Strange Daily: Issue #4

Relume launches AI site builder that exports to Webflow, Meta partners with Ultimate Fighting Championship (UFC), and a new paper proposes a way to quickly query long-form videos

Happy Tuesday! Here’s the latest in the world of AI and human creativity. There’s lots of key partnerships and growing interest from large enterprises and IP franchises. Interesting times ahead!

We’re now on Spotify and all other major podcasting platforms! Listen to our briefing while on the go.

Latest news

  • AI site generator Relume now exports directly to Figma and Webflow [Link]

  • OpenAI to add major ChatGPT updates, including example prompts, suggested replies, and the ability to upload multiple files to Code Interpreter [Link]

  • Zoom launches a new asynchronous video tool that lets users record, edit and share video clips without having to hop on a meeting [Link]

🤝 Partnerships

  • Meta partners with Ultimate Fighting Championship (UFC): Creating a dedicated UFC-themed experience within Meta’s Horizon Worlds social VR platform [Link]

  • Rapper Lupe Fiasco partners with Google to launch TextFX, an AI tool that assists rappers in their songwriting process [Link]

  • Edtech company Chegg teams up with Scale AI: Integrating generative AI with its content for a personalized learning experience for students [Link]

  • Twilio announces integration with OpenAI: Enabling brands to build customized marketing campaigns with GPT-4 model’s generative capabilities [Link]

🎓 Recent Academic Papers

  • EmotionPrompt: A study by researchers at Microsoft and CAS Institute of Software that explores emotion-enhanced interactions with LLMs [Link to abstract/full-text]

  • Stanford researchers explore emergence of simple language skills in Meta-Reinforcement Learning agents without direct supervision [Link to abstract/full-text]

  • MovieChat: An video understanding system that integrates video foundation models and LLMs [Link to abstract/full-text]

  • TokenFlow: An AI model that uses diffusion features for consistent video editing [Link to abstract/full-text]

  • UC Berkeley researchers introduce Dynalang: An AI agent that learns a multimodal world model to predict future text and image representations and learns to act from imagined model rollout [Link to abstract/full-text]

🔧 Cool Tools and Experiments

  • Music To Image: An AI tool to create images based on audio/music [Link]

  • AirBnB introduces Chronon: A solution for feature engineering in ML models [Link]

  • Pamera: An app that creates poems based on iPhone photos [Link]

💰 Latest Startup Fundings

  • Windup Mind raises $1.6M in pre-seed funding to build a game studio on virtual pets in VR/MR [Link]

  • Tutoring marketplace GoStudent secures $95M to pursue VR and AI-driven tutoring [Link]

  • Language I/O raises $8M in Series A1 funding: Expanding generative AI capabilities for its real-time translation platform [Link]

That's all for today's bulletin! Keep exploring, keep creating, and most importantly, stay strange. If you have any suggestions or feedback, feel free to hit reply and let us know.