Apple taps Gemini for Siri overhaul

PLUS: Use AI to search patents and find innovation opportunities

Good morning, AI enthusiasts. After years of delays, Apple seems to have finally picked a lane for its Siri AI overhaul — with one of its biggest rivals stepping in as a “behind-the-scenes” partner.

A reported $1B annual deal brings Google's Gemini under the voice assistant’s hood, making the anticipated Spring release a seemingly make-or-break moment for the tech giant’s already messy AI situation.

In today’s AI rundown:

  • Apple taps Google’s Gemini for Siri overhaul

  • Ex-Meta designers launch Stream Ring AI wearable

  • Use AI to find patents and innovation opportunities

  • Edison Scientific debuts Kosmos AI scientist

  • 4 new AI tools, community workflows, and more

LATEST DEVELOPMENTS

APPLE & GOOGLE

Image source: Reve / The Rundown

The Rundown: Apple reportedly finalized plans to deploy a custom 1.2T parameter version of Google's Gemini model for its long-delayed Siri overhaul, according to Bloomberg — committing roughly $1B annually to license the technology.

The details:

  • Gemini will handle summarization and multi-step planning within Siri, running on Apple's Private Cloud Compute infrastructure to keep user info private.

  • Apple also trialed models from OpenAI and Anthropic, with the 1.2T parameter count far exceeding the 150B used in the current Apple Intelligence model.

  • Bloomberg said the partnership is “unlikely to be promoted publicly”, with Apple intending for Google to be a “behind-the-scenes” tech supplier.

  • The new Siri could arrive as soon as next Spring, with Apple planning to use Gemini as a stopgap while it builds its own capable internal model.

Why it matters: After years of delays and uncertainty around Siri’s upgrade, Gemini is the model set to bring the voice assistant into the AI world (at least in some capacity). Apple views the move as temporary, but building its own solution, considering the company’s struggles and employee exodus, certainly doesn’t feel like a given.

TOGETHER WITH VANTA

The Rundown: AI is moving faster than security teams can keep up — and 59% say AI risks outpace their expertise. Vanta's new State of Trust report surveyed 3,500 business and IT leaders across the globe to reveal how organizations are navigating this growing gap.

The data reveals:

  • 61% of teams spend more time proving security than improving it

  • AI-driven attacks are growing bigger, faster, and more sophisticated

  • Nearly half of leaders say AI gives them time for strategic security work

Download the State of Trust report to see what early adopters are doing to stay ahead.

SANDBAR

Image source: Sandbar

The Rundown: Sandbar, a startup founded by former Meta designers, launched Stream Ring — an AI wearable that captures whispered thoughts through a ring device and transcribes voice into organized notes while also doubling as a music controller.

The details:

  • Cofounders Mina Fahmi and Kirak Hong developed the ring after working on neural interfaces at CTRL-Labs, which was acquired by Meta in 2019.

  • Users activate recording by holding a touchpad rather than shouting wake words, with whisper-detection microphones converting speech to text.

  • The AI assistant responds in a synthesized version of the wearer's voice using ElevenLabs speech technology, enabling back-and-forth conversation.

  • The Stream Ring is available for preorder starting at $249 (plus a $10 subscription model), with a planned summer 2026 delivery.

Why it matters: Another wearable has entered the arena, with the Stream Ring continuing the infusion of AI voice tech across form factors, joining pendants, pins, and more. Simplicity may be a differentiator, but there is no shortage of competition from both other wearables and hardware like earbuds getting AI upgrades of their own.

AI TRAINING

The Rundown: In this tutorial, you will learn how to use Perplexity's AI-powered search to quickly find patents, analyze innovation gaps, and position your invention without infringement risk.

