We're sharing an update on the advanced Voice Mode we demoed during our Spring Update, which we remain very excited about:
We had planned to start rolling this out in alpha to a small group of ChatGPT Plus users in late June, but need one more month to reach our bar to launch. For example, we’re improving the model’s ability to detect and refuse certain content. We’re also working on improving the user experience and preparing our infrastructure to scale to millions while maintaining real-time responses.
As part of our iterative deployment strategy, we'll start the alpha with a small group of users to gather feedback and expand based on what we learn. We are planning for all Plus users to have access in the fall. Exact timelines depend on meeting our high safety and reliability bar. We are also working on rolling out the new video and screen sharing capabilities we demoed separately, and will keep you posted on that timeline.
ChatGPT’s advanced Voice Mode can understand and respond with emotions and non-verbal cues, moving us closer to real-time, natural conversations with AI. Our mission is to bring these new experiences to you thoughtfully.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
We’re starting to roll out advanced Voice Mode to a small group of ChatGPT Plus users. Advanced Voice Mode offers more natural, real-time conversations, allows you to interrupt anytime, and senses and responds to your emotions.
Users in this alpha will receive an email with instructions and a message in their mobile app. We'll continue to add more people on a rolling basis and plan for everyone on Plus to have access in the fall. As previously mentioned, video and screen sharing capabilities will launch at a later date.
Since we first demoed advanced Voice Mode, we’ve been working to reinforce the safety and quality of voice conversations as we prepare to bring this frontier technology to millions of people.
Our new GPT-4 Turbo is now available to paid ChatGPT users. We’ve improved capabilities in writing, math, logical reasoning, and coding.
Source: github.com/openai/simple-…
For example, when writing with ChatGPT, responses will be more direct, less verbose, and use more conversational language.
We continue to invest in making our models better and look forward to seeing what you do. If you haven’t tried it yet, GPT-4 Turbo is available in ChatGPT Plus, Team, Enterprise, and the API.
Sora can create videos of up to 60 seconds featuring highly detailed scenes, complex camera motion, and multiple characters with vibrant emotions.
Prompt: “Beautiful, snowy Tokyo city is bustling. The camera moves through the bustling city street, following several people enjoying the beautiful snowy weather and shopping at nearby stalls. Gorgeous sakura petals are flying through the wind along with snowflakes.”openai.com/sora
We’ll be taking several important safety steps ahead of making Sora available in OpenAI’s products.
We are working with red teamers — domain experts in areas like misinformation, hateful content, and bias — who are adversarially testing the model.
Prompt: “Several giant wooly mammoths approach treading through a snowy meadow, their long wooly fur lightly blows in the wind as they walk, snow covered trees and dramatic snow capped mountains in the distance, mid afternoon light with wispy clouds and a sun high in the distance creates a warm glow, the low camera view is stunning capturing the large furry mammal with beautiful photography, depth of field.”
We are systemizing our safety thinking with our Preparedness Framework, a living document (currently in beta) which details the technical and operational investments we are adopting to guide the safety of our frontier model development. openai.com/safety/prepare…
Our Preparedness Team will drive technical work, pushing the limits of our cutting edge models to run evaluations and closely monitor risks, including during training runs. Results will be synthesized in scorecards that track model risk.
Our new safety baselines and governance process will turn these technical findings into safety decisions for model development and deployment. This involves establishing a cross-functional Safety Advisory Group to make safety recommendations.
In the future, humans will need to supervise AI systems much smarter than them.
We study an analogy: small models supervising large models.
Read the Superalignment team's first paper showing progress on a new approach, weak-to-strong generalization: openai.com/research/weak-…
Large pretrained models have excellent raw capabilities—but can we elicit these fully with only weak supervision?
GPT-4 supervised by ~GPT-2 recovers performance close to GPT-3.5 supervised by humans—generalizing to solve even hard problems where the weak supervisor failed!
Naive weak supervision isn't enough—current techniques, like RLHF, won't be sufficient for future superhuman models.
But we also show that it's feasible to drastically improve weak-to-strong generalization—making iterative empirical progress on a core challenge of superalignment