• Everyday AI
  • Posts
  • How AI has officially outsmarted humans 😳

How AI has officially outsmarted humans 😳

💡 Google DeepMind's breakthrough discovery, Midjourney teases V6, AI finds a way to self replicate, using ChatGPT to summarize YouTube videos, and more!

Author

Everyday AI
December 18, 2023

Outsmart The Future

Sup y’all 👋

Random question — should I clone my voice and give y’all an option for AI Jordan to read the newsletter in the future?

Whatchyu think?

✌️

Jordan

Should I create an AI Jordan to read you this newsletter in audio form?

Login or Subscribe to participate in polls.

Today in Everyday AI
7 minute read

🎙 Daily Podcast Episode: Some people think AI is just a simple assistant. But what happens when it discovers something humanity has spent its lifetime to understand? Give it a listen.

🕵️‍♂️ Fresh Finds: Midjourney previews V6 and it looks delish, Market research in one click, US Department of Defense’s plans for AI, and Paris using AI to prepare for Olympics. Read on for Fresh Finds.

🗞 Byte Sized Daily AI News: Pakistan's former prime minister uses AI clone from jail, AI finds a way to self replicate, and ChatGPT 4.5 is…here. Maybe? For that and more, read on for Byte Sized News.

🚀 AI In 5: One of the best use cases for a ChatGPT plugin is being able to summarize long videos. We’re showing you our favorite YouTube summarizer, and some use-cases you probably haven’t thought of. See it here

🧠 Learn & Leveraging AI: Is AI really getting THAT much smarter than us? What does it mean for the future of humanity? Keep reading for that!

↩️ Don’t miss out: Did you miss our last newsletter? We talked about what we learned from the AI Summit NYC, OpenAI invests in superintelligent AI, and creating videos inside ChatGPT. Check it here!

When AI Outsmarts Humans - DeepMind's Historic Breakthrough 🤔

When GenAI first surfaced it was a great way to help write a paper or give you some new ideas.

But some people still see AI as a simple autocomplete.

In reality, GenAI has gone from a working assistant to LLMs making new discoveries.

We're talking about Google DeepMind's historic breakthrough and what it means for the future of GenAI and humans.

Also on the pod today:

• First instance of AI outsmarting humans? 😲
• AI creating its own intelligence 🧠
• Future of GenAI 🤖

It’ll be worth your 29 minutes:

Listen on our site:

Click to listen

Subscribe and listen on your favorite podcast platform

Listen on:

Upcoming Everyday AI Livestreams

Tuesday, December 19th, at 7:30 am CST ⬇️

And Later this week …

Friday, December 22 - GenAI in 2024: What’s coming and what it means for you (Jordan’s hot takes 🔥)

Here’s our favorite AI finds from across the web:

New AI Tool Spotlight – CapGo provides market research in one click, Pons lets you create a GenAI Christmas card, and Pixplain converts your screenshots into insights.

Trending in AI – The US Department of Defense has released a new statement on its path toward adoption of AI.

AI Visuals — Midjourney unveiled a sneak peek at its newest version V6. The results were … jaw-dropping.

Future of Work - A new survey shows that AI job losses are rising but the need for new AI jobs is also rising.

AI in Society – With the 2024 Olympics set to be hosted by Paris, France is turning to AI video surveillance for security. 

AI in Science – These autonomous submarines used AI to find their way around.

1. Imran Khan's AI Clone Campaigns from Jail 🇵🇰

Imran Khan, Pakistan's former prime minister, has taken his political campaign digital from behind bars, using an AI-generated voice clone. His party, PTI, used AI technology to deliver a speech in Khan's style, reaching over 4.5 million viewers across social media platforms. While the AI voice received mixed reactions, it signifies a new era in political campaigning, blending technology with traditional methods. 

2. ChatGPT 4.5 Upgrade Sparks Confusion 🤔

ChatGPT users were abuzz with speculation about a potential stealth upgrade to "GPT-4.5," sparking debates and excitement in the AI community. This rumor emerged after users reported responses from ChatGPT suggesting it was powered by a new AI model, "gpt-4.5-turbo." OpenAI staff, however, dismissed these claims as a "hallucination" by the AI, clarifying that no such upgrade had been silently released.

3. AI's Leap Towards Self-Replication 🔬

Scientists in the U.S. have made a groundbreaking advancement, creating AI capable of generating smaller AI systems without human intervention. This development, led by a collaboration between Aizip Inc., MIT, and several University of California campuses, demonstrates that large AI models can now autonomously design more specialized AI applications. These smaller AI models could revolutionize everyday technologies, from enhancing hearing aids to monitoring oil pipelines and tracking endangered species.

4. IBM Expands AI with $2.3 Billion Acquisition 💰

IBM has announced a significant $2.3 billion deal to acquire two units from German software company Software AG, enhancing its AI and hybrid cloud services. The transaction includes the purchase of StreamSets and webMethods, which will add data ingestion capabilities and integration tools to IBM's portfolio. This strategic move, expected to close in the second quarter of 2024, is aimed at bolstering IBM's position in the competitive AI and cloud computing market.

