OpenAI is Rolling Out ‘GPT Mentions’ in ChatGPT to Invoke Custom GPTs
OpenAI is quietly testing a ChatGPT feature that allows users to mention and invoke custom GPTs during a conversation, as first spotted on X (formerly Twitter) by Dan Shipper. These ‘GPT mentions’ enable ChatGPT users to tag customized bots with an ‘@’ symbol, similar to mentioning someone on platforms like Slack or Twitter, allowing them to share access without needing to navigate to OpenAI’s GPT marketplace.
Mention GPTs
HOLY SHIT they built it pic.twitter.com/oitY6FZ6gf
— Dan Shipper 📧 (@danshipper) January 26, 2024
GPT mentions appear to be aimed at integrating the custom GPTs into conversations with as simple a process as possible. Typing ‘@’ during a ChatGPT conversation brings up a list of GPTs that users can mention and use directly in their chat. It could reduce friction in using multiple GPT models and boost the efficiency and productivity of the users by allowing them to integrate multiple GPTs into a single chat.
OpenAI has not made an official announcement about this feature yet. The selective and gradual rollout is likely set up to iron out any bugs or issues with the feature. The GPT mentions help in a lot of ways. For instance, planning a trip through ChatGPT could then be enhanced by writing ‘@TripBot to get recommendations for activities in Paris from a separate travel agent bot. Multiple bots could also be integrated together in one ongoing chat to combine capabilities.
As it continues iterating ChatGPT, OpenAI is likely aiming to balance enhancing functionality with safety. Mentions may enable more powerful multi-bot capabilities, but the company will need to ensure rigorous oversight measures are in place as well. While rapidly innovating, the research nonprofit appears cautious not to overpromise capabilities that remain in early stages.
That doesn’t mean OpenAI hasn’t made some big and public upgrades. The company recently debuted new models designed to, among other things, make ChatGPT and products derived from its API less ‘lazy.’ Major upgrades over the last year have given ChatGPT the ability to search online for answers and multimodal options that allow for verbal conversations and enable the AI to process images in addition to text as well as document and PDF analysis. Other improvements are quieter but features like chat archiving still work to improve how well the generative AI chatbot functions.
Follow @voicebotaiFollow @erichschwartz
OpenAI Showcases New Generative AI Models and Lower API Prices, Cures GPT-4 ‘Laziness’
