• Everyday AI
  • Posts
  • OpenAI’s updated model, Elon and XAI release Grok-2 and more – AI News That Matters

OpenAI’s updated model, Elon and XAI release Grok-2 and more – AI News That Matters

AMD's $49B acquisition, Google teasing new Gemini models, OpenAI SearchGPT waitlist, Trump's AI Taylor Swift controversy and more!

Author

Everyday AI
August 19, 2024

Outsmart The Future

Today in Everyday AI
7 minute read

🎙 Daily Podcast Episode: OpenAI’s new AI model, a huge flaw of xAI’s Grok-2 and new 'Personhood Credentials' are kinda scary. Cut through he fluff this week’s AI news that matters. Give it a watch/listen/read.

🕵️‍♂️ Fresh Finds: Google’s new AI feature for Androids, why Apple will win the AI game, Google teasing new Gemini version and what an AI model collapse actually means. Read on for Fresh Finds.

🗞 Byte Sized Daily AI News: AMD’s $49B acquisition, OpenAI closes SearchGPT waitlist and NVIDIA’s new weather model. For that and more, read on for Byte Sized News.

🚀 AI In 5: Here’s an AI research hack that allows you to use multiple AI chatbots at once. See it here

🧠 AI News That Matters: Did you hear about OpenAI’s new model? What major flaw does Grok-2 have? And what’s this new Personhood Credentials thing? Here’s our breakdown of the AI news that matters. Keep reading for that!

↩️ Don’t miss out: Did you miss our last newsletter? We talked about an AWS leader speaking on AI success, Google AI Overview updates and former Google CEO drama. Check it here!

AI News That Matters - July 29th, 2024 📰

Did you know OpenAI released a new model?

The one fatal flaw of Elon's new Grok-2 that no one's talking about. What is it?

And is it concerning that researchers are trying to create 'Personhood Credentials' to differentiate humans from AI pretending to be humans?

Join the conversation and ask Jordan any questions on AI here.

Also on the pod today:

• Politics and AI 🏛
• World Labs' achievement 🧪
• OpenAI’s updated GPT 4 model ✨

It’ll be worth your 38 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 – Coldreach helps you find ready-to-buy leads, Ellipsis provides AI code reviews & bug fixes and Central automates legal work to become a compliant U.S. startup.

LLM News — It appears that Google may be testing out an update for its main model, Gemini. We just spotted “Gemini-test” in the Chatbot Arena. 

Google – Google has added a new AI feature for Androids that locks out phone thieves.

Apple – Although Apple has arrived late to the AI game, Wall Street has it winning big on AI.

AI Images — Meta released a Tuning-Free personalized image generation feature.

AI in Politics - This Wyoming mayoral candidate is still pushing for an AI bot powered by OpenAI to run the capital city.

Read This – An expert explains what an AI model collapse actually means for ‘AI doom.’

AI Models – The spike in recent data center deals show that AI is headed for nuclear.

1. AMD’s $49 Billion Acquisition of ZT Systems 💰

AMD has made a significant move in the AI landscape by acquiring ZT Systems for $49 billion, combining cash and stock with potential performance payments. This strategic acquisition aims to enhance AMD’s capabilities in AI systems design and data center solutions, allowing for improved efficiency in AI training and inferencing.

With ZT Systems’ expertise, AMD is poised to compete more effectively against industry giants like NVIDIA and Intel.

2. OpenAI’s SearchGPT Waitlist Closed ❌

OpenAI has officially closed the waitlist for its much-anticipated SearchGPT, leaving many eager users disappointed after receiving rejection emails.

With only 10,000 users selected to test this prototype generative search tool, the news raises questions about its potential impact on the future of online searching. This exclusivity might signal major changes in how we interact with search technology.

(We weren’t selected. Were you?)

3. ElevenLabs Launches Global Reader App with 32 Languages 🗣

ElevenLabs has made a significant move in the synthetic voice industry by expanding its Reader app globally and adding 32 languages, including Portuguese, Spanish, French, and Hindi.

Users can now upload text, PDFs, and e-books to listen in a variety of voices, featuring iconic actors like Judy Garland and James Dean. The newly introduced Turbo v2.5 model enhances the app's performance by reducing latency and improving audio quality.

4. NVIDIA’s New AI Weather Model Could Change Forecasting ☔️

NVIDIA just unveiled its new generative AI model, StormCast, which promises to predict extreme weather events with great accuracy. This groundbreaking model operates on a small scale, tackling nasty phenomena like tornadoes and flash floods while outperforming the current US weather prediction champ, the HRRR.

With AI running circles around traditional models, experts are buzzing about how this could save lives and property with spot-on short-term forecasts.

(Bless up that weather predictions may be getting more accurate.)

5. Fortune 500 Companies Raise Concerns Over AI Risks 🚨

A recent report from Arize AI indicates that 281 Fortune 500 companies are now highlighting risks associated with artificial intelligence, marking a significant 562% increase from the previous year.

The media and entertainment sector is particularly affected, with an alarming 917% of companies expressing concerns, as industry leaders like Netflix and Disney worry about the potential impact on their revenues. Additionally, Salesforce warns that AI could pose serious reputational risks if mishandled.

Unitree Robotics G1: Ready to Roll! 🤖

Unitree Robotics is making headlines with its G1 humanoid robot, now set for mass production at a competitive price of $16,000. This sleek AI-powered model showcases impressive mobility, boasting 23 degrees of freedom in its joints, and features like a visor-like face and three-digit hands that enhance its functionality.

Following a swift evolution from a $90,000 prototype, the G1's advancements could open new doors for businesses and research institutions seeking cost-effective robotic solutions.

Could Figure, Optimus and others have some serious competition on their hands?

How to Research with Multiple AI Chatbots

Are you tired of spending countless hours on research and competitive analysis?

We’re showing you how you can run prompts through multiple AI models like GPT-4, Claude, and Gemini to save time on research.

Hardly no one noticed OpenAI’s new model. 

How come? 

Why are researchers warning us about advanced AI systems pretending to be human? 

And did…. Taylor Swift really endorse former President Donald Trump? 

So many questions, and we’ve got the answers. 

Don’t spend countless hours each week trying to keep up with AI. It’s legit not possible. (Our team spends a gross amount of time each day trying.) 

So instead of wading the AI waters in place, just come for your weekly AI swimming lesson to stay afloat. 

Each Monday, we bring you the AI News That Matters. 

We cut through the fluff, B.S. and big tech press releases to give it to you straight. 

Ready to get ahead? 

Let’s g(r)oooooooooooo(w)! 👇

1 – Grok 2: Elon's AI Chatbot Enters the Ring 🐦

Elon Musk's XAI just dropped Grok-2 and Grok-2 Mini, and they're not messing around. 

These new models are flexing some serious muscle in reasoning and performance benchmarks. 

Grok-2 scored an 87.5 on general knowledge, slightly edging out the old GPT-4 Turbo.

So although Grok-2 ain’t scoring off the benchmarking charts, it’s proven itself to be in the Top 5 LLM conversation. 

But here's the kicker - it's available for X (formerly Twitter) premium users at just $7-8 a month. That's like one of your Oat Milk Frappe lattes y’all. 

Oh, and did we mention it's integrated with Flux.1? That's the open-source image generator that's been turning heads faster than a celebrity scandal.

Why it matters:

Don’t have your company go all in on Grok just yet. Lolz. 

Grok's got some quirks that might make you raise an eyebrow. It's trained on X data, which is about as reliable as a weather forecast in Florida. Even basic X/Twitter-related AI questions fall a bit flat. 

Pluuussssss, it's got a nasty habit of getting stuck in conversational loops. 

Which to us means that Elon’s xAI team did not do basic Quality Assurance testing. 

We tested it, and if you try to summarize a long web page, Grok throws a fit and gets caught in a cycle. You'd have to start a new chat to get it working again. Not exactly smooth sailing for the average user.

2 – Digital ID Cards for Humans? AI Researchers Say It Might Be Necessary 🤔

We’re not sure if this is super weird or super needed. 

Or both. 

Researchers at OpenAI, Microsoft, MIT, Harvard, Oxford and others  - basically the Avengers of the AI world - just dropped a paper that's got everyone talking. 

