposed by Google despite the model’s support for such large inputs.

Lastly, leaker Dylan Rouseel found signs of Google letting you upload up to 10 documents simultaneously from Google Drive to Gemini, including PDFs, Word files, and Google Docs. This change should significantly speed up your workflow if you frequently upload documents to Gemini for analysis.


Multiple Gemini-related leaks in recent weeks indicate the company is working on several features for its AI chatbot, like a dedicated stop button in Gemini’s mobile app, the ability to Summarize long emails in Gmail, ‘Live Prompts’ for Assistant automation, and more. With I/O 2024 a day away, we should see Google announce these and several other Gemini enhancements at the event.