iPhone is still the best AI device?

Hola Decoder😎

If someone forwarded this to you and you want to Decode the power of AI and be limitless, then subscribe now and Join Decode alongside 30k+ code-breakers untangling AI.

🍎 Apple is making the iphone its AI hub

Apple kicked off its first fall event under new CEO John Ternus with a packed hardware lineup, unveiling its first foldable iPhone Duo, iPhone 18 Pro lineup and AirPods 5, alongside Siri AI with iOS 27 and new tools for verifying photo authenticity. 

The Decode:

Apple says the iPhone already has what AI hardware needs - Ternus argued that an ideal AI device needs personal context, cameras, microphones, on-device processing, cloud access and constant availability, calling the iPhone an “intelligent personal hub” and saying, “The best AI device is the iPhone.” 

Privacy becomes Apple’s AI differentiator - Apple Intelligence will process data on-device whenever possible, with Ternus contrasting Apple’s approach with competitors that collect and store more personal information.

Siri AI finally arrives with iOS 27 - Launching September 14, the rebuilt assistant uses models trained with Google Gemini and can handle complex questions and multi-step requests. Support starts with the iPhone 15 Pro.

Apple wants to prove photos are authentic - Apple Reference Image captures signed sensor data alongside iPhone 18 Pro photos, creating a “digital negative” that can reveal later alterations. Apple will also support SynthID for identifying AI-created or edited images.

Apple’s AI strategy is taking shape around the hardware people already carry. Rather than introducing a separate AI device, it is adding a more capable assistant, private on-device processing and content verification directly into the iPhone ecosystem.

Together with AirOps

How High-Growth Marketing Teams Secure More Resources

75.4% of marketing teams face higher pipeline or revenue targets, yet only 43% received a budget increase. If your team is expected to close that gap, you need a stronger operating playbook.

On September 17, Krithika Shankarraman, the first marketing hire at Stripe and OpenAI, joins AirOps CMO Christy Roach to unpack findings from 300+ marketing leaders.

You’ll learn how to:

• Strengthen budget and headcount requests by tying investment to business growth
• Push back when expectations exceed resources while protecting work that moves pipeline
• Design a smaller, senior team with clearer ownership and greater leverage

Krithika built both marketing functions from the ground up. Christy has held leadership roles at AssemblyAI, Airtable, Asana, and Gusto. 

You’ll leave knowing where to invest and how to make a stronger case for the people and budget needed to hit your targets. 

Reserve your free spot and learn how to secure more resources for your team.

Can't attend the session live? Register anyway, and you’ll get the recording in 24 hours.

✈️ How to find the cheapest flight for any destination using ChatGPT

ChatGPT searches live prices, compares airlines, and tells you exactly when to book. Here's how.

Step 1: Switch to Work Mode Open ChatGPT and switch to Work Mode to enable live browser access.

Step 2: Run this prompt "I want to fly from [your city] to [destination] around [dates]. Search Google Flights, Skyscanner, and Kayak right now. Find me the cheapest available option including any stopover routes that save money. Compare economy and premium economy where the price difference is under $100."

ChatGPT opens each platform, pulls live pricing, and returns a side-by-side comparison.

Step 3: Ask it to find the cheapest date window "Now check the same route for the 3 days before and 3 days after my preferred dates. Tell me which exact date combination gives me the lowest total price."

ChatGPT tests every date combination and surfaces the cheapest window automatically.

Step 4: Check for promo codes and student or card discounts "Search for any active promo codes for the cheapest airline you found. Also check if this route has any credit card portal discounts through Chase, Amex, or Capital One."

Step 5: Get the direct booking link "Give me the direct booking link for the cheapest option with the best code applied."

Click, book, done. No aggregator fees. No missed deals.

⚠️ AI labs face new pressure over how far models should be allowed to go

Two separate developments are putting AI safeguards under scrutiny: an Anthropic researcher has quit over fears of recursive self-improving AI, while newly released documents show the Pentagon sought OpenAI models with “minimal refusal rates,” although both OpenAI and the Pentagon say that language never entered the final contract.

The Decode:

Anthropic researcher quits over self-improving AI - Jacob Coxon, who worked at OpenAI and Anthropic, resigned while accusing frontier labs of racing toward self-improving superintelligence despite believing it could pose catastrophic risks.

Anthropic researchers say alignment remains unsolved - Anthropic’s Evan Hubinger said his team believes there is a greater than 10% chance AI could kill humanity within a decade and acknowledged there is no clear plan for aligning superintelligence.

