This week in AI? 

Like a who’s who A-list lineup. 

• Google’s making moves. 

• OpenAI is (kinda) dropping updates. 

• Microsoft’s drawing lines in the sand. 

Sheeeeeeeeesh. 

And that’s just the beginning shorties.

You can spend hours a week (or day!) trying to keep up and understand what it all means. 

Or, you can kick it with the Everyday AI crew every Monday as we bring you the AI news that matters.

Ready for the breakdown? 

Same. 

Awkwwwwwaaaaard. 

Yeah, you heard that right – the same OpenAI they've invested $13 billion in and reportedly own a 49% equity stake in. OpenAI's new SearchGPT is apparently stepping on Microsoft’s toes, but here's the kicker: both companies are still using each other's tech. For the first time ever, Microsoft is listing OpenAI as a competitor.

GPT's chillin' in Windows 11 in its Copilot tech, while Bing's powering SearchGPT behind the scenes. Talk about a complicated relationship! Microsoft even had a non-voting board seat at OpenAI for a hot minute, but they bounced earlier this summer.

What it means:

This move could be about more than just business rivalry. With federal regulators eyeing big tech for antitrust violations, Microsoft might be playing it safe. But it also signals a potential shift in the AI landscape, where partnerships are fluid and competition is fierce. Companies need to be ready for rapid changes in alliances and be prepared to pivot their AI strategies accordingly. 

Grab the popcorn.

Y’all know we be covering all the deets. 

Google dropped Gemini 1.5 Pro version 0801, and it's topping the charts like a late August summer blockbuster. 

(We miss Theater releases, TBH.) 

This bad boy scored a whopping 1300 on the LMSYS AI arena leaderboard, leaving OpenAI's GPT-4 (1286) and Anthropic's Claude 3.5 (1271) in the dust. The updated Geminei 1.5 pro is flexing a massive 2,000,000 token context window, which is like giving it a photographic memory for entire libraries.

But here's the catch – you can't chat with it on the regular Gemini front end. This beast is only available through Google AI Studio and the Gemini API, so it's dev playground only for now.

What it means:

Google's making a serious play for AI dominance. The massive context window could change how we handle large-scale data analysis and long-form content generation.

But the limited access suggests Google's being cautious about deployment. Businesses should start exploring how to leverage these advanced capabilities through APIs, even if they can't access them directly through a chatbot interface yet.

Cmon Google — bring this to everyone!

NVIDIA's in the hot seat with reports flying about delays in their Blackwell chip series, but they're shutting it down hard.

Word on the street was that there might be a 3-month hold-up on their most advanced AI chips, which had tech giants like Microsoft, Meta, and Google breaking a sweat. But NVIDIA's not having it – they're saying production for Blackwell is right on schedule for the second half of this year.

What it means:

Even the hint of a delay in chip production can have ripple effects across the AI industry. (And, TBH, the entire U.S. economy, as we’re seeing today.) 

Companies heavily invested in AI should diversify their hardware strategies and have contingency plans for potential supply chain issues. It's also a reminder of how crucial hardware is to the AI race – software innovations are only as good as the chips they run on.

Google just unveiled Gemma 2B, and it's punching way above its weight class.

This tiny 2 billion parameter model is outperforming GPT-3.5 (which was the world’s marquee model like 20 months ago), which is packing a whopping 175 billion parameters. We're talking David taking down Goliath here. 

Gemma 2B scored an impressive 1126 on the Elo leaderboard, edging out GPT-3.5's 1106. And get this – it's small enough to run on your smartphone.

What it means:

This is a game-changer for edge AI and on-device processing. Smaller, more efficient models could lead to faster, more private AI applications.

Businesses should start thinking about how they can leverage these compact models for mobile apps, IoT devices, and scenarios where data privacy or low latency is crucial.

And as we’ve always said — the future of large language models is small language models y’all. 

OpenAI's (slowly) rolling out an advanced voice mode for ChatGPT Plus subscribers, and it's about to make your conversations with AI feel eerily human. 

This upgrade can pick up on emotional cues through your voice, and it's not just listening – it's responding with its own inflections, tones, and even breaths. 

Cool and creepy? 

The latency is so low, you'll forget you're not chatting with a real person. It's even got party tricks like sound effects and accents, though it's keeping that American twang even when speaking other languages.

What it means:

This could completely change how we interact with AI in daily life and work. Imagine voice-controlled AI assistants that can truly understand context and emotion.

Companies should start exploring voice-based AI interfaces for customer service, accessibility features, and hands-free work environments. But they'll also need to consider the ethical implications of more human-like AI interactions.

Meta's jumping into the custom AI game with both feet, rolling out their new AI Studio. 

This bad boy lets Instagram users whip up their own chatbots, following in the footsteps of OpenAI’s custom GPTs. You can give these bots a personality, a face, and even tell 'em what topics to steer clear of. It's all powered by Meta's latest Llama 3 model, which they're dishing out for free to developers, researchers, and startups. 

What it means:

Meta’s AI Studio is its play to keep Instagram users glued to their screens even longer. For businesses and influencers, it's a whole new way to engage with followers.

Imagine having a 24/7 brand ambassador that never sleeps and always stays on brand. But it's not just about social media – this move shows Meta's serious about competing in the AI space.

Looks like the tech giants are playing Monopoly with AI startups this week.

Canva just scooped up Leonardo AI, bringing 120 AI wizards into their fold to beef up their generative AI game. No word on the price tag, but you can bet it wasn't cheap.

Meanwhile, Google's playing it cool with more of an "acqui-hire" move. They've nabbed Character AI's co-founders, including CEO Noam Shazeer, who used to be a big shot at Google DeepMind. Google's also signed a deal to use Character AI's tech, which is like getting the cherry on top of their AI sundae.

What it means:

These moves show that the big players are hungry for AI talent and tech.

Canva's aiming to stay ahead in the design world by beefing up its AI capabilities, which could mean some serious competition for other design tools.

For Google, bringing back former employees with startup experience could inject some fresh innovation into their AI research.

Startups getting acquired after only two years? Doesn’t seem like a bad strategy to us. lolz.

Refer Friends to Win a Year of ChatGPT!

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