AI New Hotness January 9th 2025
Jeremy Morgan 
Jan 9, 2025 - 9 min read
Last Update: Jan 9, 2025
Welcome to this week’s edition of the AI New Hotness Newsletter, where we talk about new and exciting stuff to happen in the world of Generative AI, particularly for software developers. It’s a fast moving, wacky world for sure.
AI in the News
The latest AI news you love!
Own Your Own AI Supercomputer for Just $3,000?!
Nvidia’s Digits: Bringing the Power of AI Home
Nvidia is revolutionizing the AI landscape by offering a personal AI supercomputer, Digits, starting at $3,000. This compact powerhouse lets you run complex AI models from the comfort of your home or office, eliminating the need for expensive cloud services or custom-built systems.
Short Summary: Nvidia’s Digits, a compact yet powerful desktop machine, is set to democratize access to AI by allowing individuals to run large language models at home. With impressive specs and connectivity options, Digits offers a cost-effective alternative to cloud computing for AI experimentation.
The Details:
- Compact Size: About the size of a small book, fitting easily on a desk.
- Powerful Hardware: Equipped with the Nvidia GB10 Grace Blackwell superchip, 128GB of unified memory, and up to 4TB of NVMe storage.
- Run Large Models: Capable of running single large language models with up to 200 billion parameters.
- Expandable Power: Connecting two Digits machines enables running even larger models, like Meta’s Llama with 405 billion parameters.
- Cost-Effective: Offers a more affordable solution compared to cloud services or building custom AI systems.
- Availability: Set to go on sale in May.
Why it Matters:
Digits marks a significant shift in AI accessibility. By offering a personal AI supercomputer, Nvidia empowers hobbyists, researchers, and students to experiment with cutting-edge AI technology without relying on expensive cloud infrastructure. This democratization of AI could lead to faster innovation and a wider understanding of this transformative technology. It also allows for greater privacy and control over AI projects, as data remains local. This could be a game-changer for independent researchers and smaller companies.
Turn your search history into a podcast
Tired of endless scrolling? Google is testing a revolutionary AI feature called “Daily Listen” that transforms your search history and news preferences into a personalized daily podcast. Imagine getting a concise, 5-minute audio briefing on the topics you care about, delivered right to your Google app. This could change how we consume information forever!
The Details:
- Personalized Podcasts: Daily Listen creates a short podcast based on your Google Search history and Discover feed activity.
- Daily Briefing: A 5-minute overview of your favorite topics appears automatically in your Discover feed each day.
- Easy Access: The podcast is labeled “Made for you” and includes playback controls, transcription, and feedback options (thumbs up/down).
- Search Labs Experiment: Currently available to US users in the Search Labs program (limited spots available!).
- Similar to NotebookLM: Works like NotebookLM’s Audio Overviews, but focuses on recurring topics and user interests rather than specific prompts.
Why it Matters:
This feature has the potential to revolutionize how we stay informed. It offers a convenient, personalized, and engaging way to consume news and information tailored to our individual interests. Imagine catching up on your favorite topics during your commute, workout, or while doing chores – all without lifting a finger. If successful, this could be the future of personalized news consumption, making staying informed easier and more enjoyable than ever before. Click to learn more and see if you can get early access!
Google’s Plan to Build an AI That Thinks Like a Human
Google DeepMind is assembling a dedicated research team to develop “world models,” a revolutionary approach to AI that aims to simulate the physical world and enable AI to predict and understand on a large scale. This initiative could be the key to achieving Artificial General Intelligence (AGI), an AI that can think and reason like a human.
Microsoft’s Tiny Genius Crushes Much Bigger Brains at Math (Seriously!)
Small AI Model Packs a Powerful Punch in Complex Reasoning
Feeling lost on your next math problem? Buckle up, because help has arrived, and it’s surprisingly small! Microsoft just released Phi-4, a tiny AI whiz kid that outperforms even much larger models when it comes to complex reasoning, especially in math.
Here’s the scoop:
- Phi-4 is a “small language model” with only 14 billion parameters (compared to giants with hundreds of billions!).
- Despite its size, Phi-4 aces math problems, even beating out some of the biggest names in AI.
- Microsoft built Phi-4 responsibly, with safety features to prevent misuse. ️
Why this matters:
Phi-4 is a game-changer for AI. It proves that smaller models can be just as powerful, if not more, for specific tasks. This could lead to faster, more efficient AI development and open doors for new applications, especially in education and scientific research.
Ready to see Phi-4 in action?
Head over to Microsoft’s Azure AI Foundry or Hugging Face to explore its capabilities!
