Welcome to Get The Gist, where every weekday we share an easy-to-read summary of the latest and greatest developments in AI—news, innovations, and trends—all delivered in under 3 minutes! ⏱

In today’s edition:

• Google Expands Gemini Live Video Features

• Meta Got Caught Gaming AI Benchmarks

• Amazon Launches New ‘Nova Sonic’ AI Voice Model

• Google Rolls Out Enhanced AI Mode Search Capabilities

• And more AI news….

Top Developments

Google

Google Expands Gemini Live Video Features

The Gist: Google is rolling out Gemini Live’s real-time camera and screensharing features to Pixel 9 and Galaxy S25 users, with more Android devices getting access soon through Gemini Advanced.

Key Details:

• Users can now ask Gemini Live questions about things seen through the camera.

• Screensharing lets Gemini analyze websites and give feedback or suggestions.

• The update is free for Pixel 9 and Galaxy S25 users, but others need a Gemini Advanced subscription.

• The feature was first demoed at Google I/O as part of Project Astra.

Meta

Meta Got Caught Gaming AI Benchmarks

The Gist: Meta’s new Llama 4 model appeared to outperform rivals on a popular AI benchmark—but it turns out the version tested wasn’t the same one released to the public.

Key Details:

• Meta submitted a chat-optimized version of its Maverick model to the LMArena leaderboard.

• Researchers found this version was customized for better conversational performance.

• Meta says this kind of testing is part of normal experimentation.

• The incident raised concerns about fairness and transparency in AI model comparisons.

Amazon

Amazon Launches New ‘Nova Sonic’ AI Voice Model

The Gist: Amazon introduced Nova Sonic, a new AI voice model that’s fast, accurate, and designed to power next-gen assistants like Alexa+, while competing with offerings from OpenAI and Google.

Key Details:

• Nova Sonic supports natural, real-time voice interaction with low latency.

• It outperforms rivals on benchmarks for speech recognition and multi-speaker conversations.

• The model is available through Amazon’s Bedrock platform for developers.

• It’s part of Amazon’s broader push toward artificial general intelligence.

Quick Gist

OpenAI's former chief research officer Bob McGrew has joined Mira Murati's AI startup, Thinking Machines Lab, which is aiming to raise $1 billion and is supported by numerous ex-OpenAI team members. (Read More)

OpenAI is exploring a $500 million acquisition of io Products, an AI hardware startup co-founded by Sam Altman, which is developing less disruptive devices like a screenless phone. (Read More)

Shopify CEO Tobi Lütke stated in a company-wide memo that all teams must demonstrate effective AI usage to justify new hires or resources, making AI fluency a core expectation. (Read More)

Google's DeepMind is using strict noncompete clauses in the U.K. to prevent AI talent from joining competitors, sparking frustration among affected researchers. (Read More)

Google is rolling out enhanced AI Mode search capabilities that allow users to upload images for more contextual, multimodal results powered by its Gemini model. (Read More)

Amazon has updated its AI video generation model, Nova Reel, to version 1.1, enabling longer, multi-shot clips up to two minutes and offering users greater control over output. (Read More)

THAT’S IT FOR TODAY!