😴 ModemWorks' Dream Recorder Pushes Sci-Fi Boundaries

Dutch design studio ModemWorks just launched its most surreal concept yet: an AI-enabled “Dream Recorder” that promises to convert your dreams into visuals. The prototype—part headband, part storytelling companion—records your sleep data and asks follow-up questions when you wake. 

Then, using AI models like Stable Diffusion and Runway, it generates a stylized video interpretation of what you might have dreamt. The project, titled Dream Out Loud, made headlines this week after going viral on several news outlets. But beneath the hype is a much bigger question: are we ready to let machines turn our unconscious mind into content?

🎨 Spatial Storytelling or Soft Surveillance?

Modem’s designers describe the experience as “spatial storytelling.” You wear the device while sleeping, then wake to a voice interface that probes your emotional state and recollections. These become prompts for image-generation AI, creating personalized visuals tied to dream themes like flying, nostalgia, or fear.

It’s a fascinating concept—but also a privacy puzzle. Your dreams are arguably your most private content. While Modem insists the device doesn’t transmit or store biometric data, skeptics warn that any product connecting dream logs with cloud-based AI models will raise ethical and security concerns.

🔮 Hope, Hype, and What Comes Next

Modem isn’t the first to try dream decoding. Neuroscience labs in Japan and California have explored fMRI + AI pairings to translate brain activity into rough visual approximations.

But Dream Out Loud is the first attempt to make it artistic—and accessible. Even if it’s more speculative art than hard tech right now, the project reflects a growing curiosity: can AI help us better understand what happens when our minds wander?

Whether this tech ends up in museums, app stores, or bedside tables remains to be seen. But it shows one thing clearly—our inner world is the next frontier in human-AI interaction.

🍎 Apple to Power Siri With Claude or ChatGPT

After years of lagging behind competitors, Apple is reportedly in talks to let ChatGPT or Claude power a revamped Siri experience. Apple is preparing to give iOS users a choice between top-tier AI models when they activate Siri—effectively transforming it from a rule-based assistant into a full-fledged LLM-powered sidekick.

Other reports suggest Apple will maintain tight on-device privacy controls but allow queries to be processed through OpenAI or Anthropic's models if users opt in. It’s a dramatic shift from Apple’s walled-garden approach to voice AI—and long overdue.

💡 Why It Matters

Apple has trailed Google, Amazon, and OpenAI in conversational AI for years. By tapping existing LLMs, it sidesteps the struggle to catch up and potentially leapfrogs ahead in user experience.

But it also raises questions. Will Apple design a unified voice interface or split Siri into a customizable tool? Will Claude and ChatGPT be rebranded or stay transparent? Either way, one thing’s clear: Siri 2.0 is coming, and it might actually be useful.

🔮 What’s Next

Apple’s refreshed Siri is targeted for iOS 26.4 in spring 2026. Internally, engineers have migrated the project to a new “V2” architecture after scrapping earlier code. Apple is benchmarking Claude and ChatGPT on its private cloud, and early tests reportedly favor Claude for stability and cost.

Although the public beta isn’t out yet, Apple previewed a Liquid Glass interface at WWDC—a hint of the design that will ship with iOS 26. If Apple can pair that UI with secure, on-device LLM smarts, Siri may finally graduate from voice assistant to true AI sidekick.

This Week’s Scoop 🍦

🎭 Artists and engineers release open letter: “Don’t panic about AI”

🧠 Microsoft AI tool outperforms doctors in diagnostic trials

⚽ China hosts first-ever AI robot football match

📈 Meta stock hits all-time high amid AI boom

🧑‍💼 OpenAI and Meta feud over poaching top AI talent

🧾 Pearson and Google deepen AI education alliance

👗 Weekly Challenge: Try Google’s Doppl for Virtual Fit Checks

Google’s newest Labs experiment, Doppl, turns a single full-body photo of you into a digital mannequin that can “wear” any outfit screenshot you upload. Think of it as a personal, AI-driven fitting room you can carry in your pocket.

🛠️ How to Get Started

1. Download the app on iOS or Android (U.S.-only for now).

2. Create your avatar by uploading a clear head-to-toe photo—or let Doppl generate an AI model if you’d rather not use your real image.

3. Add outfits by snapping pics from your camera roll, a web store, or even a magazine. Doppl maps fabric drape, lighting, and body pose in seconds.

💡 Four Ways to Use Doppl

• 🔄 Closet remix: Photograph clothes you already own, then mix-and-match them virtually before committing to an ensemble.

• 🛍️ Smart shopping: Screenshot items from online stores and see how they look on you before hitting “buy,” reducing costly returns.

• 🎥 Content creation: Fashion influencers can generate quick carousels of multiple looks without changing outfits IRL—great for “Get Ready With Me” videos.

• ✈️ Travel packing: Try potential outfits on your avatar and pack only what truly works, saving suitcase space.

🚦 Tips & Caveats

• Use bright, even lighting for your avatar photo; shadows confuse the AI cloth-mapping.

• High-resolution clothing images yield crisper fit previews.

• The tech is experimental: complex patterns and flowing fabrics (silk, chiffon) can still look stiff or distorted.

• Privacy note: avatar data is stored in your Google account and can be deleted anytime from the app settings.

💌 Respond to this email with your favorite look
Share the side-by-side before-and-after, and tell us—did Doppl nail your style or miss the mark?

Want to sponsor Jumble?

Click below ⬇️

That’s it for this week! AI’s reaching into our dreams and finally fixing Siri. What frontier will it explore next? Hit reply and let us know. See you next time!

Stay informed, stay curious, and stay ahead with Jumble!

Zoe from Jumble