Take a week off AI, and you might not understand how search engines work. 

Or your computer. 

Real talk — big tech has been at it with AI-powered updates, shoving them in our faces whether we’re stoked or scared. 

Meta, Microsoft, Google, OpenAI and others are bringing us new tech all at once. 

Yay or Yikes? 

Don’t worry — we’ve got the 101 on what’s happening and what it means. 

Don’t waste countless hours each day trying to keep up with how AI updates impact your biz. We do that for you with our weekly AI News that Matters on Mondays. 

So, let’s cut through the fluff, see what’s actually cracking, and dissect what it all means. 

Let’s get it. 👇

Meta's unveiled Movie Gen, a generative AI video model that might give OpenAI's Sora a run for its money. 

(Well, if either company actually releases these models publicly.) 

This new Meta Movie Gen powerhouse can create 16-second video snippets from a single text prompt and personalize them with just one photo. It's got serious editing chops too, letting users swap objects in videos with ease.

Movie Gen's not available to the public yet, but Meta's teasing us with a 90-page research paper. 

(We’d use NotebookLM to just turn that into a podcast, TBH.)

Meta being kinda coy about the training data, only mentioning a mix of "licensed and publicly available content."

Pretty typical for AI video generators. 

What it means: 

The video creation landscape is gettin wild y’all. 

Sora was first announced about 8 months ago, and just about everyone else has jumped on board. 

While Meta and OpenAI have just teased releases, we’ve seen pretty impressive AI video models from the likes of Runway, Kling, Luma Dream Machine and others. 

Businesses need to prepare for a world where high-quality, AI-generated video content is the norm. 

This will change marketing, entertainment, and social media. Buuuuut, it also raises serious questions about copyright and the potential for misuse. Companies should start developing policies now for AI-generated content and its verification.

OpenAI just secured a financial package that's got everyone's jaws on the floor. 

We're talking a $4 billion revolving credit line, bringing their total liquidity to over $10 billion. 

But wait, there's more! 

As we talked about last week, the LLM Giant also closed a $6.6 billion fundraising round, valuing the company at a mind-boggling $157 billion. 

Their revenue? It shot up 1,700% year-over-year, raking in $300 million last month alone. They're projecting $11.6 billion in sales for next year, but don't pop the champagne just yet – they're also eyeing a $5 billion loss this year. 

Turns out, those GPUs and (kinda?) maintaining top talent ain't cheap, shorties.

What it means: 

OpenAI's gearing up for some serious AI flexing. This cash influx could widen the gap between OpenAI and its competitors. 

Expect new models, features, and maybe even some AGI whispers in the near future. 

Businesses, take note: OpenAI's moves could reshape the AI landscape. Keep a close eye on their developments and start thinking about how to adapt or compete.

And knowing that Microsoft AND Apple are relying on OpenAI’s models, you have no choice but to pay close attention. 

Maybe Microsoft updates of yesteryear made you think of a couple of small new bells and whistles? 

Some recent updates are changing how your PC works, with the Windows 11 2024 Copilot updates. 

They're rolling it out gradually, starting with Qualcomm-powered Arm devices in October. 

The star of the show? The often-delayed but much-hyped feature called "Recall" that gives your PC AI-powered photographic memory. 

(Don't worry, it's off by default. Microsoft knows we're all paranoid about our digital footprints.) 

There's also this nifty "Click To Do" feature for quick AI actions, and Windows Search is finally getting smart enough to understand what you're actually looking for. Oh, and they're jazzing up Photos and Paint with some AI magic too.

What it means: 

Your PC's about to get a whole lot smarter, whether you like it or not. 

(But we all want it, right?) 

Businesses, brace for impact – this could seriously boost productivity, but it'll also mean a learning curve for your team. IT departments, get ready for some interesting support calls.

Microsoft's rolling out a new Copilot interface and voice chat feature, aiming for a "warm and inviting" vibe. 

They've borrowed some design cues from Inflection AI, hoping to make Copilot feel more like a friendly neighborhood AI. 

It looks like it’s slowly rolling out, first to paid users and then eventually to free. 

What it means: 

Microsoft's pushing hard to make Copilot more user-friendly, but they've still got some catching up to do. We gave the new Copilot voice mode a try and weren’t super impressed. 

It’s an early version, but Microsoft has got some work to do in order to catch up to OpenAI’s counterpart, Advanced Voice Mode.

Bet no one had this on their AI bingo card? 

Nvidia just pulled a move so bold, it's got the whole AI world doing a double-take. 

