- Everyday AI
- Posts
- NVIDIA’s $700M acquisition changes everything 😳
NVIDIA’s $700M acquisition changes everything 😳
How GenAI helps recycling, Apple's new AI models, OpenAI's enterprise security features
Everyday AI
April 24, 2024
👉 Subscribe Here | 🗣 Hire Us To Speak | 🤝 Partner with Us | 🤖 Grow with GenAI
Outsmart The Future
Today in Everyday AI
7 minute read
🎙 Daily Podcast Episode: What would happen if we could use GenAI for recycling? We learn how GenAI can help recycle carbon emissions. Give it a listen.
🕵️♂️ Fresh Finds: Meta’s Q1 earnings revolve around AI, OpenAI’s Instruction Hierarchy and Snowflake releases the biggest open-source LLM. Read on for Fresh Finds.
🗞 Byte Sized Daily AI News: Nvidia acquires Run:ai in $700 million deal, Apple releases open source AI models and OpenAI unveils enterprise AI security and control. For that and more, read on for Byte Sized News.
🚀 AI In 5: Here’s a simple ChatGPT hack to get better and more in-depth data analysis. See it here
🧠 Learn & Leveraging AI: We break down how GenAI is helping with recycling and what it means for our society. Keep reading for that!
↩️ Don’t miss out: Did you miss our last newsletter? We talked about concerns with AI deepfakes, Perplexity raises another $250 million, Microsoft's Phi-3 mini AI model and a hidden AI PDF feature in Copilot Check it here!
GenAI: Turning trash into treasures? ♻️
When we think recycling, we might think of plastics.
Probably not Generative AI, right?
Well, that's actually one of the ways that Lanzatech is fighting global warming -- by using Generative AI to help recycle carbon emissions.
How do they do it?
And how has Lanzatech created an internal Large Language Model that's giving them ridiculous in-office efficiencies?
James Daniell, VP of AI and Computational Biology at LanzaTech, joins us to answer these questions and more.
Join the conversation and ask Jordan and James questions on GenAI and recycling here.
Also on the pod today:
• How generative AI helps recycling 🔄
• Use and benefits of AI within LanzaTech 🏢
• GenAI solving environmental problems 🌎
It’ll be worth your 27 minutes:
Listen on our site:
Click to listen
Subscribe and listen on your favorite podcast platform
Listen on:
Upcoming Everyday AI Livestreams
Thursday, April 25th at 7:30 am CST ⬇️
Here’s our favorite AI finds from across the web:
New AI Tool Spotlight – Dart is an AI project management tool, LangWatch helps you improve and iterate LLM pipelines, and Wizad creates social media posters in one click.
OpenAI - OpenAI has released an Instruction Hierarchy, which aims to improve security by preventing quick injections and other methods of misleading LLMs into performing dangerous actions.
LLMs - Snowflake, the cloud computing company, has introduced Arctic LLM, the biggest open-source model to date with 17 billion active parameters and 128 experts, trained on 3.5 trillion tokens.
Big Tech – Meta’s Q1 earnings show that AI is front and center for the company.
Read This – Claude 3 has AI researchers stunned at its level of ‘self-awareness.’
Business of AI - Tesla is ramping up its AI development using NVIDIA chips.
AI in Society – Cisco’s CEO has met with Pop Francis to sign an AI ethics pledge.
Trending in AI – Some Etsy listings for crochet patterns are using AI-generated images, leading to dissatisfied buyers when the final products don't match the pictures.
1. Nvidia Snatches Run:ai in a $700 Million Deal 🤑
Nvidia has made a significant move in the AI industry by acquiring Run:ai, a Tel Aviv-based company, for $700 million. This acquisition will enable Nvidia to integrate Run:ai's cutting-edge products into its DGX Cloud AI platform, providing enterprise customers with access to advanced compute infrastructure for training AI models.
The collaboration is aimed at optimizing generative AI deployments across multiple data center locations, offering customers enhanced flexibility and performance in managing AI workloads.
2. Apple Releases Open Source AI Models 🍎
Apple recently unveiled four OpenELM models, boasting a staggering 3 billion parameters and employing a layer-wise scaling strategy for superior accuracy in AI queries. These models, hosted on Hugging Face, mark a significant leap in large language model technology. With a focus on transparency and reproducibility, Apple's move not only advances open research but also paves the way for groundbreaking AI developments at WWDC in June.
3. OpenAI Raises the Bar in Enterprise AI Security and Control 🛂
OpenAI unveils new enterprise-grade features like Private Link and Multi-Factor Authentication, enhancing security for businesses using their AI models. The upgraded Assistants API now handles 500X more files and introduces Projects for better administrative oversight, revolutionizing how enterprises deploy and manage AI. With cost management features and higher rate limits, OpenAI aims to offer a seamless experience for organizations navigating the AI landscape.
4. UK Investigates Microsoft and Amazon's AI Partnerships 🇬🇧
The UK's Competition Markets Authority (CMA) is delving into the hiring practices of Microsoft and Amazon's collaborations with AI startups to ensure they comply with merger rules. This scrutiny comes amidst concerns that these partnerships could impact competition in the UK's AI market. The CMA's move signals a shift in regulatory focus towards Big Tech's maneuvers in the AI landscape, raising questions about fair competition and innovation.
