ChatGPT+ Will Soon Be Able To See, Hear, Speak, And Produce Images!

Breaking OpenAI News:

1. OpenAI is delivering on its commitment to a multi-modal ChatGPT+ by integrating both the upgraded DALL·E 3 image generator into the platform as well as the ability to speak with ChatGPT and have it talk back to you!
2. Starting as early as October, users can look forward to an integrated, single-interface experience where ChatGPT serves as an intermediary to DALL-E 3. This setup aligns well with a conversational approach, making it particularly user-friendly for those new to the world of Generative AI.Additionally, a brand new blog post from OpenAI indicates that voice interaction capabilities are on the horizon too.
3. Soon, you'll be able to engage in dynamic, spoken conversations with your AI assistant through ChatGPT.

DALL·E 3 Breakdown

Enhanced Listening to Prompts: It appears that DALL·E 3 is designed to pay closer attention to the specific details in text prompts. Unlike previous versions (or other models and techniques), it doesn't lose essential parts of the prompts.

This will apparently allow for much more accurate and detailed image generation and is going to be especially interesting to Art Directors and creators who really want to bring their own vision to life (every word of it, and rapidly)!

Competitive Again: DALL·E 3 is now far closer to the abilities of Midjourney and Stable Diffusion compared to DALL·E 2. DALL·E 3 even appears to offer as much detail, definition, and life in the images it generates as current leading platforms and will no doubt rapidly improve with user interaction.

Generative Text Support in Images: DALL·E 3 promises better text support in the images it generates, a feature that has been somewhat lacking in previous versions and in other Models like Stable Diffusion & Midjourney. Only Ideogram is currently competing with the generative text in image capabilities of DALL·E 3.

Below is a selection of images that the team at OpenAI shared to showcase the capabilities of DALL·E 3, obviously these will be cherry-picked to show the platform at its best, but we really can't wait to get access and see how consistently it's making this level of quality too!

In Wider OpenAI News:

OpenAI has introduced a fine-tuning UI for their large language models. Users can train these models on their data, but it's still monitored by OpenAI.

ChatGPT's knowledge cutoff seems to have been updated to September 2022, but OpenAI denies updating its models based on user input ??♂?

OpenAI announced a new GPT-3.5 instruct model, which is designed to complete text inputs with instructional capabilities. It appears to be considerably more powerful than the current chat models! ??

Take Midjourney To Crazy New Heights!

Our comprehensive Midjourney training sessions and workshops will massively elevate your ability to create stunning images - for users at all levels!??

谷歌 News:

1. Google's Bard has introduced extensions or plugins that connect to Google's major entities like Maps, YouTube, Flights, and Hotels.
2. It also integrates with Google Workspace. Google is currently testing a powerful new chat-based AI system called Gemini with a small select group of companies. It's seen as Google's answer to OpenAI's models GPT4 and is expected to be a significant competitor in the space (launching soon)!

ElevenLabs

1. Possibly the best text-to-speech platform, has introduced a new "projects" formatting allowing for the creation of long-form audio editing workflows.
2. It can now transform entire books into audiobooks and this allows users to assign multiple speakers to different text fragments giving a greater variety of back-and-forth dialogue and a more natural conversation style.
3. And suppose you need to fix just a specific section of the audio. In that case, projects allow users to seamlessly regenerate parts of audio without disrupting flow or internation, allowing for more control over the output.
4. Because of this longer workflow and the inclusion of chapters, you can now also save and resume your progress and pick up where you left off later!