This week, Google announced the imminent release of Gemini Ultra, which they say has surpassed GPT-4 in 7 out of 8 benchmarks. I’m excited to try it. Also, this means we’ll probably see GPT4.5 from OpenAI very soon.
Personally, the most exciting news this week was that Nat Friedman’s contest to unlock the secrets of ancient Roman scrolls using AI has already had a major breakthrough.
They’ve made the first part of an over 2,000-year-old scroll on papyrus readable using AI. It’s amazing to think about the world we’re heading into where even secrets we thought were lost to time may be rediscovered with new technologies.
Ambience Healthcare raised $70M to create an “operating system” for healthcare organizations to help clinicians complete administrative work (Link)
Colossyan raised $22M for AI video for workplace learning (Link)
Jua raised $16M to build a foundational AI model for the natural world, starting with the weather (Link)
Synthetaic raised $15M to unlock insights from image data (Link)
Quindar raised $6M for AI-driven satellite management (Link)
Arkeus raised $4.45M to expand its AI-enhanced search & surveillance technology (Link)
Octup raised $4M for operational AI-driven e-commerce platform (Link)
Camb AI raised $4M for its AI dubbing platform (Link)
AudioStack raised €2.9M to scale studio-quality audio production for enterprises (Link)
Sable Bio raised £1.5M for AI drug testing (Link)
OpenHouse raised $1.5M to deliver real solutions to real problems in the home building industry (Link)
Humanoid robot startup Figure AI in funding talks with Microsoft, OpenAI (Link)
AI startup Runway is raising a venture capital fund (Link)
Step into the forefront of AI innovation with Lore Labs. As a pioneering AI Video Agency, we're reshaping how stories are told and experiences are created.
Creative Storytelling: Using advanced AI, we bring any concept to life, from advertisements to movie pitches.
Cutting-edge Technology: With tools like Runway and Pika, we're redefining the bounds of digital creativity.
Customized Solutions: Tailoring AI video to fit your unique needs, ensuring memorable and impactful results.
Diverse Applications: Whether it’s for social media, real estate, or entertainment, our AI expertise caters to all.
Embark on a journey of limitless possibilities with Lore Labs. Join our waitlist and transform your vision into reality with the next generation of AI video.
Google announces Gemini Ultra
Google's Gemini Ultra, set for release this Wednesday, has surpassed GPT-4 in 7 out of 8 benchmark tests, marking a significant achievement as the first model to outperform human experts in massive multitask language understanding (MMLU).
OpenAI is exploring the potential risks of large language models (LLMs) aiding in the creation of biological threats. Through a study involving biology experts and students, it was found that GPT-4 might slightly improve the accuracy of creating biological threats, though the effect was not statistically significant. This research is part of a broader effort to develop evaluation methodologies for AI-enabled safety risks, specifically focusing on the misuse potential of AI in biological threat creation.
Roblox launched an AI-powered tool enabling instant text translation for its 70 million daily users across 16 languages. This feature, designed to facilitate seamless communication among players from 180 countries, leverages a custom AI model to translate 2.4 billion daily messages in real-time. This highlights the role of AI in transforming multiplayer gaming into a truly global experience, leading to future developments like live voice translation.
Meta: Labeling AI-Generated Images on Facebook, Instagram and Threads (Link)
China’s generative video race heats up (Link)
EU’s AI Act passes last big hurdle on the way to adoption (Link)
AI-generated fake IDs can now beat security checks
Text-to-3D becomes almost unrecognizable
DomoAI turns home videos into anime
Finding your soulmate with ChatGPT?
Kaiber: Turn text, videos, photos, and music into videos (Link)
2short: Extract the best moments of your videos and turn them into performing short clips (Link)
Syllaby: Create your video marketing strategy in 10 minutes (Link)
InVideo: Create YouTube Shorts using text directly within ChatGPT (Link)
OLMo: Accelerating the Science of Language Models (Link)
Dolma: an Open Corpus of Three Trillion Tokens for Language Model Pretraining Research (Link)
BlockFusion: Expandable 3D Scene Generation using Latent Tri-plane Extrapolation (Link)
Specialized Language Models with Cheap Inference from Limited Domain Data (Link)
If you want to support Lore, please: