New models and modes everywhere. 

Kinda. 

The past week in AI news was a mixture of hot drops, models announcements, and 

Meta was like Snoop Dogg at the Olympics and dropped Llama 3.1 like it was hot. 

On the other end? 

Apple’s hitting pause on its AI while OpenAI is just kinda teasing things. (But currently not releasing a whole lot.) 

So, what’s it all mean?

We gotchyuuuuuu. 

Almost every Monday, we bring you the AI news that matters. We cut through the fluff, find the hidden gems, and give it to you straight. 

Cuz we know that so much of your company and career growth depends on how you use AI. 

Without further chit-chat, let’s chat about the Chats. 

Lolz. 

Here’s what you need to know. 👇

Meta just dropped Llama 3.1, and it's a whole mood.

We're talking a 405 billion parameter model that's giving GPT-4 a run for its money. They beefed up their 8B and 70B models too, all open-source. It's like Meta's handing out free AI superpowers.

Now, here's where it gets juicy: Llama 3.1's 405B model is going toe-to-toe with GPT-4 Omni on the MMLU benchmark. That's like your high school's chess club suddenly taking on Garry Kasparov. And get this - it's all open-source. Meta's either feeling extra generous or playing 4D chess while we're all still figuring out checkers.

What it means:

Meta's not just dipping its toes in the AI pool; they're doing a cannonball and splashing everyone. This open-source move could be a total game-changer, letting developers and businesses tinker with top-tier AI without selling their firstborn or signing over their data souls. It's like Meta just handed out free tickets to the AI revolution.

But here's the real tea: Meta's confidence in going open-source might be the biggest flex since Elon Musk sent a car to space. They're basically saying, "Our AI's so fire, we dare you to make it hotter." It's a bold strategy, Cotton. Let's see if it pays off for 'em.

Wanna know more on Llama 3.1? Go listen/watch/read our breakdown y’all.

Want Apple’s new AI? You’re gonna have to wait. Even more.

Looks like Apple's "Intelligence" (because calling it AI is too mainstream for the cool kids in Cupertino) is pulling a diva move.

According to reports, they're pushing back their iOS 18 AI features to October, which means we won't see it strutting its stuff with the iPhone 16 launch in September. Classic Apple - always fashionably late to the party.

Important to remember here that there’s a lot going on under the Apple hood when it comes to ‘Apple Intelligence.’ They've got a local 3 billion parameter JAX/XLA Small Language Model for the quick stuff, and a direct integration with GPT-4 for when things get heavy.

But now we’ve all gotta wait even longer to give it a real-life test run.

However, Apple did JUST release iOS 18 to developers who are building on top of the platform.

What It Means:
Apple's doing the "fashionably late" thing to the AI party, but they're aiming to be the coolest kid in the room when they finally show up.

It's classic Apple - underpromise with great marketing, (maybe?) overdeliver, and make us all drool in the process.

But real talk: if you're gonna be THIS late to the game, you better bring the heat. No pressure, Apple. We're just all watching. No biggie.

If you want the full rundown on ‘Apple Intelligence,’ go read our big breakdown here.

Wall Street's finally caught the AI bug.

Morgan Stanley's got "AI at Morgan Stanley debrief" summarizing meetings and drafting emails. JPMorgan Chase is rolling out "LLM Suite" to 50,000 employees.

Morgan Stanley's flexing with their new toy, "AI at Morgan Stanley debrief." This bad boy summarizes video meetings and drafts follow-up emails faster than you can say "buy low, sell high." They've partnered up with OpenAI for this one, 'cause why reinvent the wheel when you can just slap some AI on it and call it innovation?

But hold onto your portfolios shorties, cuz JPMorgan Chase is going even harder. They're rolling out something called "LLM Suite" to a whopping 50,000 employees. That's 15% of their workforce suddenly getting AI superpowers. It's like Oprah, but instead of cars, everyone gets an AI assistant.

"You get an AI! You get an AI! Everybody gets an AI!"

What It Means:
Wall Street's finally trying to play catch up with Silicon Valley, and it's about dang time. These AI tools could flip the script on how financial advice is given and how investments are managed.

Remember when JPMorgan banned ChatGPT? At the time, we kinda scoffed and said, ‘yeah, this won’t last long.’

Sure enough, they just did a 180 so fast it'd make a day trader's head spin. Turns out, FOMO is a powerful motivator, even for the big banks as now they’ve gone all in on LLMs. They realized that banning AI was like trying to stop a bull market with a feather - pointless and potentially expensive.

OpenAI just dropped a search bombshell with SearchGPT, and it's got Google sweating more than a novice coder during a live demo. This isn't just another search engine; it's a whole new way of finding information that could make Google look like AskJeeves.

In short — OpenAI is taking the Perplexity route and trying to redefine how humans search. (And taking on Google in that process.)

SearchGPT is designed to give you more conversational, context-aware search experiences. It's like having a genius librarian, personal assistant, and mind reader all rolled into one. Right now, it's about to launch beta with just 10,000 lucky users, but the waitlist is open. So if you want to be part of the revolution, get in line, shorties.

What It Means:
This could be the biggest shake-up in search since Google told AltaVista "bye Felicia" back in the day. If SearchGPT takes off, it could change how we find information online faster than you can say "I'm feeling lucky."

But let's not get ahead of ourselves. Google's still the 800-pound gorilla in the room, and it’s not going down without a fight. Plus, there are some serious questions about privacy, web traffic, and digital business models that need answering.

And that whole OpenAI-New York Times lawsuit thang, which could really throw a monkey wrench in OpenAI’s long-term plans.

We give you an even more in-depth look at SearchGPT here.

According to reports, OpenAI may be a little cash-strapped.

A new report from The Information showed that OpenAI could be facing a potential $5 billion loss in 2024.

Yeah, you read that right. Billion with a 'B'.

OpenAI is spending a whopping $7 billion on training and inference for their models. That's like buying a small country's worth of GPUs. And $4 billy of that going straight to its partner Microsoft. (Sweet deal there, Satya!)

And get this - they're dropping $1.5 billion on staffing. For less than 600 employees. That's some expensive talent, y'all. Each employee is basically worth more than some startups.

What It Means:

OpenAI's playing the long game, but they're betting big.

Spending like there’s no tomorrow, and also preparing that tomorrow will exclusively be powered by GPT. lolz.

(But it for reals will be.)

This kind of spending shows OpenAI is all-in on AI domination, but it also raises questions about sustainability. Are they the Netflix of AI, burning cash now to rule later? Or are they the WeWork, flying too close to the sun on wings of neural networks?

Here's our hot take: that $20/month ChatGPT subscription? It's a steal. We’d expect these LLM subscriptions prices to go up across the board in the coming quarters. (Or, you get it for free in our Thanks a Million contest, that is.)

Refer Friends to Win a Year of ChatGPT!

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