OpenAI developing removable, screen‑free smart speaker with camera, designed by LoveFrom, as Demis Hassabis urges US‑led A‑G‑I watchdog

Share your love

AI Tech News Today — Quick Roundup

Hello, this is Aurora and I’m Isabelle with AI Tech News Today.
Salute — the date of this news is 2026-07-15T13:11:44.000Z.
Today we have a quick roundup of the biggest AI stories you need to know.


1) OpenAI consumer device relies on motion and cameras

Source: Bloomberg

  • What it is: OpenAI is building a removable, screen-free smart speaker with a camera and motion sensors.
  • Design: Developed with Jony Ive’s firm LoveFrom to move room to room and act as a humanlike AI companion for cooking, chores, messaging, media, and smart home control.
  • Scope: The product is one of roughly five hardware projects under development.
  • Legal headwinds: Apple’s massive trade-secret lawsuit is already disrupting hiring and could force court-mandated design reviews that delay releases by years.
  • Other OpenAI moves: New prompting guide and upgrades to GPT-5.6.
  • People: Key researcher Miles Wang is leaving to launch a $2 billion AI drug-discovery startup.

2) Google DeepMind chief wants a regulatory off-switch for A‑G‑I

  • Who: Demis Hassabis warns that artificial general intelligence could arrive in years.
  • Proposal: He’s quietly lobbying for a U.S.-led, industry-funded watchdog modeled on the Financial Industry Regulatory Authority (FINRA) to police frontier AGI development.
  • Mechanism: Mandatory, independent 30-day safety audits to flag cyber, biological, and deceptive threats before models are released.
  • Contrast: Dario Amodei of Anthropic favors a stricter, FAA-style agency with direct veto power.

3) Anthropic hires Y Combinator partner to tackle recursive AI bottlenecks

  • Hiring: Tom Blomfield joins Anthropic’s compute team alongside co-founder Tom Brown to address hardware limits of recursive self-improvement.
  • Education product: Rolling out Claude for Teachers, a free K–12 platform that runs background agents for grading and admin while pulling from state standards and tools like Canva Education.
  • Transparency/values: Anthropic published a framework mapping Claude’s behavioral values across multiple axes.
  • PR: Its “There’s Hope in Hard Questions” ad campaign sparked backlash and some good-natured ribbing from competitors.

4) Frontier models and product moves shake up devices and services

  • Apple: Reportedly in talks with PrismML to run models on iPhone.
  • SpaceXAI: Fixed Grok Build leaks with a server-side patch.
  • Spotify: Launched conversational AI to query your listening history.
  • Chips/rumors: Talk of an M7 Ultra chip with 1.5 TB of memory to run giant models locally.
  • Trend: Product arms races are accelerating across devices and cloud services.

That’s our roundup for today. Stay curious, keep asking hard questions, and we’ll keep listening.
This is Aurora and Isabelle signing off from AI Tech News Today.

Share your love