Pentagon documents sought AI that rarely refuses - A document released through a FOIA lawsuit described OpenAI “Mission Models” for national security with “minimal refusal rates,” suggesting the Pentagon wanted fewer model rejections during military use.

OpenAI disputes agreeing to those terms - OpenAI says it rejected the “minimal refusal rates” language and that the released document was an earlier draft. The Pentagon ultimately also said the phrase appears in no active OpenAI contract.

Both stories expose the same unresolved tension from different directions: AI systems are becoming more capable while governments and labs are simultaneously exploring how much control should remain around them. 

Together with Playbook Pro

Your Next 6 Months of Ad Spend Could Disappear in 6 Weeks

CAC is up 60% this year, and 70% of first-time buyers never return. Keep leaning on ads, and the budget you’re planning for the next half year could be gone before spring.

Rewriting the DTC Rulebook is the shortcut top operators are using to escape the trap, a 24-page report distilled from 200+ hours of real brand research and field tests.

Inside, you’ll learn how to:

🔥 Build emotional campaigns that double repeat-purchase rates in 90 days.
💰 Raise prices without losing volume using the Brand Value Canvas.
🧠 Cut 30% of paid-ad dependency with the Owned Media Checklist.

Download today and get 2,000+ viral video hooks already tested to grab attention in under 3 seconds, ready to plug into your next launch.

Stop feeding the ad machine. 

Grab the free report now and lock in sustainable growth before your next dollar disappears!

🏆 Tools you Cannot Miss:

🖥️ 49agents IDE – Organizes AI coding agents, terminals, repositories, files, and connected machines on one zoomable 2D canvas so developers can manage parallel agent sessions without juggling terminal windows.

🧭 Frigade Assist API – Gives an existing AI agent up-to-date knowledge of your product so it can answer user questions and guide people through actual in-app workflows.

📊 AdScope – Combines Meta and Google Ads performance into one dashboard, showing live creatives alongside spend, results, and cost-per-result data.

🛡️ Harden – Adds a security layer to AI coding workflows to identify risks introduced while agents generate and modify software.

🧩 Noodle Seed – Connects business data and workflows to AI conversations through a no-code agent-building environment.

🚀 Quick Hits

🌐 OpenAI says its AI agents wrote to several internet sites beyond the German wiki, with reports identifying at least 10 other websites including public wikis, text-storage sites and university link shorteners.

🔒 Microsoft, AFT and UFT introduced enforceable AI protections for U.S. schools, preventing student data from training AI, requiring human oversight and giving schools control over how data is collected, retained and deleted.

🎬 Amazon Prime Video launched AI-powered lip-sync dubbing that adjusts actors’ mouth movements to match translated human-recorded dialogue, starting with the English dub of German series Maxton Hall.

📚 OECD data from 760,000 students found AI users generally scored worse in science, but moderate, intentional use improved outcomes when students were taught to critically evaluate AI-generated information.

🛡️ OpenAI appointed prominent AI safety researcher Paul Christiano to its Foundation board, where he will join the Safety and Security Committee that has final authority over releasing new models. 

🦾 Prompt of the day

Create a Surreal Oversized-Face Fashion Portrait

Image Prompt: Create an ultra-realistic surreal editorial fashion photograph using the uploaded face image as the subject’s exact identity reference.

Context

• Subject: [Upload face reference image]
• Location: Exterior of a luxury clothing boutique
• Wardrobe: White sleeveless top, dark wide-leg trousers, black formal shoes
• Style: Surreal high-fashion editorial photography
• Key effect: Oversized head with multiple realistic arms

Instructions

Show the subject standing outside a luxury boutique, leaning slightly forward against a wooden wall beside a glass storefront. Replace the normally proportioned head with an intentionally oversized, hyper-realistic version of the same face, dramatically larger than the body, wearing a textured brown patterned cap. Add multiple anatomically realistic human arms emerging naturally around the upper body and oversized face, gently touching the cheeks and chin in an artistic editorial pose, with minimal silver rings and bracelets. Preserve the uploaded subject’s exact facial identity and recognizable features. Keep the surreal proportions intentional but physically believable, with realistic skin texture, shadows, anatomy, fabric detail, and interaction between the hands and face.

🤳AI Nugget of the Day

Thanks for Decoding with us🥳

Your feedback is the key to our code! Help us elevate your Decode experience by hitting reply and sharing your input on our content and style.

Keep deciphering the AI enigma, and we'll be back with more coded mysteries unraveled just for you!