Apple in talks to use Google’s AI for new Siri

Good morning, AI enthusiasts. Apple is approaching a pivotal moment that could reshape Siri’s trajectory: whether to continue relying on their proprietary AI frameworks or to embrace external expertise from a competitor.

With Google’s Gemini now a strong contender and Apple already experimenting with custom models, the tech giant may soon lean on one of its fiercest rivals to revitalize Siri amid mounting AI challenges.


Today’s AI Highlights:

  • Apple considers integrating Google’s Gemini to enhance Siri
  • Meta collaborates with Midjourney to elevate AI-generated aesthetics
  • How to create an AI-powered email assistant that schedules meetings automatically
  • OpenAI and Retro Biosciences use AI to rejuvenate aging cells
  • Four innovative AI tools, community workflows, and more
APPLE & GOOGLE
Apple and Google collaboration concept

Overview: According to recent reports, Apple is in advanced talks with Google to potentially power a revamped Siri using Google’s Gemini AI, after delays pushed Siri’s major overhaul to 2026.

Key points:

  • Google is developing a tailored Gemini model to operate on Apple’s private servers, with testing already underway.
  • Apple is simultaneously advancing two Siri prototypes internally: one based on Apple’s own AI (Linwood) and another leveraging external technology (Glenwood).
  • Apple has also explored partnerships with Anthropic and OpenAI, with ChatGPT already contributing to Siri’s response capabilities.
  • The final decision on whether to adopt internal or external AI models, and which partner to choose, is expected within weeks.

Significance: While Apple’s AI development has faced criticism and setbacks, collaborating with a leading AI lab like Google could accelerate improvements for iPhone users. Given the talent migration and technical hurdles, relying solely on in-house solutions may delay progress significantly.

PARTNER SPOTLIGHT: ARTISAN

Introducing Ava: An AI-powered Business Development Representative that continuously monitors your leads and initiates outreach as soon as prospects show buying intent. Ava detects signals such as funding announcements, job openings, and website visits, then enrolls leads into tailored, multi-channel campaigns.

Features within the Artisan platform include:

  • Access to over 300 million verified B2B contacts
  • Automated enrichment of lead data from more than 10 sources
  • Comprehensive multi-channel outreach with email deliverability management
  • Real-time intent signal tracking including website visitor identification and funding news

Empower your sales team with these integrated tools to boost efficiency and conversion rates.

META & MIDJOURNEY
Meta and Midjourney partnership

Summary: Meta has announced a strategic alliance with Midjourney to incorporate the startup’s distinctive aesthetic AI technology into Meta’s future AI products, marking a shift from Meta’s traditional in-house creative AI development.

Details:

  • Meta’s Chief AI Officer, Alexandr Wang, described the collaboration as a “technical partnership” aimed at enhancing visual AI capabilities across Meta’s platforms.
  • Meta currently offers various visual generation tools such as Imagine and Movie Gen, alongside research models.
  • Midjourney’s founder, David Holz, emphasized that despite the partnership, Midjourney remains an independent, community-driven research lab without external investors.
  • In June, Midjourney launched its first video generation feature, enabling users to transform images into extendable five-second video clips.

Why it matters: Integrating Midjourney’s unique artistic style into Meta’s ecosystem could significantly elevate the quality of AI-generated visuals for billions of users. This move also signals Meta’s openness to external collaborations beyond talent acquisition to accelerate AI innovation.

BUILDING AI-POWERED EMAIL ASSISTANTS
AI email assistant scheduling meetings

Overview: This guide walks you through creating an AI email agent that you can CC on emails to automatically identify meeting times and schedule them directly in Google Calendar.

Step-by-step instructions:

  1. Register for Lindy AI (members of Rundown University receive a discount) and duplicate the provided template.
  2. Set up triggers by assigning your AI’s email address (e.g., [email protected]) and whitelist your personal email as an authorized sender.
  3. The agent supports two modes: Direct Booking (you specify a time like “Tuesday 2 PM”) and Time Finding (the agent suggests available slots based on calendar data).
  4. Configure your availability in the “Find Available Times” section by defining work hours and days.
  5. Test both modes, then simply CC the AI assistant on any scheduling email to automate meeting coordination.

