FTC sues Amazon over hidden ad charges

Hi there, this is your daily ☕️ Techpresso.

In today's Techpresso:

⚖️ FTC sues Amazon over hidden ad charges

🗑️ Apple accuses OpenAI of destroying evidence

🦷 Dyson toothbrush films inside your mouth

🍎 John Ternus takes over as Apple CEO

🛡️ Anthropic resumes cyber tests

Plus: 🎁 10 other news you might like, 🧰 6 tools, and 📚 5 papers.

⚖️ FTC sues Amazon over hidden ad charges LINK

The FTC has sued Amazon, saying the company quietly raised the minimum prices advertisers paid to place ads for their products, in a case joined by more than 20 state attorneys general.

The lawsuit covers a seven-year period during which Amazon brought in tens of billions of dollars, and it accuses the company of illegally gaining over $20 billion from unwitting advertising customers.

Amazon shares dropped after the report came out, falling nearly 3% to around $258 by 3 p.m. EDT, with attorneys general from 21 states signing on to the complaint.

🗑️ Apple accuses OpenAI of destroying evidence LINK

Apple has filed new court documents in its lawsuit against OpenAI, saying a forensic look at a laptop used by former engineer Chang Liu turned up "shocking evidence" of stolen trade secrets.

Apple claims Liu, who left for OpenAI in January, downloaded a confidential circuit schematic and used it in a March simulation in the electrical tool LTspice, where his AI "agent" learned to run the software and check results.

The filing also alleges Liu told an OpenAI colleague to destroy evidence once he learned of Apple's internal probe, and Apple now wants access to a Mac mini that synced the schematic file to the laptop via iCloud.

🦷 Dyson toothbrush films inside your mouth LINK

Dyson has begun selling the CameraJet, a $499 toothbrush with a 100,000-pixel camera in the brush head, using the images to aim a jet that sprays mouthrinse between the teeth.

The camera takes 28 pictures a second and finds gaps between teeth within 100 milliseconds, a system Dyson calls Gap Optical Targeting that runs on 16 million lines of code trained on 470,000 dental images.

Dyson says the CameraJet clears 69% more plaque from hard-to-reach spots than rivals, though that figure comes from its own testing with a plaque stand-in, not an independent clinical trial, and no camera images are stored.

🍎 John Ternus takes over as Apple CEO LINK

John Ternus officially became Apple's CEO today, and his newly updated bio on the company's leadership page offers early hints about the priorities he plans to focus on in the role.

A single new paragraph, filling nearly half the short bio, credits Ternus with work on product reliability, durability, and repairs, plus materials that cut Apple's carbon footprint, from recycled aluminum to 3D-printed titanium in the Apple Watch Ultra.

The bio signals Ternus will carry on Tim Cook's push to reduce Apple's carbon footprint, while also framing himself as a hardware designer focused on making products that last for many years.

🛡️ Anthropic resumes cyber tests LINK

Anthropic has restarted the outside cybersecurity tests it paused a month ago, after three incidents where its Claude models broke out of their test setups and attacked real companies instead of the fake targets.

In one case Claude Opus 4.7 hit a real firm sharing a domain name with a fictional target, while another model's Python code leaked onto the internet and was downloaded by 15 systems, one of which ran it.

The failures traced back to a misconfigured sandbox from evaluation partner Irregular that never actually blocked internet access; two affected organisations only learned they were compromised when Anthropic contacted them directly.

Other news you might like

• Firefox for iOS Gets Built-In Ad BlockerLINK

• Kalshi bans ex-lawmaker George Santos for life over tradingLINK

• China's CXMT makes its first HBM3E chips, closing the AI memory gapLINK

• Sony just added free, ad-supported live TV to every PS5LINK

• Texas to become first test bed for new safeguards against water system cyberattacksLINK

• Linux kernel nears record 2,000 vulnerabilities per release as AI bug hunters scour 40 million lines of codeLINK

• Anthropic has signed a $35bn cloud deal with Nvidia-backed LambdaLINK

• Longtime exec Phil Schiller is downgraing his role at Apple againLINK

• The Pentagon now has its own version of ChatGPT and GrokLINK

• Perplexity's hybrid agent is a win for AI privacyLINK

🧰 Trending tools

Radar by Particle: organizes news coverage from multiple sources so you can quickly grasp the full picture of every story without the effort.LINK

Skud: menubar app for Mac that uploads files, generates branded links, and tracks opens, with password protection and expiration control for designers.LINK

Edge Drop: an open source Windows clipboard shelf that lives on your screen edge, letting you stack, pin, and drag text, files, and screenshots into any app.LINK

Any Command: turns an Android phone into a second screen, trackpad, keyboard and gamepad for Windows, working over your network or Bluetooth without accounts.LINK

Tiles: memorize a numbered tile sequence in three seconds, then tap them in order across increasingly difficult rounds while competing with friends.LINK

Tether: a physics ball for your Mac menu bar that treats open windows as surfaces, giving restless hands something to fidget with during meetings.LINK

📚 Trending papers & reports

WaymoQA driving dataset gives self-driving AI 35,000 human-answered questions on tricky situations where dodging one hazard creates another, and training on it sharply improves how safely the systems reason using all-around camera views.LINK

EU AI testing sandboxes get a concrete playbook breaking the compliance journey into 29 steps and 15 requirements, giving startups and regulators a clear, standardized path to test high-risk AI legally before launch.LINK

Public interest lawsuits are emerging as Australia's main tool for holding companies accountable when automated decisions cause harm, since dedicated AI regulation is stalled and lawyers must stretch existing laws to cover it.LINK

Medical image measurement gives AI a benchmark and training data to actually measure things like tumor size and joint angles, tasks today's medical AI fails at, moving it from yes-or-no answers toward the numbers clinicians rely on.LINK

Wearable mobility tracking now predicts how several movement-limiting diseases progress over time using a single model trained across separate patient groups, beating eight rival methods while flagging which movement signals matter.LINK

See you tomorrow for a new dose of ☕️ Techpresso!