The AI Canvas Newsletter #17
Dive into the latest AI innovations: Stable Audio 2.0, Grok-1.5V's long-context comprehension and OpenAI's Voice Engine.
This Week in AI: Innovations, Reports, and Features
🎶 Stable Audio 2.0: Generate full-length tracks and modify audio with natural language.
🤖 Grok-1.5: Enhanced long-context understanding and reasoning with a 128,000 token window.
🗣️ OpenAI's Voice Engine: Creating custom synthetic speech for diverse applications, with a focus on ethical use.
đź’» DBRX: Databricks' efficient and high-performing open large language model, challenging GPT4 Turbo.
AI Music Generation: Introducing Stable Audio 2.0
All the way back in September last year, the first edition of The AI Canvas featured Stable Audio. Since then, Stable Audio has advanced to version 2.0, now enabling users to generate full-length, high-quality tracks and modify audio samples using natural language. This latest iteration respects artist rights with a licensed dataset and offers its enhanced creative tools for free on the Stable Audio website.
Grok-1.5: Extended Context and Reasoning Abilities
Grok-1.5, the latest AI model from xAI, boasts significant advancements in long context understanding, with a context window of 128,000 tokens, and enhanced reasoning skills, evidenced by its high scores on various benchmarks. The model, supported by a robust infrastructure, is set to be released on the đť•Ź platform, promising improved capabilities for complex problem-solving tasks.
Read more from xAI here.
Voice Engine Preview: The Path to Ethical Synthetic Speech Deployment
OpenAI shares insights from the preliminary testing of Voice Engine, a model designed to create custom synthetic voices from minimal audio input. The article discusses the model's applications, from aiding non-readers to assisting non-verbal individuals, and addresses the importance of responsible deployment and safety measures, such as watermarking and consent policies, to mitigate misuse risks.
Read more on OpenAI’s blog.
DBRX: The New Benchmark in Open Large Language Models
Databricks introduces DBRX, an open-source large language model (LLM) that challenges the performance of GPT-4 Turbo while setting new efficiency standards. DBRX's mixture-of-experts architecture enables rapid inference and a compact model size, offering a robust alternative for both the open community and enterprises.
Read more on Databricks’ blog.
Technical Reads
Mamba Explained – Kola Ayonrinde
“Practically all the big breakthroughs in AI over the last few years are due to Transformers. Mamba, however, is one of an alternative class of models called State Space Models (SSMs).”
Your AI Product Needs Evals – Hamel Husain
“How to construct domain-specific LLM evaluation systems.”
LLM Task-Specific Evals that Do & Don't Work – Eugene Yan
“If you’ve ran off-the-shelf evals for your tasks, you may have found that most don’t work. They barely correlate with application-specific performance and aren’t discriminative enough to use in production. As a result, we could spend weeks and still not have evals that reliably measure how we’re doing on our tasks.”
Notes on how to use LLMs in your product. – Will Larson
“I’ve been working fairly directly on meaningful applicability of LLMs to existing products for the last year, and wanted to type up some semi-disorganized notes. These notes are in no particular order, with an intended audience of industry folks building products.”
Projects, Code and Discussions
AI Building Stuff in Minecraft
“This video shows off AI agents that build stuff in minecraft. I used chatgpt 4 turbo, claude 3 opus, and gemini 1.0 to power three different agents and tested their ability to follow commands, build complex structures, and not annoy me (impossible).”
The Next Big Step in Mojo🔥 Open Source
“At Modular, open source is ingrained in our DNA. We firmly believe for Mojo to reach its full potential, it must be open source. We have been progressively open-sourcing more of Mojo and parts of the MAX platform, and today we’re thrilled to announce the release of the core modules from the Mojo standard library under the Apache 2 license!”
Business and Trends
- How Stability AI’s Founder Tanked His Billion-Dollar Startup
- Amazon spends $2.75 billion on AI startup Anthropic in its largest venture investment yet
- Jony Ive and OpenAI's Sam Altman Seeking Funding for Personal AI Device
- Microsoft AI opens new AI hub in London to attract talent and drive advanced LLM research
- OpenAI and Microsoft reportedly planning $100 billion datacenter project for an AI supercomputer
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 #15
Explore the latest in AI news: Anthropic's Claude 3, Pi-2.5, and NVIDIA's StarCoder2.
The AI Canvas Newsletter #16
Explore AI breakthroughs: NVIDIA's efficiency leap,Grok-1's open-source release, Stable Video 3D's innovation, and DeepMind's SIMA mastery...
The AI Canvas Newsletter #18
Delve into the latest AI advancements: Boston Dynamics' Electric Atlas, Llama 3, Udio, Grok-1.5V, The Silicon Shift and more.
Looking for more specialised consultancy?
At ADSP we’re a team of data experts who build AI products with purpose.