Wednesday, January 14, 2026

☕️ Gemini introduces 'Personal Intelligence'

Receive a daily summary of what happened in tech, powered by ML and AI.

Thank you! We sent you a verification email.
Oops! Something went wrong while submitting the form.
Join 1,500+ thinkers, builders and investors.

Hi there, this is your daily ☕️ Techpresso.


In today's Techpresso:

🤖 Gemini introduces 'Personal Intelligence'

📱 Apple to customize Gemini AI without Google branding

🚫 China blocks Nvidia H200 imports despite US approval

🚗 Tesla FSD will soon become subscription-only

🕶️ Meta considers doubling Ray-Ban smart glasses production

📹 Google’s Veo now turns portrait images into vertical AI videos

🎁 + 21 other news you might like

🧰 + 6 trending tools

📚 + 4 trending papers

🤖 Gemini introduces 'Personal Intelligence' LINK

  • Google is launching Personal Intelligence, a beta feature in the Gemini app that lets the AI assistant connect across Gmail, Photos, Search, and YouTube history to deliver tailored responses.
  • The feature can reason across your data without being told where to look, such as suggesting all-weather tires after spotting family road trip photos or pulling a license plate number from Pictures.
  • Personal Intelligence is off by default and rolling out to Google AI Pro and AI Ultra subscribers in the U.S., with plans to expand to more countries and Gemini's free tier later.
  • 📱 Apple to customize Gemini AI without Google branding LINK

  • Apple will customize its version of Google's Gemini AI model independently and may remove Google branding from the experience, according to a new report from The Information about the partnership.
  • The Gemini-powered Siri will run on Apple devices and Private Cloud Compute, meaning Google won't have access to user data, and Apple expects it to answer questions directly instead of listing links.
  • Apple also wants the new Siri to handle unclear requests by searching through messages to identify contacts, like figuring out which contact is your mother even if not labeled that way.
  • 🚫 China blocks Nvidia H200 imports despite US approval LINK

  • China's customs officials blocked Nvidia H200 chip shipments on January 14, just one day after the US Commerce Department approved exports, leaving over $54 billion in orders stuck at checkpoints.
  • Chinese tech giants including Alibaba and ByteDance have ordered 2 million H200 chips, nearly three times Nvidia's current inventory, after three years of US export restrictions created massive pent-up demand.
  • The H200 outperforms Huawei's Ascend 910C by 32% in processing and 50% in memory bandwidth, forcing Chinese AI labs to choose between unreliable US supply chains or slower domestic chips.
  • 🚗 Tesla FSD will soon become subscription-only LINK

  • Tesla will stop selling its Full Self-Driving package as a one-time purchase after February 14, switching to a monthly subscription model only, CEO Elon Musk announced on X Wednesday.
  • The FSD package currently costs $8,000 extra for Model 3, Model Y, and the All-Wheel Drive Cybertruck, while other models like the Model S include it in the base price.
  • Despite Musk's repeated promises that FSD would turn Teslas into fully autonomous vehicles, the feature still requires active driver supervision and remains unavailable in most of Europe due to regulations.
  • 🕶️ Meta considers doubling Ray-Ban smart glasses production LINK

  • Meta and its manufacturing partner EssilorLuxottica are considering doubling production of Ray-Ban smart glasses to 20 million pairs by the end of 2026, responding to strong demand for the devices.
  • High demand in the US has forced Meta to delay the international launch of its Meta Ray-Ban Display smartglasses to fulfill existing orders, while the company also plans to cut 10% of Reality Labs staff.
  • Warby Parker stock rose 10% after the news, likely because the company is partnering with Google on incoming smartglasses, showing strong market interest in the overall smartglasses category this year.
  • 📹 Google’s Veo now turns portrait images into vertical AI videos LINK

  • Google's Veo 3.1 AI video tool can now turn portrait images into vertical videos, letting users create clips ready for TikTok and YouTube Shorts without needing to crop them manually afterward.
  • The Ingredients to Video feature, which generates clips from up to three reference images, now keeps characters looking the same across different scenes and lets users reuse objects and backgrounds consistently.
  • Google is also adding 4K upscaling for generated videos, up from the previous 1080p limit, and bringing these updated tools to the Gemini app, YouTube Shorts, and the YouTube Create app.
  • Other news you might like

    🧰 Trending tools

    0xCal: a minimalist calorie tracking app that uses AI to log meals from natural language descriptions instead of manual database searches.LINK

    Undiscord: bulk leave multiple Discord servers at once to declutter your sidebar without manually exiting each server individually.LINK

    nia-epstein-ai: indexes and searches 100M words of Epstein files using natural language queries with source document references and semantic search.LINK

    Redlight Greenlight for Claude Code: a keyboard shortcut overlay that lets you approve or reject Claude's terminal commands without switching windows, showing queued requests and execution context.LINK

    Lyra Music: a subscription-free music app that lets you listen to tracks without recurring payments, offering a straightforward alternative to monthly streaming services.LINK

    Nogic: a VS Code extension that visualizes codebase structure and relationships as an interactive graph for faster code exploration.LINK

    📚 Trending papers & reports

    Random game theory formula reveals structure: researchers proved that randomly generated competitive games have predictable average values, providing mathematical foundation for game analysis.LINK

    HiGP speeds up Gaussian process calculations by 100x: researchers created a Python tool that makes statistical predictions run 100 times faster than existing methods on large datasets.LINK

    Macaronic language tricks bypass AI image safety filters: researchers used mixed-language prompts to generate banned content, fooling safety systems 88% of the time across models.LINK

    AI chatbots function as entertainment products: researchers argue that people use AI assistants primarily for enjoyment rather than productivity, similar to social media or games.LINK


    Want to get the latest news differently? Find us on:

    twitter instagram spotify apple-podcasts


    See you tomorrow for a new dose of ☕️ Techpresso!

    Feeling behind on AI?

    You're not alone. Techpresso is a daily tech newsletter that tracks the latest tech trends and tools you need to know. Join 400,000+ professionals from top companies like OpenAI, Apple, Google and more. 100% FREE.
    Thank you! We sent you a verification email.
    Oops! Something went wrong while submitting the form.
    Join 1,500+ thinkers, builders and investors.
    You're in! Thanks for subscribing to Techpresso :)
    Oops! Something went wrong while submitting the form.
    Join 5,000+ thinkers, builders and investors.