- Everyday AI
- Posts
- AI’s Energy Crisis - Can Quantum Save the Day?
AI’s Energy Crisis - Can Quantum Save the Day?
Mistral's edge-ready AI models, Perplexity's financial analysis tool, Boston Dynamics humanoids get AI and more!
Everyday AI
October 16, 2024
👉 Subscribe Here | 🗣 Hire Us To Speak | 🤝 Partner with Us | 🤖 Grow with GenAI
Outsmart The Future
Today in Everyday AI
8 minute read
🎙 Daily Podcast Episode: With AI’s rapid boom, its energy consumption needs are rising every day. So what can we do as a solution? An expert breaks down how quantum computing can help. Give it a listen.
🕵️♂️ Fresh Finds: Alibaba’s new translation tool, NVIDIA’s finetuned Llama 3.1 model outbenching GPT-4o on some benchmarks, YouTube’s AI updates, Amazon Kindle gets AI and more. Read on for Fresh Finds.
🗞 Byte Sized Daily AI News: Mistral unveils edge-ready AI models, Perplexity announces financial analysis tool and Boston Dynamics adds AI intelligence to humanoid robot. For that and more, read on for Byte Sized News.
🚀 AI In 5: We found a secret hack that lets you talk to webpages using Microsoft Edge and Copilot. See it here
🧠 Learn & Leveraging AI: Wondering what we can do to fix AI’s energy problem? We break down a solution and how it may affect us all. Keep reading for that!
↩️ Don’t miss out: Did you miss our last newsletter? We talked about NVIDIA reaching an all-time high, Copilot getting OpenAI’s o1 model, NYT issuing a C&D to Perplexity and Google investing in nuclear energy. Check it here!
AI’s Energy Crisis: Can Quantum Save the Day? ⚡️
AI is sucking up energy at an alarming rate.
Gartner predicts that AI could consume up to 3.5% of global electricity by 2030.
But what if quantum computing could change that?
Peter Chapman of IonQ, breaks down how quantum tech could reduce the power needed to fuel AI’s explosive growth and why it’s the next big thing in computing.
Join the conversation and ask Jordan and Peter questions on AI and energy here.
Also on the pod today:
• Mechanics of Quantum Computing ⚙️
• Barriers to Adopting Quantum Computing 🤔
• Quantum Computing’s Role in Energy Efficiency 🔋
It’ll be worth your 34 minutes:
Listen on our site:
Click to listen
Subscribe and listen on your favorite podcast platform
Listen on:
Here’s our favorite AI finds from across the web:
New AI Tool Spotlight – Focus Buddy is an AI productivity co-pilot, FAQWidget is an AI-powered FAQ generator and SeyftAI is a multi-modal content moderation platform.
Trending in AI – Alibaba’s international arm is claiming that its new AI translation tool beats Gemini and ChatGPT.
YouTube – YouTube is adding a ‘captured with camera’ label to videos that comply with C2PA Content Credentials.
YouTube is also testing “AI-enhanced” suggestions for comment replies.
Google - Google.org has announced $15 million in AI training grants for the government workforce.
LLMs – LatticeFlow’s LLM framework is attempting to benchmark compliance with EU’s AI Act.
Also, NVIDIA’s new Nemotron model (a tuned version of Llama 3.1 70B) is outperforming GPT-4o on some benchmarks.
AI Tech - Amazon has revamped its Kindle line with AI.
1. Mistral Unleashes Edge-Ready AI Models 🚀
Mistral has just launched “Les Ministraux,” a new line of generative AI models designed for edge devices, allowing everything from local translations to smart assistants without the need for internet access. With two models available—Ministral 3B and the more powerful 8B—these tools boast an impressive context window capable of processing around 128,000 tokens, equivalent to a 50-page book.
Priced competitively for developers, these models promise to outperform existing competitors like Llama and Gemma on essential benchmarks, making them a compelling choice for businesses prioritizing privacy and efficiency.
2. Perplexity Unveils Financial Analysis Tool 💸
Perplexity has just pulled back the curtain on its latest project: a finance-centric platform poised to revolutionize how users access and analyze financial information. This innovative tool not only provides real-time stock quotes and historical earnings reports but also facilitates detailed industry comparisons, allowing users to dive deep into company financials with an intuitive and visually appealing user interface.
A recent video demonstration showcased how users can effortlessly explore financial data for companies like Nvidia, track stock price trajectories, and even pose questions about company performance—all of which positions this tool as a potential game-changer for finance professionals and investors alike.
3. OpenAI Welcomes New CISO Amid Security Focus 🧑💼️
Dane Stuckey, former CISO of Palantir, has joined OpenAI as its new Chief Information Security Officer. Stuckey's extensive background in digital forensics and incident response positions him to enhance OpenAI's security measures, especially as the company aims to strengthen its ties with the U.S. Department of Defense.
His appointment comes at a crucial time when OpenAI is actively seeking to build secure AI infrastructure, reflecting a growing emphasis on compliance and user protection in the AI sector.
4. Major AI Models Struggle with EU Compliance 😬
A recent study from Reuters highlights that major AI models, including those from OpenAI and Meta, are facing significant hurdles in meeting new European regulations, particularly around cybersecurity and bias. The newly developed LLM Checker revealed concerning scores for discriminatory outputs, with OpenAI’s "GPT-3.5 Turbo" scoring just 0.46, raising alarms about human biases embedded in AI.
With fines looming at up to 7% of global turnover for non-compliance, tech companies may need to pivot their resources to address these gaps before the full implementation of the EU AI Act.
