AI Academy
Posts
🤖 Google Goes Nuclear To Power Its AI

🤖 Google Goes Nuclear To Power Its AI

Plus: Fighting AI Misinformation and AI Helps Win Nobel Prize

October 17, 2024

Hello AI Enthusiast,

This week in AI, DeepMind's founders won a Nobel Prize in Chemistry, showing how AI can help science evolve, while Anthropic's CEO's bold predictions paint a bright picture of AI's future. But it's not all smooth sailing. As the 2024 US elections get closer, Anthropic and OpenAI declared they are working hard to stop AI from spreading false information. And powering all this AI? It takes a lot of energy, so much that Google is looking into nuclear. We'll dive into these stories shortly.

But first, save the date: October 23rd!
Join our free webinar where Gianluca Mauro will unveil a new opportunity for those who want to get practical with AI. We're working on something special to help you stay constantly updated, keep learning, and keep doing. Don't miss it.

The Big Picture 🔊

AI's Promise in Science and Beyond

Anthropic's CEO predicts "powerful AI" by 2026, surpassing humans in biology, engineering, and literature, potentially solving global issues and extending lifespans. This vision gains credibility as DeepMind's founders win half the 2024 Nobel Prize in Chemistry for their AI-driven protein structure breakthrough, AlphaFold.

Gianluca MauroCEO and AI Rockstar

AlphaFold wins a Nobel prize with "traditional" AI solving real problems, while Amodei sketches a rosy AI future. His optimism is personal, not objective truth.

He's right about needing diverse expert input, but his vague predictions seem tailored more for investors than the public. It's the kind of big-picture talk that tends to catch an investor's ear.

How do Anthropic CEO's bold predictions about AI's future make you feel?

Anthropic and OpenAI Try to Combat Election Misinformation

With the 2024 US elections approaching, Anthropic is blocking political campaign use of its AI and running human audits to catch misuse, while OpenAI is tracking AI's election influence, finding limited impact so far. Both declared they are steering users to reliable voting info and monitoring AI content trends to safeguard the electoral process.

Helin YontarCPO and Polyglot

While it's great to see AI companies taking steps to prevent misuse, we can't ignore the cases they might miss. They're showing they're doing whatever they can to prevent this, which is what regulators want, but complete control is impossible.

I wonder if "AI did it" will become the new excuse for election losses.

Google to Power Data Centers with Nuclear Energy

Google has partnered with Kairos Power to build seven small nuclear reactors by 2030 to power its data centers with 500 megawatts of carbon-free energy, addressing the rising demand from AI and cloud services. This initiative is part of Google’s strategy to secure sustainable energy.

Gioele MottarliniCOO and Image Addict

Energy is clearly a bottleneck for AI development. While Google goes nuclear, OpenAI's CEO Altman invests in fusion, and Musk tried gas turbines for XAI.

Each approach has its critics. The race for AI power shows we have an energy problem, but the ideal solution remains elusive.

Bits and Bobs 🗞️

Adobe unveiled new AI-powered features at the Adobe Max conference, including enhanced tools in Photoshop like automatic distraction removal and an updated Firefly generative AI model.
Adobe has introduced new generative AI video tools, including Generative Extend for small tweaks in video editing and Text-to-Video and Image-to-Video for creating short clips.
At a recent Tesla event, CEO Elon Musk unveiled its Robovan concept, an autonomous vehicle that can carry up to 20 passengers, designed for both personal transport and commercial use. Production is not expected before 2027.
At the same event, Musk showcased the Optimus Gen 2, a humanoid robot envisioned as a versatile household companion capable of tasks like babysitting, dog walking, and grocery shopping, with an estimated future price between $20,000 and $30,000.
Wikipedia editors are increasingly battling AI-generated content, leading to fake entries, challenging Wikipedia's quality control efforts.
Anthropic has introduced a Batches API allowing developers to process up to 10,000 queries per batch with a 50% cost reduction, completed in under 24 hours.
Walmart is testing its new AI tool, Wallaby, a suite of retail-focused language models trained on Walmart data to enhance customer and employee interactions.

From Our Channels 🤳

Google's plan to power AI with small nuclear reactors has sparked discussion. Here's Gianluca's take on this energy strategy in his TikTok video:

@gianluca.mauro
Google’s Nuclear powered AI data centers #Ai #learnontiktok #artificialintelligence #business #machinelearning #product #ux #entrepreneur

LOLgorithms 😂

At Tesla's reveal, it looked like the entire "I, Robot" cast showed up, but Will Smith must have missed the memo.

That's a wrap on our newsletter! Here’s a quick recap before you go:

Generative AI Project Bootcamp: Learn to craft optimal prompts and integrate them into automated AI prototypes to accelerate processes and test business ideas.
Practical Introduction to ChatGPT: A free course on using ChatGPT confidently, understanding its workings, and exploring its potential.
Customized Corporate Training: Equip your team with the skills they need to unlock the potential of AI in your business.

Catch you next week! 👋