Strange Daily: Issue #1
Highlights in today's issue: 7 Big AI companies collaborate to develop watermarking technology, making music from brainwaves, Alibaba introduces their diffusion model, Composer
Welcome to the first edition of Strange Daily — a bi-weekly newsletter sharing bite-sized updates on AI and the future of human creativity. We trawl the web for the most interesting partnerships, research, and projects, and will be sharing them every Tuesday and Thursday. Any feedback? Drop us a note at research@strangevc.com.
Fable, the creators behind Showrunner AI, shows off their ability to automatically generate episodes based on IP. Check out their demo. The idea seems simple, if you are an IP holder of rights to a franchise brand like, says, Cocomelon, you could quickly spin out episodes at scale once the creative direction and production is established. Fable proposes that you could even choose to tweak future episodes based on viewers’ feedback. Read their paper here.
🤝 Partnerships
No more deepfakes? Collaboration between seven major companies, including OpenAI, Microsoft, Google, Meta, Amazon, Anthropic, and Inflection to develop technology for watermarking AI-generated content [Link]
Open AI, Google and Microsoft collaborate to form Frontier Model Forum: An initiative focused on responsible AI practices [Link]
OpenAI partners with American Journalism Project: Invests $5M to establish an AI studio for application of AI in local newsrooms [Link]
🎓 Recent Academic Papers
A new study explores the potential of Large Language Models (LLMs) in open text generation [Link to abstract/full-text]
Brain2Music: A research from Google that reconstructs music from human brain activity [Link to abstract/full-text]
TypeChat: Microsoft releases an AI library to facilitate creation of type-based natural language interfaces to LLMs [Link]
Samsung AI researchers introduce Neural Haircut: A novel AI method to reconstruct strand-based geometry of human hair from video or images [Link to abstract/full-text]
🔧 Cool Startups, Tools, and Experiments
Wix introduced an AI text-to-website generator [Link]
Disco AI: A collaboration between researchers and Microsoft, that can create dance videos from a single photo [Link]
Runway ML releases online text-to-video tool that which generates 4-second clips based on your prompts [Link]
Composer: Alibaba introduces AI-powered image generation with a 5 billion parameters controllable diffusion model [Link]
Genesis: Google tests AI tool that produces news stories and content [Link]
💰 Latest Startup Fundings
HumanFirst secures $5M led by Panache Ventures: Harnessing conversational data’s potential for AI-driven insights [Link]
Sports assets manager ScorePlay raises $5M in seed funding led by SevenSevenSix and Kevin Durant's 35V [Link]
Spiritt secures $13M seed funding: Building an AI platform that builds apps through verbal descriptions [Link]
If you have any suggestions or feedback, feel free to comment and let us know. We're all ears! 🐰 Stay strange.