5. Deloitte Embraces AI to Navigate Workforce Challenges 🏢

Deloitte is innovatively using AI to potentially reassign existing employees to more promising roles within the firm. This strategic move aims to prevent mass layoffs while balancing new hiring growth, a challenge faced by many consulting firms in the current economic climate. By leveraging AI to assess employee skills and identify suitable new positions, Deloitte is betting on AI's ability to moderate hiring and enhance productivity.

YouTube Summaries ChatGPT Plugin Review 🎥

One of the best use cases for a ChatGPT plugin is being able to summarize long videos.

YouTube Summaries by eightify is a great plugin that gives you quick insights on any YouTube video.

We’re showing you how to use it!

🤷‍♂️ What’s Going On and Why It Matters:

AI is outsmarting humans now. 

Not as in — “Oh, this large language model can write faster than humans.”

Nah shorties. 

More like — “this language model solved a math problem that was unsolvable for decades.”

Sheesh. So why does this matter? 

That’s what we broke down today on Everyday AI. Google’s discovery was legit groundbreaking, marking the first time a language model discovered a solution to a previously unsolvable scientific problem. 

AI > humans. (Sorry, fellow humans.)

Here’s the dorky stuff we talked about: Google Deepmind’s used its FunSearch language model to discover a solution to correctly solve the cap set math problem. 

Simply put, the cap set problem involves finding the maximum number of points that can be placed on a grid so that no three points are in a straight line. 

(Yeah, we don’t really understand it either, but that’s probably why we wear sweatpants more than lab coats.) 

The cap set theory is important because it relates to fundamental questions in combinatorics, a branch of mathematics, and has implications for other areas such as computer science and matrix multiplication.

(Whew, tech speak is over) 

So, Google Deepmind’s achievement isn't just about AI being smarter than us; it's about AI creating new intelligence and pushing the boundaries of what's possible in problem-solving.

What else did we talk about in today’s show? 

  • AI creating new intelligence and discoveries, potentially outsmarting humans

  • DeepMind's new discovery, not derived from existing data, and the potential of large language models and generative AI to create new intelligence

  • AI moving beyond autocomplete to making new discoveries in generative AI and large language models

  • Large language models and generative AI rejecting incorrect answers at a scale 1,000 times that of humans

  • AI's transformative impact on various applications

Yeah, today’s show was a bit deep. We had to get some techy stuff in to close out 2023, right? 

Now, let’s talk about what it all actually means. 

🦾How You Can Leverage:

K. Got it. 

AI is outsmarting the smartest of humans now. 

(Cue sad music.) 

If you tuned in to today’s show, you now know that AI is outsmarting us and creating new discoveries. Yeah, a language model solved a (previously) unsolvable problem.

Don’t be discouraged, though. If you’re reading this newsletter, this is actually a good thing for your and your business. 

Here’s the scoop y’all. 

Heading into 2024, use this recent breakthrough as the cog to get GenAI and LLMs integrated into your day-to-day. 

If you STILL needed a reason or were looking for a GenAI sign in the sky, this recent Deepmind discovery is it. 

Now, here’s what it means for us all. 👇

1 – Go Deeper with Deepmind 🤿️

TBH, we’re Everyday AI is for Everyday people.

We fit that description too.

So although we did an entire episode on this new Deepmind discovery, we’re not experts in the cap set problem. Or the FunSearch language model. Or mathematical combinatorics. (Still not even sure what that means.)

Try this: After you hit replay on today’s show, make sure to check out the recent Deepmind news and read the entire study here.  

2 – Look at LLMs in a new light 💡

Hopefully this one has already been crossed off your list.

But if you STILL aren’t leveraging a LLM in your day-to-day, it’s time to pencil Large Language Models in to your 2024 plan.

Seeing how Google Deepmind used a language model to solve the cap set problem, think of the creative use-cases on how you can use LLMs to better analyze your offerings and solve customer problems.  LLMs are just content creators. They’re problem solvers. 

Try this: We’ve compiled dozens of videos (short and long!) on creative ways to use ChatGPT to grow your business. Check em out! 

3 – Prepare for what’s next 🔮

Today’s show changing your outlook on what can be accomplished with LLMs? 

Maybe.

Here’s the reality — breakthroughs like Google Deepmind’s cap set discovery are monumental. But by this time next year, we see breakthroughs like this being routine. 

Try this: Wanna see what’s actually gonna happen in the GenAI space moving forward?

Make sure to join Jordan Friday as he dives into GenAI in 2024: What’s coming and what it means for you. 

Now This …

We asked you all last Friday what type of content you wanted to see from us during our slight holiday break.

Between our email poll and from our livestream feedback, replays of Top Episodes of 2023 barely won out. Don’t worry, we’ll still be serving up the AI news and fresh finds each and every day in the newsletter!

Reply

or to participate.