The AI Canvas Newsletter #13
Revealing the newest AI advancements: Google's Gemini Advanced Ultra 1.0, Apple's MGIE Magic and MobileDiffusion art.
This Week in AI: Innovations, Reports, and Features
🌟 Gemini Advanced: Google's latest stride in AI, introducing Ultra 1.0.
🖌️ Apple's MGIE Magic: Transform images with the power of words using Apple's open-source AI for intuitive image editing.
🌙 Dream Weaver Tech: Step into the realm of conscious dreaming with 'The Halo' and 'Morpheus-1', the frontier of lucid dream exploration.
🤖 EVE Evolves Autonomy: 1X's EVE android intelligence, mastering tasks through sight.
📲 Instant Artistry with MobileDiffusion: Google's MobileDiffusion turns your text into stunning images on your phone, in a flash.
😇 GOODY-2's Gentle Calculations: Say hello to GOODY-2, the AI so polite it thinks twice before solving 2+2, setting a new standard in AI ethics.
Gemini Advanced: The Next Step in Google's AI Evolution
Sundar Pichai, CEO of Google, details the latest developments in AI with the introduction of Gemini Advanced, an upgrade to the company's AI capabilities. The new model, Ultra 1.0, achieves a milestone in language understanding and is poised to integrate with consumer and business products, including Workspace and Google Cloud. Gemini Advanced aims to assist with tasks ranging from personal tutoring to business planning, accessible through the new Google One AI Premium plan.
Find out more on the announcement page and the blog page.
Apple Unveils MGIE: Open-Source AI for Language-Directed Image Editing
Sora is an AI model designed to create videos from textual prompts, producing scenes that range from realistic cityscapes to imaginative animations. The model, which is being tested by visual artists and red teamers, can generate videos up to a minute long, with a focus on adhering to the details of the user's instructions. Despite its capabilities, Sora is still being refined to overcome challenges in physical simulation and temporal consistency.
Checkout OpenAI’s announcement here.
Navigating the Mind's Labyrinth: The Halo & Morpheus-1's Lucid Dreaming Odyssey
Discover 'The Halo' and 'Morpheus-1', pioneering neurotechnologies designed to unlock the mysteries of lucid dreaming. 'The Halo' is a cutting-edge wearable that simulates the brain's natural lucid dream states, while 'Morpheus-1' uses ultrasonic holograms to induce these vivid dreams.
1X's EVE: Pioneering Vision-Guided Autonomy in Androids
1X unveils the latest strides in android autonomy with EVE, an android capable of learning diverse tasks through vision-based neural networks. The article delves into the end-to-end data-driven training process that enables EVE to perform without pre-programming, emphasising the transformative role of 'Software 2.0 Engineers' in robotics. Additionally, it announces career opportunities for those eager to contribute to the future of physically embodied intelligence at 1X.
Read more here.
Google's MobileDiffusion: Quick and Easy Image Creation from Text on Your Phone
Google's latest innovation, MobileDiffusion, brings the magic of creating images from text descriptions right to your smartphone, almost instantly. This technology simplifies the process, using a method that allows it to quickly turn text into images in a single step, making it practical and fast for everyday use on mobile devices. With its relatively small size, MobileDiffusion promises to be a game-changer for creative mobile apps while keeping in line with Google's ethical AI guidelines.
Read more here.
Meet GOODY-2: The AI That's Too Polite to Compute 2+2
GOODY-2 is the latest in AI that's so ethically cautious, it won't even risk the maths on a preschool worksheet. With a commitment to safety that borders on the comical, this AI takes no chances, dodging even the most innocent questions to steer clear of potential controversy.
Have a play around here.
Technical Reads
Self-Reflective RAG - LangGraph – LaingChain
“In practice, many have found that implementing RAG requires logical reasoning around these steps: for example, we can ask when to retrieve (based upon the question and composition of the index), when to re-write the question for better retrieval, or when to discard irrelevant retrieved documents and re-try retrieval? The term self-reflective RAG (paper) has been introduced, which captures the idea of using an LLM to self-correct poor quality retrieval and / or generations.”
The pain points of building a copilot - Austin Henley
“Given the sudden surge in generative AI being integrated into products, we wanted to know what is the process that software engineers follow to build these products, what are the pain points, and what are the opportunities for tools.”
AI Design Patterns - Tomasz Tunguz
“As we’ve been researching the AI landscape & how to build applications, a few design patterns are emerging for AI products. These design patterns are simple mental models. They help us understand how builders are engineering AI applications today & which components may be important in the future.”
RAG vs Fine-tuning: Pipelines, Tradeoffs, and a Case Study on Agriculture
“There are two common ways in which developers are incorporating proprietary and domain-specific data when building applications of Large Language Models (LLMs): Retrieval-Augmented Generation (RAG) and Fine-Tuning. RAG augments the prompt with the external data, while fine-Tuning incorporates the additional knowledge into the model itself. However, the pros and cons of both approaches are not well understood.”
Projects and Code
Ask HN: What have you built with LLMs?
“Curious what people have been building with LLMs.”
LLM Prompting
Build prompt templates for Chat GPT, Bard, Claude2 and others.
Datatrove
“DataTrove is a library to process, filter and deduplicate text data at a very large scale. It provides a set of prebuilt commonly used processing blocks with a framework to easily add custom functionality.”
Learning
Supervised Machine Learning for Science
“Machine learning has revolutionized science, from folding proteins and predicting tornadoes to studying human nature. While science has always had an intimate relationship with prediction, machine learning amplified this focus. But can this hyper-focus on prediction models be justified? Can a machine learning model be part of a scientific model? Or are we on the wrong track?”
Build a Large Language Model (From Scratch)
“Implementing a ChatGPT-like LLM from scratch, step by step.”
Business and Trends
Journey Through the AI Canvas Podcast
Latest Episode of the AI Canvas
The AI Canvas - Generative AI in the Classroom: The Future of Learning with Francisco Recalde
In this enlightening episode of the AI Canvas podcast, host David Foster sits down with Francisco Recalde, Head of the Department of Languages at Dixon's Unity Academy, to explore the transformative effects of AI on education. They discuss AI’s potential in teaching and learning, the fear of AI replacing teachers, and the role of AI as a guide for students.
David Foster
Founding Partner, ADSP
In Case You Missed It: Explore Our Recent Newsletters
The AI Canvas Newsletter #12
Explore the latest AI advancements: Google's Lumiere, Olympiad-level geometric reasoning with AlphaGeometry, NVIDIA's HDR video enhancement!
The AI Canvas Newsletter #14
Discover the latest in AI innovation from Google's Gemini 1.5 and ethical AI ensemble Gemma, OpenAI's model Sora, Stable Diffusion 3's creative surge...
The AI Canvas Newsletter #15
Explore the latest in AI news: Anthropic's Claude 3, Pi-2.5, and NVIDIA's StarCoder2.
Looking for more specialised consultancy?
At ADSP we’re a team of data experts who build AI products with purpose.