Pro tip: Add your team members’ emails as authorized senders and link their calendars to enable the AI assistant to coordinate meetings across multiple schedules seamlessly.

INSIGHTS FROM GALILEO

How do you assess AI-generated content? Instead of relying on subjective impressions, learn how experts use Large Language Model (LLM) judges to evaluate AI outputs quickly and accurately.

Download Galileo’s latest eBook, featuring 70 pages of expert knowledge on:

  • Automating evaluation processes to score, explain, and flag quality issues
  • Advanced methods such as token-level scoring, Chain-of-Thought reasoning, and pairwise comparisons
  • Practical frameworks and code samples for building your own LLM-based evaluators
BREAKTHROUGHS IN AI-DRIVEN BIOLOGY
AI rejuvenating aging cells

Summary: OpenAI has partnered with Retro Biosciences to develop a specialized AI model that redesigns proteins capable of reverting cells to a stem-like state, achieving a 50-fold increase in efficiency compared to the original Nobel-winning proteins discovered in 2012.

Details:

  • The team created GPT-4b micro, an AI trained on biological datasets rather than general internet text, to optimize ‘Yamanaka’ proteins responsible for cellular reprogramming.
  • The AI-engineered proteins demonstrated dramatically enhanced DNA repair and rejuvenation capabilities.
  • Multiple independent laboratories have validated these findings, effectively reversing key cellular aging markers.

Why it’s important: While general AI models empower users across various domains, domain-specific AI trained by experts can unlock breakthroughs in biology, chemistry, and materials science-compressing years of experimental research into weeks.

QUICK UPDATES

  • 📊 Your AI Data Analyst: Connect your datasets, ask questions, and receive insights instantly-no coding needed.*
  • 🧠 Cohere launches a new enterprise-grade reasoning model.
  • 🌿 ByteDance releases a suite of open-source reasoning models with extended context capabilities.
  • ⚙️ Alibaba unveils a free agentic coding platform to streamline software development.

*Sponsored Listing

Industry News:

  • Recent court documents reveal Elon Musk requested Meta CEO Mark Zuckerberg’s support for a $97.4 billion bid to acquire OpenAI, which Meta declined.
  • xAI has deprecated its Grok 2.5 model, with plans to open-source Grok 3 within approximately six months.
  • OpenAI is expanding with a new office in New Delhi, alongside launching a $5/month ChatGPT GO subscription tailored for the Indian market.
  • Elon Musk and xAI have introduced MacroHard, an AI software company focused on simulating competitors like Microsoft using AI agents.
  • Meta FAIR researchers developed DeepConf, a deep reasoning method achieving 99.9% accuracy on the AIME benchmark using open-source models.
  • Baidu released MuseStreamer 2.0, an enhanced image-to-video model family featuring multi-character coordination, synchronized audio, and reduced pricing.

COMMUNITY SPOTLIGHT

Each edition, we highlight how readers harness AI to boost productivity, creativity, or simplify daily tasks.

Today’s feature comes from Mark H. in Detroit, MI:

“I transform imaginative ideas into tangible creations using AI. Starting with a rough sketch, I generate a detailed prompt via ChatGPT, then feed it into Krea AI to produce a realistic image. Next, I convert that image into a 3D model and finally 3D print the object. It’s incredible to bring concepts from my mind into the physical world in just a few steps-I’m excited about this future.”

How do you leverage AI? Share your story with us.

Additional Resources:

  • Explore our previous AI newsletter
  • Catch up on the latest Tech newsletter
  • Discover insights from our Robotics newsletter
  • Browse today’s curated AI tool guide
  • Reserve your spot for our upcoming workshop at 4 PM EST this Friday

Until next time,

Rowan, Joey, Zach, Shubham, and Jennifer – the team behind The Rundown

The Rundown editorial team

More from this stream

Recomended