Seattle Public Library: “Watch: Why Libraries are on the Front Lines in Supporting Mental Health”
Via Chatlie Guo at Artificial Ignorance
“GPT-4o
OpenAI held their Spring Update event on Monday, with a new model, GPT-4o.
Between the lines:
- At a high level, the model is cheaper, faster, and beats GPT-4 on various benchmarks. For a deep dive, check out Conrad Gray’s excellent post.
- The model is also natively multimodal, meaning it understands your emotions and can offer some of its own – even Sam Altman has compared it to the film Her.
- OpenAI is also releasing a desktop app with GPT-4o, with the ability to view screenshots and presumably work with your files someday.
Elsewhere in AI apps:
- ElevenLabs released Reader, its first consumer app to read web pages and other documents aloud.
- Anthropic launches its Claude Pro and Team subscription plans in Europe, available via web and iOS.
- And Hugging Face commits $10M in free GPUs to help indie developers and academics build more AI apps.
But the biggest news this week came from Google I/O, which announced a slew of AI updates and features across its products and services.
All the announcements I could find:
- A new version of Gemini Pro that can handle up to 2 million tokens.
- Gemini 1.5 Flash, a new model with similar capabilities to Pro but faster and cheaper.
- Gemini Gems, personalized chatbots similar to GPTs.
- Veo, a video generation model that outputs 1080p videos over a minute long.
- Project Astra, an experimental real-time AI agent.
- LearnLM, a new family of AI models for education.
- Music AI Sandbox, an unreleased set of music tools to create loops from prompts.
- AI-powered scam detection for Android phones.
- Ask Photos will search your photo library using Gemini.
- Gemini on Android will integrate with Gmail, Messages, YouTube, and more.
- The SynthID watermark system can work with generative text and video.
- Project IDX, an AI-centric browser IDE, is now in open beta.
- Firebase Genkit, a new open source framework for building AI-powered apps.
- Trillium, the sixth generation of Google’s Cloud TPU chip.
- Google Lens can now search using audio and video input.
- Gemini Nano is coming to Chrome on the desktop.
- An AI-centric redesign of Google’s classic search pages.
- And Gemini Pro is coming to Workspace Labs and educational institutions.”
OpenAI and Reddit Partnership
We’re bringing Reddit’s content to ChatGPT and our products.
https://openai.com/index/openai-and-reddit-partnership/
Via Unwind AI
| “Imagine having a conversation with ChatGPT that brings in the vibrant real-time discussions from Reddit! Content on Reddit is an important source of training data for companies training their AI models. To leverage this position, OpenAI has partnered with Reddit to use its data and provide the company with AI-powered features for Reddit users.
Key Highlights:
- Content Integration: OpenAI will incorporate Reddit’s content into ChatGPT and other products, offering users a more engaging way to explore Reddit communities.
- Real-Time Data Access: OpenAI will use Reddit’s Data API for real-time insights, ensuring up-to-date and relevant content is available for users, particularly on trending topics.
- AI-Powered Reddit Features: Reddit will leverage OpenAI’s platform to create new AI features for users and moderators, improving community management and interaction. This announcement has made Reddit’s falling shares jump around 12% in the market after-hours!”
|
OpenAI and Reddit unite
https://www.aitoolreport.com/articles/openai-and-reddit-unite
ChatGPT To Surface Reddit Content Via Partnership With OpenAI
Reddit partners with OpenAI to integrate content into ChatGPT.
https://www.searchenginejournal.com/chatgpt-to-surface-reddit-content-via-partnership-with-openai/516605/
OpenAI inks deal to train AI on Reddit data
OpenAI inks deal to train AI on Reddit data
Google’s red team takes on misinformation and cyber security.
Recent Comments