9 NVIDIA Announcements From CES 2025
NVIDIA’s CES 2025 keynote was a whirlwind of groundbreaking AI announcements, from personal supercomputers to advanced robotics and autonomous vehicles. These advancements, coupled with the power of blockchain, are poised to transform industries and redefine how we interact with technology.
The Details:
- Project DIGITS: A handheld AI supercomputer packing the power to handle massive AI models, bringing unprecedented processing power to individual developers.
- Cosmos World Foundation Models: Revolutionizing autonomous vehicle training by turning real-world driving data into billions of virtual simulations.
- AI Blueprints for Video, Agents, and RTX PCs: Democratizing access to powerful AI tools for video analysis, intelligent agents, and everyday computing, boosted by blockchain for security and authenticity.
- Isaac GR00T for Humanoid Robotics: A new framework for training humanoid robots, paving the way for widespread automation in manufacturing and logistics.
- Omniverse “Mega” Blueprint: Enabling the creation of detailed digital twins of entire facilities, optimized for robotic system testing and enhanced by blockchain for data integrity.
- Autonomous Driving Innovations: Combining DRIVE Thor and DRIVE Orin platforms with key partnerships to accelerate the development of safe and efficient self-driving vehicles.
Why it Matters:
NVIDIA is not just building chips; they are building the future. Their advancements in AI, robotics, and digital twins are poised to revolutionize industries like manufacturing, logistics, transportation, and even personal computing. The integration of blockchain technology adds a layer of security, transparency, and trust, further amplifying the impact of these innovations. This is a crucial moment for anyone interested in technology, as these advancements are set to reshape the world around us. Click to learn more about how NVIDIA is leading the charge into the next industrial revolution.
AI Tools
Tools you can use as a developer, and more.
Build Software Without Writing Code - Lovable Makes Development Easy
Sure, here is a catchy headline, a short summary of what the tool does, and why someone would want to use this tool based on the webpage https://lovable.dev/:
Headline: Build Software Without Writing Code - Lovable Makes Development Easy
Summary: Lovable is a tool that allows users to build software without writing code. It is a faster and easier to use than traditional coding methods.
Why someone would want to use this tool:
- No coding required: Lovable allows users to build software without writing code, making it accessible to people with no coding experience.
- Faster development: Lovable is said to be faster than traditional coding methods, which can save developers time and money.
- Versatility: Lovable can be used to build a variety of different types of software, including websites, mobile apps, and desktop applications. It can also be used to build software for specific industries.
Create Engaging Video Training Content in Minutes with Lupo
Summary: Lupo is a platform that helps businesses create high-quality video training content quickly and easily. With Lupo, you can use Markdown and Marp to create courses, and even leverage their agency services if needed.
Why someone would use this tool:
Create high-quality video training content quickly and easily Use Markdown and Marp to create courses Leverage their agency services if needed
Learn about AI
Courses and resources for learning more about AI.
Learn by Doing - Prompt Engineering 101
“Learn by Doing - Prompt Engineering 101” is an interactive course on prompt engineering with AI. It features an application that connects to the Mixtral-8x7B model, using a ChatGPT-like interface. The goal? To help users craft effective prompts for AI. In practical labs, learners will discover ways to get accurate and detailed responses from AI. This course welcomes everyone, from AI beginners to pros wanting to upgrade their tech interactions.
Study with us and get NVIDIA Certified!
Happy 2025 Everyone! Michael Forrester and I are looking for a few good people who might want to join us as we study for the NVIDIA-Certified Associate Generative AI LLMs (NCA-GENL).
https://www.nvidia.com/en-us/learn/certification/generative-ai-llm-associate/
This is a basic AI certification with NVIDIA and is the first of six possible certifications with them.
We are thinking of a once-a-week hourly check-in (maybe one time for Eastern US/EST and another for India/IST) for 6-8 weeks as a study group on our KodeKloud Community Discord server!
We have created and taught several AI-related courses, from AWS AI practitioner to Coding assistants, and were looking for some allies in studying for this basic NVIDIA certification.
From there we will probably chase the other NVIDIA certifications. Reply here if you are interested, or DM me or Michael Forrester on LinkedIn, and we will add you to the list.
Looking to start mid-January on the 15th, so let us know if you want to join us for this.
We will provide some structure, you will study, we will offer support and exam questions, and you help us pass as we help you pass. :) 1 hour a week and a great addition to your budding AI-DevOps resume!!!
Thanks for reading!! I started this newsletter to share all the cool stuff I’m finding in AI over time. Keep updated on the latest stuff with this newsletter. I’m glad you’re here.
Share this on your LinkedIn page if you think your friends might be interested in this stuff.
BTW Check out my YouTube Channel for more cool stuff with Generative AI. Tips, tutorials, reviews of AI tools and more.
Follow @JeremyCMorgan