Lead story / launch

Meta launches an AI outfit planner built around clothes you own

Meta has launched Style Sketch, an AI fashion app that builds outfit suggestions around clothes users already own. It can turn Instagram photos into a digital wardrobe, add virtual try-ons, and organize looks for a trip—not just recommend a single outfit.

Instagram is one starting point, rather than a requirement for supplying every item. Users can also photograph clothing, record short videos, or upload images from their camera roll. They can let the app suggest combinations or choose a specific piece and ask it to build an outfit around that item.

For virtual try-ons, the app description says outfits appear on a model matched to the user’s body shape and skin tone. That is Meta’s product description, not an independently verified finding about how accurately the preview represents clothing appearance or fit.

Travel planning adds another layer: users enter a destination, trip length, and planned activities, and Style Sketch produces day-by-day outfits that account for the itinerary and weather forecast. The app is available on iOS and Android only in Indonesia, Singapore, and Thailand. Meta has not said whether it will bring the app to more countries.

platform shift

Reddit will test making accounts the default as search referrals fall

Reddit will test a primarily logged-in experience in select markets outside the U.S. in Q4, with no global rollout date or settled access rules. Search referrals fell by more than 1 million per week on average in October. The company wants more control over AI scraping and better ad targeting, but account requirements are not proven to stop scraping. It says the logged-out front page, signed-in anonymity, and incognito browsing will remain.

security risk

CrowdStrike confirms nine intrusions in AI-assisted espionage campaign

CrowdStrike says China-linked spies used AI agents to target more than 200 organizations worldwide, with access and follow-on activity confirmed at nine. A human operator set objectives and redirected agents as they pursued multiple attack paths. The framework ran mainly on DeepSeek, plus an unidentified commercial model. The findings show coordinated use of established intrusion techniques, not an unattended operation or more than 200 confirmed breaches.

acquisition

OpenAI-backed Thrive Holdings agrees to buy ADT in a $15.3 billion deal

Thrive Holdings agreed to buy ADT for $10 per share in cash, with the deal valued at approximately $15.3 billion in enterprise value. Shareholder and U.S. antitrust approvals remain, with closing expected in Q1 2027. OpenAI is Thrive’s investor and technology partner, but no new AI security products were announced. ADT says its data policies will stay unchanged and new sharing will require customer permission.

Also worth knowing

•  Trump declares factory robot imports a security threat, but sets no tariffs

•  Google adds hundreds of free textbooks to Gemini Notebook

•  Proton CTO says AI wrote 90% of October code amid staff security concerns

•  Pika releases tools to generate video with sound locally

•  Mozilla fixes Firefox security flaws with help from OpenAI researchers

•  OpenAI updates model behavior rules to emphasize intent and human control

•  Researchers find gaps in AI agents’ identity and human approval records

•  Google launches an AI certificate for non-clinical healthcare tasks

•  DeepSeek and Moonshot used OpenAI outputs to train models, report alleges

•  Rivian licenses AI voice technology to Volkswagen through their joint venture

Daily tool drop

6 AI tools worth knowing today

The Internet Had a Point

Meme from the stashAI promised labor relief, delivered the opposite

Reader check-in

Help shape tomorrow's briefing

One click tells us what to keep, improve, or tighten.