Y'all, this week in AI has been hotter than a single GPU trying to run Dall-E 3. 

(Why you still look so cartoonish though?)

We've got Google employees raising eyebrows over military contracts, Salesforce dropping AI sales assistants, and California cooking up some controversial AI legislation. 

Aaaaaannnnnd, Microsoft's reportedly unleashing that feature that had privacy advocates on high alert. 

Grab your favorite caffeinated beverage, shorties. 

Let's dive into this week's AI rollercoaster.

Here’s what ya need to know. 👇

At least 200 DeepMind employees are giving Google the side-eye over its military contracts, especially with the Israeli army.

Employees of Google DeepMind circulated an internal letter back in May, arguing that military involvement undermines DeepMind's position as a leader in responsible AI.

What it means:

This isn't just Google's headache.

It's a wake-up call for the whole tech industry.

As AI gets more powerful, these ethical dilemmas are gonna pop up more often. Companies need to start thinking hard about where they draw the line, or they might find their top talent heading for the exits.

You gonna hand over all your CRM duties to a Salesforce AI helper? 

The company kinda wants you to. 

Salesforce just dropped two AI tools that might make your sales team feel like they've got superpowers.

First up, the Einstein SDR agent – a 24/7 inbound lead nurturing machine.

Then there's the Einstein sales coach agent, an AI mentor for your pitches.

These AI sidekicks aim to tackle a big problem: sales pros reportedly spend 70% of their time on non-selling tasks. 

That's a lot of coffee runs and spreadsheet wrestling.

What it means: 

This could be a game-changer for sales teams. 

But it's not just about boosting numbers. It's about freeing up your human talent to do what they do best – build relationships and close deals that need that human touch.

Companies that find the right balance between AI efficiency and human ingenuity are gonna come out on top.

The new AI bill in California has officially divided Silicon Valley.

California's trying to regulate AI with the Safe and Secure Innovation for Frontier Artificial Intelligence Models Act (SSIFAIMA, if that’s a real acronym)

The Safe and Secure Innovation for Frontier AI Models Act (SB 1047) is a California bill aimed at regulating high-cost AI models. Here’s the high level.

• Targets AI models that cost over $100 million to train and need a lot of computing power.

• Requires safety checks and plans to reduce risks.

• Companies must report any safety incidents.

• Offers protections for whistleblowers.

• Creates a new Frontier Model Division for oversight.

Silicon Valley is still mixed.

Meta, OpenAI, and some startups are against it, while Anthropic supports it with some changes.

Google and Microsoft have a more divided view. The big issue is finding a balance between innovation and safety.

What it means: 

This isn't just California's show. 

If this bill passes, it could change the AI game globally.

 Companies might need to pump the brakes on their AI development and think more about the consequences. It's time to start baking ethics into your AI strategy now, before the law makes you.

According to reports, Google's dropping $250 million over 5 years on California newsrooms to sidestep a bill that could've cost them billions.

 It's $180 million for news outlets (sorry, broadcasters) and $70 million for AI tools to boost journalism.

This deal helps Google avoid a proposed state bill that would've required them to compensate for linking to news articles. 

That bill estimated Google and Meta could owe US publishers up to $13.9 billion annually. 

Yikes.

That $250 milly seems like a bargain, no? 

What it means: 

This is about more than Google saving some cash. It's about figuring out how news and AI can coexist. 

For businesses in media and tech, this is a glimpse into the future. We're watching new models for content creation and distribution emerge in real-time. Time to start thinking about how you'll navigate this new landscape.

Keep an eye on how similar legislation might find its way to Perplexity and OpenAI’s doorsteps. 

Microsoft's launching Recall for Windows Insiders in October, reports are showing.

 It's like giving your PC a photographic memory, automatically taking screenshots for later searches.

Privacy folks are raising red flags, but Microsoft says it'll be off by default and wrapped in tight security.

TBH, we don’t hate it at all. 

It'll only work on compatible "Copilot Plus PCs" designed to handle AI workloads locally.

 No timeline for wider release yet, but Microsoft's eyeing that holiday season hype.

What it means: 

This could be super useful, but it comes with some pretty big privacy concerns and grey area. 

For businesses, it's time to think about policies for these kinds of tools before your employees start using them. It's a reminder that the line between helpful AI and privacy concerns is pretty thin.

A recent report highlights that over half of Fortune 500 companies are now mentioning AI as a potential risk in their SEC filings. 

A study by Arize AI found that nearly two-thirds of these companies mentioned AI in their latest reports, with one in five specifically referencing generative AI.

Only 31% discuss AI benefits outside the risk sections.

The most likely industries to disclose AI risk?

Media, technology, telecom, healthcare, and financial services.

What it means: 

Studies are showing that the AI honeymoon is winding down. 

Companies are realizing that just slapping "AI-powered" on everything isn't a magic solution. 

It's time to get real about AI.

Focus on how it actually fits into your strategy, and for Pete's sake, train your dang people, y’all! 

Throwing AI at untrained employees is a recipe for disaster, yet that’s just about how everyone company is doing it. 

(We can help with that.) 

Matt Garman, AWS's big head honcho, says AI could take over coding tasks within two years. 

He's predicting that being a dev in 2025 will look totally different than in 2020.

But don't panic yet. 

Garman's not saying devs are going extinct; he's saying they need to evolve.

It's time to focus on the stuff AI can't do (yet) – like innovation and understanding customer needs.

What it means: 

This is your wake-up call, tech world. 

AWS said that AI-assisted coding and development has saved them 4,500 developer years worth of work.

(Seriously.) 

The days of just churning out code the old school way could be numbered. 

For businesses, it's time to rethink how you build and manage tech teams. You need people who can work with AI, not compete against it. 

Start investing in upskilling your devs now, or you might find yourself with a team that can't keep up with the machines.

What do you think? 

Refer Friends to Win a Year of ChatGPT!

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