They've unleashed NVLM 1.0, a family of open-source AI models that's making proprietary systems sweat. 

The star? NVLM-D-72B, a 72 billion parameter beast that's flexing hard on both vision and language tasks. Some researchers are saying it's going toe-to-toe with bigger models like Llama 3.5 in math and coding. 

NVIDIA’s not just releasing the model – they're promising to drop the training code too.\

What it means: 

NVIDIA just threw a wrench in the AI industry's gears. 

This open-source/open weights move could spark an innovation boom that makes the app explosion look like a firecracker. 

Companies relying on proprietary AI might need to rethink their strategy. For smaller businesses and startups, this could be a game-changer – high-quality AI models without the hefty price tag. 

But with great AI comes great responsibility – we're gonna need some serious ethical guidelines to keep things from going sideways. Cuz when you make a model open source and open weights, the bad guys benefit, too. 

OpenAI just turned DevDay into an all-you-can-eat buffet for coders. 

(Don’t worry… us everyday people will ultimately benefit. Lolz.) 

They've introduced model distillation, letting devs give their mini models a glow-up using the big boys' makeup. 

There's a new prompt caching feature that'll save some serious cash on those common prompts. The real showstopper? 

The Realtime API. It's about to make speech-to-speech apps smoother than a freshly waxed surfboard.

What it means: 

The AI app explosion is about to go kaboomy. 

Every dev from Silicon Valley to Silicon Alley is gonna be pumping out AI-powered apps faster than flipping flapjacks. 

(LOVE pancakes, BTW.) 

Businesses, take note: integrating these new APIs could give you a serious edge in efficiency and user experience. 

Especially the real-time API. That could become the new standard for the future of customer service. 

Google's been feeling the heat since ChatGPT burst onto the scene, but they're not throwing in the towel. 

According to reporting from the Wall Street Journal, Google is cooking up some serious AI reasoning software, aiming to go toe-to-toe with OpenAI's latest and greatest. 

They're using this fancy thing called chain-of-thought prompting, but on the front-end of the model instead of the user prompt engineering on the back end. 

This forces the AI to ponder before spitting out answers. 

What it means: 

Everyone loves strawberries. 

(Oh, the famous? Mr. Strawberry even popped by the livestream audience today.) 

Google mighta been slow out of the gate, but they're making up for lost time after OpenAI released its ‘Strawberry’ o1 reasoning model. 

With Google joining the reasoning model game? This could lead to more sophisticated problem-solving tools and AI assistants across Google's product lineup.

Google's not just playing in the AI sandbox – they're rebuilding the whole AI search playground with these updates. 

They've added generative AI to Lens, so you can point your camera at stuff and ask questions about it. They're also rolling out AI-organized search results, starting with recipes and meal ideas and even sticking ads in the AI overviews. 

What it means: 

Searching for stuff is about to get a whole lot smarter and more intuitive. 

TBH — about time. Tired of pretending Alexa and Siri know left from right. 

What should you be focused on? 

Businesses need to adapt their SEO strategies for this new visual and AI-powered search world. 

Content creators, get ready to optimize for AI-generated snippets and visual search. 

(We just barely knew ya, voice only search.) 

The line between online and offline information is blurring – this could change how we interact with the world around us.

A new Canvas for your next AI masterpiece? 

OpenAI's latest update is rolled out with its Canvas mode now live, and it's about to make coding and content creation smoother than a fresh jar of skippy. 

(Or the Costco version.) 

OpenAI’s newest side-by-side mode is a direct competitor to Anthropic Claude's 'Artifacts' feature, and is now available to all ChatGPT Plus users. 

Canvas can convert code between programming languages, debug your code on the fly, and even suggest ways to improve it in an interactive way. 

 But here's the kicker – it's not just for coders. Canvas also brings a whole new interface to ChatGPT with split-screen and in-line editing for any type of content creation. 

What it means: 

We're looking at a world where writer's block becomes as outdated as a floppy disk. 

(And also the death of a few dozen more GPT Wrappers. TBH, the only thing they still had going for them was the ability to work together with AI in a document. Which is now the main draw of Canvas in our view.) 

Developers could see a serious productivity boost with the code conversion and review features. 

Content creators — and literally anyone — get ready for a more intuitive way to collaborate with AI. 

This might be one of the first mainstream LLM ways to “work with” AI in real-time, and not just sending short bursts of tasks to a model then finessing the output. 

Do you wanna see us do a deeper dive in Canvas? 

Lettuce know by voting in the poll at the top of this newsletter, or hit us with a reply.