Step-by-step:

  1. Go to Perplexity and search naturally: "Are there any patents related to AI automations?" - Perplexity automatically activates Patent Research (beta), showing relevant filings, owners, and dates

  2. Refine with conversational queries: "Find active patents for AI-driven industrial automation and model drift detection", then follow up with "Summarize main claims" or "Show whitespace in this field"

  3. Toggle on Agent Mode for advanced analysis - the AI automatically retrieves patents from multiple jurisdictions, creates tables, and builds visualization charts (showing "12 steps completed")

  4. Review generated PNG charts showing patent clusters and risk zones, plus CSV files with patent IDs, titles, owners, and claims - identify which companies dominate and where opportunities exist

  5. Use results to inform product design by identifying saturated areas to avoid, high-opportunity/low-risk zones for innovation, and specific technologies or claims requiring caution

Pro tip: Start with a broad query to capture the full patent landscape. Then iterate: ask the agent to list patents by company, summarize claims, or visualize whitespace.

PRESENTED BY SYNK

The Rundown: AI coding tools are boosting productivity, but they're also introducing new security concerns. Join Snyk's live session on Nov. 20 at 11 AM ET to learn how to embed security in GenAI-powered development workflows before vulnerabilities make it to production.

Snyk Staff Developer Advocate Sonya Moisset will cover:

  • How to spot security vulnerabilities hidden in AI-generated code

  • Secure coding best practices tailored for GenAI workflows

  • Strategies for building AI-native apps with security from the start

Register now. Plus, ISC2 members will earn 1 CPE credit for attending live!

AI RESEARCH

Image source: Reve / The Rundown

The Rundown: Futurehouse just announced the launch of its commercial spinout Edison Scientific, alongside the debut of Kosmos — an autonomous AI research system that beta testers report can complete six months of scientific work in a single day.

The details:

  • Kosmos coordinates cycles of literature review, data analysis, and hypothesis generation, processing 1,500 papers and executing 42k lines of code per run.

  • All of Kosmos’ generations maintain full citation traceability for every claim, making it easily auditable down to specific lines of code.

  • 79% of Kosmos’ outputs were validated as accurate, with the AI reproducing unpublished findings and making new discoveries across multiple fields.

  • Edison Scientific will commercialize the platform following pharma demand, while FutureHouse continues nonprofit foundational research development.

Why it matters: Edison Scientific says the “era of AI-accelerated science is here,” with Kosmos continuing the trend of AI models removing the human-bandwidth limitation for research and analysis. These timeline-compressing abilities are set to completely transform the pace of progress across scientific domains.

QUICK HITS

  • 🧪 Kosmos - Edison Scientific’s next-generation AI scientist

  • 🗺️ Codemaps - Windsurf’s coding tool to understand & navigate codebases

  • 🎥 Sora App - OpenAI’s AI video platform, now available for Android users

  • 🧭 Google Maps - New Gemini integration for conversational assistance

OpenAI said it now has 1M+ business customers, becoming the fastest-growing platform in history, with ChatGPT for Work growing 40% in two months to 7M+ seats.

Stability AI won a UK High Court case against Getty Images over trademark infringement related to AI training, with Getty saying the ruling shows that even well-resourced companies “face significant challenges in protecting their creative works.”

xAI reportedly required employees to submit biometric data to train its "Ani" and other AI companions, telling staff the collection was a mandatory job requirement.

Google integrated its Gemini AI into Maps, enabling conversational navigation, multi-step questions, and directions based on visible buildings instead of just distances.

Famed ‘Big Short’ investor Michael Burry disclosed over $1B in put options against Nvidia and Palantir, following cryptic social media warnings about an AI bubble.

Snap is partnering with Perplexity to integrate its AI into Snapchat starting in 2026, with Perplexity paying $400M to reach the platform’s nearly 1B monthly users.

COMMUNITY

Every newsletter, we showcase how a reader is using AI to work smarter, save time, or make life easier.

Today’s workflow comes from reader Elaine in Toronto, Canada:

"I’ve found a simple but powerful way to learn from The Rundown AI. Every day, you feature a user story — and instead of just reading it, I copy their experience straight into ChatGPT and ask the model to teach me how to do the same, step by step. It feels like recreating a mini experiment every morning. Over time, this has become my favorite way to learn: turning other people’s discoveries into my own hands-on lessons."

How do you use AI? Tell us here.

That's it for today!

Before you go we’d love to know what you thought of today's newsletter to help us improve The Rundown experience for you.

Login or Subscribe to participate in polls.

See you soon,

Rowan, Joey, Zach, Shubham, and Jennifer — the humans behind The Rundown

Reply

or to participate.