🎧 Google turns headphones into translators

Good Morning! Google just turned every pair of headphones into a real-time universal translator with its latest iOS rollout. Plus, you’ll learn how to edit images using Photoshop inside ChatGPT.

Plus, in today’s AI newsletter:

• Google Turns Headphones Into Translators

• A New Library That Eliminates the DOM Layout Bottleneck

• Spotify Founder Is Selling Full-Body Scans

• How to Edit Images Using Photoshop Inside ChatGPT

• 4 new AI tools worth trying

AI NEWS

🎧 Google Turns Headphones Into Translators

Google has officially rolled out its Gemini-powered "Live Translate" feature to iOS, allowing users to understand over 70 languages in real-time. Unlike Apple's built-in translation which requires specific high-end AirPods, Google's feature works right through any connected headphones.

• The feature uses Gemini's speech-to-speech capabilities to preserve the original speaker’s tone, cadence, and emotion, ditching the flat, robotic voice of traditional translators.

• Users simply open the Translate app, tap "Live translate," and connect their headphones to start listening to real-time translated audio.

• Originally launched on Android, the feature is now expanding on both iOS and Android to more countries including the UK, France, Germany, Japan, and Spain.

• Social media users are already calling it a "killer feature," with many joking that instant AI translation might put language-learning apps like Duolingo out of business.

This is the real-world utility AI was built for. By capturing not just the exact words, but the actual vibe and emotion of the speaker, Google is breaking down massive global communication barriers. More importantly, making this hardware-agnostic gives Google a massive strategic edge over Apple's walled-garden approach, putting frontier AI translation into the hands of anyone with a smartphone and a pair of standard earbuds.

AI NEWS

⚡️ A New Library That Eliminates the DOM Layout Bottleneck

Cheng Lou has released Pretext, a pure JavaScript/TypeScript library for multiline text measurement and layout. The repository is already going viral with over 11,000 stars on GitHub by completely side-stepping the need for expensive DOM measurements.

• Eliminates the need for DOM layout and reflows (like getBoundingClientRect or offsetHeight), which are some of the most expensive operations in the browser

• Uses the browser's own font engine as ground truth to calculate exact dimensions, making it a highly AI-friendly iteration method

• Blazing fast performance: preparing a batch of 500 texts takes just ~17ms, while the actual layout calculations happen in a blistering ~0.10ms

• Fully supports all languages out of the box, including emojis, complex scripts, and mixed-bidi text

• Unlocks complex UI capabilities like true virtualization, custom masonry layouts, and preventing layout shifts when new text loads

Text measurement has historically been a massive headache in front-end development because it forces the browser to pause and recalculate the entire page layout. By moving this logic into pure JS/TS without ever touching the DOM, Pretext gives developers, and AI coding agents, the ability to perfectly calculate and render complex text boundaries instantly.

AI NEWS

🩺 Spotify Founder Is Selling Full-Body Scans

Daniel Ek is bringing his next ambitious project stateside. Neko Health, the Swedish startup offering comprehensive full-body health scans, is preparing to launch its first US clinic in New York as early as this spring, pending regulatory approval.

• Uses a combination of advanced imaging and blood tests to check for skin cancer, heart disease, stroke, and diabetes risk.

• Currently operates in Sweden and the UK, charging around £299 ($400) for a comprehensive scan.

• US pricing is expected to be higher due to the increased costs of medical staffing and real estate.

• Neko Health was founded in 2018 as Ek began planning his next massive venture following Spotify's IPO.

The transition of tech billionaires from software into deep health tech is rapidly accelerating. If Neko Health can successfully navigate the massive complexities of the US healthcare and regulatory systems, it could normalize walk-in, comprehensive health data collection, effectively turning preventative medical care into a highly scalable consumer tech product.

HOW TO AI

🎨 How to Edit Images Using Photoshop Inside ChatGPT

In this tutorial, you will learn how to seamlessly integrate Adobe Photoshop into your ChatGPT workspace, allowing you to make quick, natural-language image edits without ever needing to open the full desktop software or pay for a subscription.

🧰 Who is This For

• People who hate coding but need dashboards

• Founders tracking metrics without devs

• Product managers monitoring KPIs

• Marketers tracking campaigns in real time

STEP 1: Connect the Photoshop App

Head over to ChatGPT and click on the "Apps" button located in your left sidebar. Locate "Adobe Photoshop" in the apps list, and click on the "Connect" button in the top right corner of the page to link the tool to your account.

STEP 2: Upload and Describe Your Edit

Once the connection is established, open a fresh chat window. Click the "+" icon in the prompt bar and select "Adobe Photoshop" from the menu. Upload the image you want to work on, and simply type out your desired changes using natural language. For example, you can tell the AI to “Make the background black and white.”

STEP 3: Adjust and Fine-Tune with Sliders

ChatGPT will process your request and display the edited image directly in the chat, complete with an interactive Intensity slider so you can manually control the strength of the applied effect. You can stack further edits by continuing the conversation. For instance, if you type "Adjust the exposure on the image," ChatGPT will bring up Photoshop's native exposure sliders, allowing you to fine-tune the lighting to perfection.

STEP 4: Export or Expand Your Edits

Because this integration prioritizes speed and ease of use, it intentionally leaves out the overwhelming complexity of the full desktop app. It is perfect for fast, everyday tweaks. However, if your image needs more advanced retouching, every creation includes an "Open in Photoshop" button. Clicking this will seamlessly launch your file in Photoshop on the web, where you can continue refining your work with a more comprehensive set of tools.

Bluesky's CEO talks about Attie, a new agentic social app built on Bluesky's AT Protocol that uses Claude and lets users build custom feed.

Report analyzing payments of 28M US consumers shows Claude adding paid subs at a steadily increasing pace; Anthropic: paid subs have more than doubled this year.

An AI-generated TikTok parody of reality series Love Island, called Fruit Love Island, averaged 10M+ views across its first 21 episodes after debuting last week.

Cohere launches Transcribe, its first voice model; the 2B-parameter, open-source speech recognition model handles tasks like notetaking and speech analysis.

🎶 Lyria 3 Pro: Google’s AI that makes longer music tracks

🌐 MolmoWeb: Ai2’s open-source AI that can browse the web

🔎 Model Council: Perplexity tool to query multiple AI models at once

⚙️ OpenAI Codex: OpenAI’s coding assistant, now with automations and customizable themes

THAT’S IT FOR TODAY

Thanks for making it to the end! I put my heart into every email I send, I hope you are enjoying it. Let me know your thoughts so I can make the next one even better!

See you tomorrow :)

- Dr. Alvaro Cintas