🤖 Google's Custom AI Supercomputer Leaves NVIDIA in the Dust

Was this email forwarded to you? Make sure to subscribe for more!

Greetings, fellow humans. 👋

This is Not A Bot - the newsletter about AI that was definitely not written by AI. I’m Haroon, founder of AI For Anyone, and I share the latest news, tools, and resources from the AI space.

Happy Monday! I have some exciting news to share with you this week about the future of Not A Bot. Stay tuned.

But for now, let’s dive into it…

Have a product, service, job, event, newsletter, app, book, movie, tool, or anything you'd like to share with over 40k subscribers? 

Advertise with us | Job board | Follow us on Twitter

🧵 In today's edition:

• 💾 Google's Custom AI Supercomputer Leaves NVIDIA in the Dust

• 🔎 Google to Add Conversational AI to Search

• ✍️ Microsoft's AI Copilot Coming to OneNote

• 🤑 AI Fundraising News

🤖 Top AI News

💾 Google's Custom AI Supercomputer Leaves NVIDIA in the Dust

Google has been working hard on its Tensor Processing Unit (TPU) chip, which is now in its fourth generation. And boy, has it paid off! Google has connected over 4,000 TPUs to create a supercomputer that's faster and more power-efficient than similar systems from Nvidia. They even threw in some custom optical switches to help connect individual machines. Talk about next-level engineering!

But wait, it gets even better. A comparison with Nvidia's A100 chip revealed that Google's system is 1.7 times faster and 1.9 times more power-efficient. That's some serious AI muscle right there! And to top it off, Google hinted that they might be working on a new TPU to compete with Nvidia's H100 chip. Looks like the AI arms race is heating up, folks!

So, what's the takeaway from all this? Google is definitely a force to be reckoned with in the world of AI supercomputers, and we can't wait to see what they come up with next.

Read more: Live Mint

🔎 Google to Add Conversational AI to Search

Google CEO Sundar Pichai is not one to shy away from a challenge, as he believes that the potential for AI chat is even bigger than before. In fact, Google is working hard to integrate large language models into its search engine. But, let's face it, Google is currently lagging behind Microsoft in this area. So much so, that they've declared a "code red" and are looking to more capable language models to catch up.

But fear not, Google is not one to give up without a fight! They're making efforts to increase collaboration between their AI divisions, Google Brain and DeepMind, to become more efficient. Who knows what kind of AI magic they'll come up with next?

Read more: Engadget

✍️ Microsoft's AI Copilot Coming to OneNote

Microsoft is rolling out its AI-powered Microsoft 365 Copilot assistant to OneNote, making your note-taking experience even smoother.

With Copilot, you can draft plans, generate ideas, create lists, and organize information like a pro. And the best part? Copilot can rewrite, format, or summarize existing text in OneNote based on prompts in a sidebar interface. Talk about a virtual assistant that's got your back!

Now, we know you're itching to get your hands on this AI goodness, but hold your horses. Microsoft is currently testing the private preview of Copilot in Word and Outlook, and hasn't announced when it will be available in OneNote. But hey, good things come to those who wait, right?

Read more: The Verge

🤑 AI Fundraising News

Engin Sciences Raises $2.5M - a New York-based AI-powered recruitment software startup

Native AI raises $3.5M - a market intelligence platform startup

A MESSAGE FROM PIXIEBRIX

Drive productivity with AI and automation

Leverage the power of AI and automation to streamline workflows and take productivity to new heights. PixieBrix is fully customizable and works on any site, allowing users to personalize any web application.

PixieBrix also offers quick-bar shortcuts and contextual sidebar information, and ready-to-use mods in its marketplace.

🗞️ Byte size: AI article summaries

Disclaimer: AI is (partially) used to summarize these articles.

AI Doesn’t Hallucinate. It Makes Things Up - The article debunks the idea that AI models can "hallucinate" and argues that it's not a correct explanation for AI mistakes. It also warns about the potential risks of releasing AI models and suggests that the language of AI "hallucinating" should be used carefully to avoid incorrect assumptions about AI capabilities. (Read more)

India Says It Won't Regulate AI - India's Ministry of Electronics and Information Technology announced no plans to regulate AI use, focusing instead on standardizing responsible AI guidelines and encouraging business. This contrasts with calls to pause large language model development due to security concerns. India aims to become a major player in the AI race by taking a hands-off approach. (Read more)

An Interview with Daniel Gross and Nat Friedman about the AI Product Revolution [Stratechery] (Link)

🐦️ Tweet of the day

And that does it for today's issue.

As always, thanks for reading. Have a great day, and see you next time! ✌️

— Haroon: (definitely) Not A Robot and @haroonchoudery on Twitter