The AI Canvas Newsletter #9
Explore the latest AI's news: OpenAI DevDay, The Samsung AI Forum, Runaway Gen-2, Humane AI Pin...
The AI Canvas Newsletter #9
The AI Canvas: Your weekly palette of inspiration, insights, and innovation in the world of AI.
- 🚀 OpenAI Unveils Game-Changing Upgrades at DevDay: GPT-4 Turbo, Assistants API, and More…
- 🎬 Runaway Gen-2: Redefining Creative Possibilities in AI-Generated Videos
- đź“Ś Humane Unveils AI Pin: A Glimpse into the Future of Wearable AI
- 🖼️ LRM: Transforming Single Images into 3D in Just 5 Seconds
- 🔍 Samsung Gauss at the Samsung AI Forum 2023
Written by Oli Wilkins.
OpenAI Unveils Game-Changing Upgrades at DevDay: GPT-4 Turbo, Assistants API, and More…
OpenAI's DevDay announcement introduced huge enhancements to their platform, including the powerful GPT-4 Turbo with 2 times the content window (128K tokens) , a versatile Assistants API enabling agent-like experiences and innovative modalities like vision capabilities for GPT-4 Turbo and the integration of DALL·E 3 for image generation. With reduced prices, higher rate limits, and features like “Function Calling”, the developments aim to empower developers with enhanced AI capabilities for various applications.
Find out more OpenAI’s blog here.
Runaway Gen-2: Redefining Creative Possibilities in AI-Generated Videos
Explore the next frontier in generative AI with Runaway’s Gen-2, a multimodal system that seamlessly transforms text, images, or video clips into captivating films. From turning text prompts into vivid scenes to stylising images for dynamic videos, Runaway Gen-2 sets a new standard for creativity, making the magic of movie-making accessible to all.
Read more here. The paper for Runaway Gen-2 can also be found here.
Humane Unveils AI Pin: A Glimpse into the Future of Wearable AI
Humane introduces the AI Pin, a $699 wearable powered by a Snapdragon processor. With voice control, gestures, and a built-in projector, the device connects to AI models, including ChatGPT, through the AI Mic software. Stripping away traditional interfaces, the Pin aims to simplify interactions with AI, offering features like messaging, email summarisation, nutritional information from food images, and real-time translation. While Humane envisions it as a starting point, the device marks a significant step towards a streamlined, AI-driven user experience.
Read more here.
LRM: Transforming Single Images into 3D in Just 5 Seconds
Introducing the groundbreaking Large Reconstruction Model (LRM), a transformative approach to generate high-fidelity 3D object meshes from a single image in 5 seconds. Utilising a scalable transformer-based architecture with 500 million parameters, LRM breaks away from category-specific training, achieving generalisability by training on a vast dataset of around 1 million objects. Dive into the future of rapid 3D reconstructions, spanning real-world captures to generative model outputs.
Read the summary here and the paper for LRM can be read here.
Samsung Gauss at the Samsung AI Forum 2023
The Samsung AI Forum 2023 delved into the advancements of generative AI, featuring presentations by industry leaders and a showcase of Samsung Gauss, a comprehensive generative AI model developed by Samsung Research. The discussions covered topics such as the operation of large language models, new paradigms in the large language model renaissance, trends in multimodal AI technology, and innovative research from leading universities. Samsung Gauss, comprising Gauss Language, Gauss Code, and Gauss Image, aims to improve work efficiency, support software development and generate creative images.
Read more here.
Posts and Projects
Out-of-Domain Finetuning to Bootstrap Hallucination Detection – Eugene Yan
“It’s easy to finetune models for our specific tasks; we just need a couple hundred or a few thousand samples. However, collecting these samples is costly and time-consuming. What if, we could bootstrap our tasks with out-of-domain data? We’ll explore this idea here.”
WebcamGPT – Roboflow
“just created the WebcamGPT 🤗 @huggingface space”
Runaway/Midjourney/StableAudio Demo - @infiniteyay
@infiniteyay shows off their demo of Runway (featured above), Midjourney and StableAudio (featured all the way back in the AI Canvas Edition #1!).
First Physical Device for Video Editing Generated by AI – Cristóbal Valenzuela
“Introducing 1stAI Machine, the first physical device for video editing generated by AI.”
GPT-4V + TTS = AI Sports Narrator – Gonzalo Espinoza Graham
"Passed every frame of a football video to gpt-4-vision-preview, and with some simple prompting asked to generate a narration."
AI Timelines – LESSWRONG
“How many years will pass before transformative AI is built? Three people who have thought about this question a lot are Ajeya Cotra from Open Philanthropy, Daniel Kokotajlo from OpenAI and Ege Erdil from Epoch. Despite each spending at least hundreds of hours investigating this question, they still disagree substantially about the relevant timescales.”
The Computing Pioneer Helping AI See – Quanta Magazine
“Alexei Efros has spent his career learning how machines see differently from humans. Now he’s helping to bridge the gap.”
Quick Links
Amazon is reportedly racing to build an AI model called Olympus to take on ChatGPT and Bard
Microsoft is renting GPU power from Oracle to push its AI work forward
AI that reads brain scans shows promise for finding Alzheimer’s genes
YouTube is testing a chatbot that will appear under select videos
🚀 Don't miss your weekly dose of cutting-edge AI innovations with The AI Canvas newsletter!
Subscribe now to ensure you never miss out on these transformative insights.
Looking for more specialised consultancy? At ADSP we’re a team of data experts who build AI products with purpose.
We deliver data science projects for companies who want to harness the power that AI can bring to their organisation. Get in touch at hello@adsp.ai.
Stay tuned with The AI Canvas podcast for in-depth episodes exploring Generative AI's transformative role across various industries.