What I learned from ChatGPT-4o’s launch event

ChatGPT-4o is here. The “o” stands for omni, hinting at the combination of text, audio, video and image outputs that the latest version of ChatGPT offers.

OpenAI CEO Sam Altman posted on X that GPT-4o is “natively multimodal” with its native combination of voice, text and vision. At the launch livestream of GPT-4o earlier this week, OpenAI CTO Mira Murati explained how GRP-4o addresses the latency problems with GPT’s previous voice mode implementation. GPT-4o will be freely available and users will be able to do things like convert text into image outputs.

Image Credit: screenshot

Image Credit: screenshot

During the livestream the folks at OpenAI demoed the improved voice capability of GPT-4o, showing how one no longer has to wait for the voice assistant to finish; you can but in and it won’t disrupt the conversation. The model offers real-time responsive voice mode, removing the 2.3 second wait time before getting a response. Whilst you can still tell that the assistant’s voice is computer generated, I was impressed by how natural the conversation sounded. With GPT-4o Open AI has improved the model’s ability to both perceive and generate emotions.

And finally, I really liked GPT-4o’s ability to interact visually, with a phone camera pointed at a math problem, with GPT talking the user through solving the math problem. Another good example of how ChatGPT-4o combines different inputs and outputs.

Main learning point: I know that some people were hoping for the release of ChatGPT 5, but I was already very impressed with the public demo of ChatGPT-4o and I look forward to having a play myself, and experience ChatGPT-4o’s ‘omni’ approach first hand.

Related Links:

1. https://openai.com/index/spring-update/
2. https://openai.com/index/hello-gpt-4o/
3. https://blog.samaltman.com/gpt-4o
4. https://openai.com/index/gpt-4o-and-more-tools-to-chatgpt-free/
5. https://twitter.com/gdb/status/1790071008499544518
6. https://www.lifewire.com/openai-reveals-gpt-4o-8647637
7. https://www.theverge.com/2024/5/13/24155493/openai-gpt-4o-launching-free-for-all-chatgpt-users
8. https://arstechnica.com/information-technology/2024/05/chatgpt-4o-lets-you-have-real-time-audio-video-conversations-with-emotional-chatbot/