5. Qualcomm Unveils Snapdragon X Chips for Laptops 💻
Qualcomm is stepping up their game with the Snapdragon X Plus and Elite processors for laptops, challenging the likes of Apple, Intel, and AMD in speed and performance. The Snapdragon X Plus boasts 10 cores, 34GHz frequency, and a whopping 45 TOPS for AI applications, while the Elite chips offer up to 12 cores with impressive speeds and graphics capabilities.
6. Adobe Unveils VideoGigaGAN: Video Upscaling 🎥
In a recent paper, Adobe introduced VideoGigaGAN, a cutting-edge AI model that promises to upscale blurry videos up to eight times their original resolution with unprecedented detail and realism. Unlike traditional upscaling methods, VideoGigaGAN aims to maintain sharpness and clarity while minimizing flickering and distortion issues across frames. The demos provided by Adobe exhibit remarkable enhancements, such as lifelike skin textures and eyebrow hairs, showcasing the potential of this technology in revolutionizing video enhancement.
A Simple ChatGPT Hack For Better Data Analysis!
If you want the most out of ChatGPT, here’s a hack to get better info out of Data Analysis.
Get better breakdowns of your data from spreadsheets and PDFs for more accurate data analysis with this hack.
Check out today's AI in 5.
🦾How You Can Leverage:
Think Generative AI and you probably don’t think about Gucci perfume.
Or tires. Or carbon recycling. Or fighting climate change.
But maybe you should.
If it were up to LanzaTech, that might be the lens through which we view Generative AI.
Because for certain companies, Generative AI can literally mean turning trash into treasures.
James Daniel gave us the blueprint.
James is the VP of AI and Computational Biology, a sustainable technology company that converts waste carbon into valuable products like sustainable aviation fuel or consumer goods.
So, literally turning trash into treasures.
Landfill junk into Gucci perfume.
By recycling carbon extracted from waste products above the ground, LanzaTech aims to reduce (or eliminate) the amount of carbon that needs to be extracted below the ground.
You know, that whole reduce, reuse, recycle thang we learned in 2nd grade?
And their approach is working. Their revenue is up. Productivity is up. And they’re making a difference to fight climate change.
One secret to their success?
Generative AI.
Don’t worry, their GenAI journey isn’t a secret locked up in a lab.
Here’s what you need to know:
1 – Your IP is everything 🔒️
James talked us through the process of how (and why!) LanzaTech created its own version of ChatGPT, dubbed internally as LanzaTech.
In short: LanzaTech has a LOT of knowledge, expertise and in-house intellectual property.
(You know, those smart scientists who somehow made trash smell like Gucci perfume? That secret sauce ain’t easy to reproduce.)
How much?
James said LanzaChat has more than 30,000 pages of internal knowledge base. All of those secrets, findings and expertise?
Now available to all. With the (proper) clicking and clacking of the keyboard.
Try this:
What’s it look like to turn your company’s domain expertise into a large language model?
Here’s a guide on how to fine-tune an open source model on your company’s data.
2 – Know your audience 👥
James droppin hot knowledge like a cat knocking stuff off the dresser at 3 AM.
This is a common problem we hear about at Everyday AI. Whether it’s companies who have their own models or companies just trying to get their employees to use GenAI, adaption can be low.
James said you have to flip the script.
Having a fine-tuned, custom model for your company isn’t a tech tool. It’s a people tool.
(Dude. Rocky if you knock over one more glass…)
You have to set up a safe place to learn and play.
You have to train and teach your people.
And you have to find wins.
(Not just throw tech jargon in an email and hope for the best.)
Did it work?
Yuuuuuuup.
James gave the example of a huge win from someone who wasn’t even a techy early adapter.
A painful documentation task that would normally take 3-4 months took only 6 days with LanzaChat.
Saving about 90% of your time on a daunting knowledge-task with GenAI?
Win.
Try this:
Still struggling to implement GenAI? A little jealous of LanzaTech’s big gains?
Don’t worry. Try this trifecta of episodes (in this order!) that’ll get you going down the right path:
3 – Accelerate and do more 🚀
Ready for more golden nuggets from James?
Well, we’ll mix it with one of our go-to sayings. Instead of ‘use GenAI to focus less on the mundane and more on the meaningful’ you can swap it out with mission.
James reiterated that LanzaTech isn’t just using AI to reduce tires and create more eco-friendly perfumes. They’re focused on a mission of fighting climate change.
He said, "AI accelerates the scientific process... if AI is a great accelerator of science, then that means we can do more science."
For LanzaTech, it’s a simple equation.
More AI = More science = More mission.
Try this:
How can you make that happen?
Humans need to drive. At times, AI can either be a rocket ship or a scapegoat. (Something broke? Oh, the LLM made me do it!)
James said that humans need to take an active role and not just drive AI, but also be the last line of defense and the ultimate owner of outcomes.
Try this:
Wanna do more?
Mission-focused?
This gem of an episode with Nathan Chappell packs more punch than Tyson-Holyfield
Now This …
Let us know your thoughts!
Vote to see live results
What do you want more of at Everyday AI? |
Reply