Major reversals šŸ”

Apple heard the whispers

Welcome back! Here’s something I’ve been thinking about a lot lately: Hollywood can’t quite make up its mind on AI. For every headline about the promise of fully or partially AI-generated films, there’s another about how automation represents the industry’s downfall. I’m curious what you think.

Would you watch a movie that was made exclusively with AI tools?

Login or Subscribe to participate in polls.

Quick programming note: We’re off on Friday for the Fourth of July. Enjoy your long weekend and I’ll see you back here next week.

US Senators scrap AI regulation proposal from budget bill

Sen. Marsha Blackburn (R-TN) pulled her support for the provision on Monday in a deal with other lawmakers. Via Anna Moneymaker / Getty Images.

It’s not often that US Senators vote 99-1, but they just did on a controversial AI proposal in President Donald Trump’s sweeping tax bill. 

In the version of the bill passed by the Senate yesterday, lawmakers nearly unanimously voted at the last minute to scrap a proposal that would have prevented states from regulating AI for a decade.

The details: 

  • The original legislation proposed a 10-year ban on state-level AI regulation of any kind. If a state passed any AI regulation, it couldn’t tap into key broadband funding from a $500 million federal program.

  • Proponents said: Patchwork state-by-state legislation around AI hinders progress and keeps the US chasing China on tech. Microsoft, Meta, and Andreessen Horowitz were among the tech companies pushing for the moratorium.

  • Critics said: The regulation would have been a hall pass for an industry avoiding accountability for the impact of its products.

The bigger picture: North of 1,000 AI bills have been proposed at the state level, but Congress hasn’t passed any of its own larger-scale regulations. Expect the tech industry to keep pushing for limits on state- and local-level AI legislation to make the job of following the law of the land (and scaling their businesses) a little easier.

Apple may ditch its own AI for Claude or ChatGPT

Apple is reportedly considering using OpenAI or Anthropic LLMs to power its next-gen Siri. This is a major reversal that could signal Apple is reading the writing on the wall: It’s behind on generative AI.

What happens now? Apple has asked both OpenAI and Anthropic to train custom versions of their models to run on Apple’s Private Cloud Compute servers, Bloomberg reports. While no deal has been finalized, executives reportedly think Claude is better suited to handle Siri’s needs than Apple’s homegrown models.

Quick context: Apple delayed its AI-enabled Siri (originally slated for 2025) until next year at the earliest following a slew of technical challenges. Bringing in third-party models could fast-track Apple’s effort to match Android competitors and close the feature gap. Licensing third-party AI is already a play run by Samsung (with its features leveraging Gemini) and Amazon (which works with Anthropic to help power the new Alexa+).

Why it matters: Handing the reins to Anthropic or OpenAI would mark a major shift for Apple. But after news the company recently considered acquiring Perplexity to beef up its AI efforts, is it really that surprising?

Microsoft tests multi-agent AI for medical diagnosis

Microsoft’s latest AI system diagnosed patient cases four times more accurately than human doctors in a recent benchmark test.

How it works: 

  • MAI-DxO uses a ā€œchain-of-debateā€ orchestration system that mimics how multiple physicians might collaborate on a diagnosis.

  • It draws on models like GPT-4, Gemini, Claude, and Llama to analyze symptoms, select follow-up tests, and arrive at a conclusion with 80% accuracy. 

  • Human doctors scored just 20% under the same constraints.

The system also reduced costs by 20% by recommending more efficient diagnostic routes. Microsoft researchers say the approach could be a major step toward lowering healthcare costs—if the tool can prove itself in real-world clinical trials.

Why it matters: While AI has already helped radiologists and specialists, this is one of the clearest examples of a general-purpose diagnostic model that mirrors doctors’ reasoning. Microsoft hasn’t committed to commercial rollout yet, but the company is exploring integrations that could eventually put this tech into consumers’ hands.

Missed Cisco Live 2025? Get Caught Up Now

The future of AI isn’t just hype—it’s architecture. Cisco is leading the charge, helping enterprises everywhere unlock AI’s full potential with speed, scale, and trust.

Now, you can see it all unfold, even if you couldn’t make Cisco Live 2025.

  • Watch on-demand keynotes + deep dives

  • See how Cisco’s AI-ready infrastructure powers the next era

  • Learn how to scale securely, intelligently, and fast

Ready to meet the future now?

Turn static visuals into motion with AI

Magic Animator

Magic Animator analyzes your uploaded image and uses AI to identify elements like text, shapes, and logos, then automatically applies animation patterns like fades, bounces, and slides. You can customize the motion, preview results instantly, and export in various formats.

How you can use it:

  • Animate social media posts without manual keyframing

  • Add motion to app UIs for prototypes

  • Create quick animated logos for presentations or branding

  • Leverage directly inside Figma, Canva, and Adobe Express

Pricing: Free & Paid

Turn any room into a design showroom

Decofy is a one-click interior design assistant that turns any room photo into a shoppable, AI-styled makeover. Upload a photo of your room and watch it transform instantly into a fully styled space.

How you can use it:

  • Upload a photo of your room to get instant, AI-generated redesigns

  • Customize the space with text prompts

  • Explore curated design themes for inspirations

  • Shop the exact pieces shown in your design with product links

Pricing: Free & Paid

Jobs, announcements, and big ideas

  • Baidu just dropped the latest version of its Ernie model—and it’s being called China’s biggest public AI release since DeepSeek. The new model is now open to the public, and according to state media, it boasts over 300 million users across Baidu’s ecosystem.

  • OpenAI sticks with Nvidia, passing on Google’s AI chips (for now).

  • Cloudflare locks out AI bots and rolls out paywalls for data access.

  • Meta is adding voice calls for businesses on WhatsApp and is testing AI-powered voice recommendations.

  • Google is using its Veo model to animate iconic photographs from the Harley-Davidson Museum.

Udio unveiled a visual editing suite mid-lawsuit. But that’s only one of many wild AI moves this past week—catch up here.

That’s a wrap! See you next week. Have a great Fourth of July weekend!

—Matt (FutureTools.io)

P.S. This newsletter is 100% written by a human. Okay, maybe 96%.