5. Lenovo Unveils AI Innovations at Tech World 2024 🤖
At Lenovo's Tech World 2024, the tech giant introduced a suite of AI innovations aimed at revolutionizing how we work, learn, and connect. Among the standout announcements is Lenovo AI Now, a local AI agent designed to turn PCs into personalized assistants while ensuring robust data protection, and the ThinkPad™ X1 2-in-1 Gen 10 Aura Edition, which boasts impressive hybrid work capabilities.
These advancements are set to streamline workflows and enhance productivity, catering to professionals and students alike in an increasingly digital landscape.
6. Boston Dynamics and TRI Join Forces for AI Humanoid Robotics 🤯
In a significant move for the robotics industry, Boston Dynamics has teamed up with the Toyota Research Institute (TRI) to integrate AI-driven intelligence into its electric Atlas humanoid robot. This collaboration aims to harness TRI's expertise in large behavior models, enhancing robots' capabilities to perform complex tasks with autonomy.
With Boston Dynamics already making strides in robotic hardware, this partnership could accelerate the development of versatile robots that tackle everyday tasks, potentially reshaping how we interact with machines.
Secret Microsoft Copilot Hack: Talk live with a webpage
We found a secret hack to be able to talk to a website using the edge browser and Copilot to interact with the web in a new way!
We show you how it works and ways to use it.
Check out today's AI in 5.
🦾How You Can Leverage:
There’s no denying the power of Generative AI.
But what about the actual POWER that’s …. Powering Generative AI?
As the demand for GenAI and large language models everywhere skyrockets, so too does the energy demand.
In short, the average request out of a LLM requires about 10X more energy than a traditional Google search.
So to keep up, we both need more energy to power the compute and also have to change the way computing works.
Luckily, we had an industry expert help break it down for us.
And trust us, understanding quantum computing is a tall task, yet Peter Chapman did an admirable job today on Everyday AI.
Peter is the President & CEO of IonQ, a quantum computing tech company that develops powerful quantum computers to help solve the future of technology.
So, will quantum computing solve the AI energy crisis?
Or, are there too many technical hurdles to clear?
Ready to power up?
Here’s what ya need to know.
1 – The big picture = LLMs need more power 💪
Generative AI is kinda like a growing teenage boy.
A bottomless pit you gotta keep feeding.
Feeding the future of AI is kinda one of IonQ’s ambitions, but Peter said quantum computing is tackling the problem from a different angle.
Because right now, everyone’s focused on the actual energy that GenAI needs, Peter said.
Big players like Google and Microsoft have recently shifted their focus to Nuclear power, but there might be a dual solution to the nonstop energy demands.
(You know… continuing to feed the teenage boy instead of addressing how hunger starts.)
In other words, IonQ is trying to change the way computers compute.
Try this:
Here’s a quick history lesson on traditional computing vs. quantum computing, and how it could impact Generative AI.
Peter likens traditional computing to trying to solve a maze.
With bits and bytes (0s and 1s), traditional computing would have to aimlessly wander down every tunnel, one at a time.
Quantum computing is different.
While traditional computers use 0s and 1s—checking each option one by one—quantum computing takes it further. Instead of only being 0 or 1, qubits (quantum bits) can be 0, 1, or both at the same time.
(Mind blown yet?)
This allows quantum computers to look at multiple possibilities all at once.
Peter explained it like this: a regular computer would wander through a maze, checking every path one by one.
But with quantum, it looks at all the paths at once, solving the maze in a single step.
When it comes to AI, especially large language models that eat up a ton of energy, this makes a big difference.
Change how it operates, and the need for power shrinks faster than humans in ‘Honey I shrunk the kids.’
Quantum computing could potentially swap out power-hungry GPUs for QPUs, cutting down energy use while speeding up performance.
Instead of just feeding the "hungry teenage boy" with more power, quantum changes how we approach the problem altogether.
2 – Quantum computing = not theoretical 🧠
Quantum computing isn’t some theoretical pie in the sky.
Forget everything you thought you knew about quantum computing.
It's not trapped in a lab or existing only on a chalkboard anymore. It's here, it's real, and it's probably running in a cloud near you right now.
Try this:
For as little as $10, Peter said you can run quantum algorithms on the cloud.
Platforms like Amazon Braket, Microsoft Azure, or Google Cloud offer this and you won’t even need a PhD in quantum.
You can even use Jupyter notebooks to write and run your first quantum programs, making it super easy to get started.
3 – How Quantum can feed GenAI 🍽
Forget the sci-fi — quantum computing is powering AI today.
(Ya know… kinda like flipping the hungry teenager conundrum.)
Quantum excels at tasks like linear algebra, a core component of AI, where classical computers struggle.
Try this:
IonQ is already building quantum processors, QPUs, to replace GPUs, aiming to deliver better AI models with less data.
Like that maze analogy.
And what comes with less data? Less power required.
In just 6-9 months, Peter said quantum could outperform traditional methods, transforming GenAI's efficiency.
It’s not about reinventing AI; it's about translating current problems into quantum's strengths, cutting energy use and enhancing AI performance.
If ya really wanna dive into the nitty gritty, go check out this great guide on why QPUs are the next GPUs.
⌚
Numbers to watch
40%
Out of the $79.2 billion total raised by cloud firms, 40% of all funding went to generative AI startups, according to a study by Accel.
Now This …
Let us know your thoughts!
Vote to see live results
Do you attend our livestreams?Every weekday, we bring you fresh AI insights, exclusive interviews, and breaking news with our Everyday AI livestream. |
Reply