They're proposing "personhood credentials," which is fancy talk for digital ID cards to prove you're not a bot.

Why it matters:

This isn't just about spotting fake accounts. 

It's about creating a digital world where you can trust that the "person" you're talking to isn't an advanced AI in a trench coat trying to sell you snake oil. 

The researchers are worried about advanced AI systems that can mimic human behavior online, distort public discourse, and spread misinformation faster than gossip at a high school reunion.

But here's the real piping hot tea: This proposal came out right after the internet exploded over some Twitter accounts that seemed too smart to be human. 

Coincidence?

Maybe. 

Maybe not. 

As we barrel towards the US election, this kind of AI-powered misinformation is going to be everyone's problem, whether you like it or not.

Even current AI systems are sophisticated enough to fool the most astute humans looking for scams. 

But future models? 

Sheeesh. 

That’s gonna be one area that you’ve gotta pay real close attention to. 

3 – World Labs: The $1 Billion Baby of AI 🦄

World Labs just pulled off a Usain Bolt in the startup world, sprinting to unicorn status in just 4 months. 

They secured a $100 million funding round, valuing the company at over $1 billion.

That's "billion" with a "B," folks.

Founded by Fei-Fei Li, often called the "godmother of AI," World Labs is tackling one of AI's biggest headaches: accurately estimating 3D characteristics of real-world objects without needing a truckload of data.

Why it matters:

This could be the secret sauce for everything from next-level gaming to robots that don't trip over their own feet.

But more importantly, many AI big brains think this 3D understanding is one of the last big hurdles to achieving AGI (Artificial General Intelligence). 

You know, the kind of AI that makes us humans start updating our resumes.

Right now, getting good 3D data is about as easy as finding a needle in a digital haystack. 

It's mostly collected through costly methods by autonomous vehicle companies. If World Labs cracks this nut, it could revolutionize how we train AI to understand and interact with the real world.

4 – Trump's AI Campaign Strategy: Swifties and Shenanigans 🎭

Former President Trump is stirring the AI pot again. 

He posted AI-generated images claiming Taylor Swift fans (Swifties) support his campaign, complete with a picture of Taylor pointing and saying, "Taylor wants you to vote for Trump."

But wait, there's more! (*Infomercial voice*)

He also shared an AI image of Kamala Harris speaking at a "communist rally" (spoiler: it wasn't real) and last week claimed a real crowd photo of Harris supporters was AI-generated (plot twist: it wasn't).

Why it matters:

Lettuce get this one outta the way.

We’re NOT getting political here.

These are facts. (And troubling at that.) 

Regardless of political affiliation, if a U.S. Presidential candidate misuses AI, we’ll be covering it here at Everyday AI. 

And this ain't just about Trump, y'all. 

It's a wake-up call that AI-powered political shenanigans are here, and they're not going away.

We're talking AI-generated images, cloned voices, and bots that can understand and steer online conversations. As we barrel towards the 2024 election, this stuff is going to be everywhere.

Pro tip: Before you share that mind-blowing political post, take a hot second to fact-check. The internet's about to get trickier than a house of mirrors.

5 – ChatGPT Reclaims the AI Throne 👑

OpenAI just pulled a sneaky upgrade on us, and ChatGPT is back on top like it never left.

They released a new model called "GPT-4-0-latest" (catchy, right? Lolz), and it's crushing it in the LMSYS chatbot arena.

How crushing?

Its Elo score is about 17 points higher than the next closest model. In the world of AI, that's like lapping your competition in a race.

Why it matters:

This quiet upgrade is packing more punch than a double espresso.

OpenAI announced this with all the fanfare of a library whisper - just one tweet.

No fancy event, no press release barrage. Just a "oh, by the way, we're top dog again" kind of vibe.

⌚

Numbers to watch

74%

74% of IT pros see AI making their skills obsolete.

Now This …

What are your thoughts?

Vote to see live results

If our newsletter was sent at a more consistent time, would you read it more?

Be honest. Our feelings won't be hurt

Login or Subscribe to participate in polls.

(Or your fave LLM like Claude, Gemini, Copilot, etc)

Reply

or to participate.