Siri gets Gemini’s brain

Plus: The AI agent takeover in Slack

Welcome back! Picture this: You're a contractor for OpenAI, and they just asked you to raid your old work folders. That PowerPoint deck you made for a work itinerary? The Excel model from your last finance gig? Send it over—well, strip out the confidential stuff first. 

This is how OpenAI is preparing its AI agents for the real world. Not with sanitized test cases or academic benchmarks, but with the actual messy artifacts of knowledge work. We're talking about a fundamental rewiring of how companies operate. Are you ready for 2026? Because we’re just getting started.

Apple Hands Siri's Brain to Google

Via The Information

After years of watching from the sidelines, Apple is making a big AI move: a multiyear deal with Google to power the next generation of Siri using Gemini models. 

What's happening: 

  • Apple 🤝 Google: Apple will lean on Google's Gemini and cloud infrastructure for its foundation models, though processing will still happen on-device and through Apple's private cloud compute. 

  • Siri’s struggles: The deal comes after Apple delayed its Siri AI upgrade from 2025 to 2026, struggling to deliver on features it had already started advertising. 

  • The lead up: Google was reportedly in talks for this since August, with Bloomberg suggesting Apple could pay around $1 billion annually for the partnership.

The OpenAI complication: Apple already integrates ChatGPT into Siri for complex queries that need broader world knowledge. Apple says it's not changing the OpenAI agreement, but the message is clear: When it came time to build the core intelligence layer, Apple chose Google.

Validation for Google: After lagging OpenAI in mindshare for two years, Google is now winning the trust (and checkbooks) of the world's most cautious tech company. The partnership signals Google, not OpenAI, is the partner Apple trusts the most.

The bigger picture: Apple's legendary control-freak instincts are giving way to pragmatism. In the AI era, even the most self-reliant companies are realizing they need partners to compete and can't build cutting-edge AI alone.

Microsoft Promises Your Power Bill Won't Pay for AI

Microsoft made a public pledge that sounds almost too good to be true: As it builds out massive new data centers for AI, it won't let your electricity bill go up. The announcement comes amid growing backlash against data center construction.

What Microsoft is promising: The company says it will "pay its own way" by working with utility companies and state commissions to ensure the cost of powering its data centers isn't passed on to residential customers. It's also pledging to create local jobs and minimize water usage. 

Community pushback: 142 activist groups across 24 states are organizing against data center projects. Microsoft has already abandoned plans for a Wisconsin facility in October after overwhelmingly negative community feedback. Meanwhile, Meta announced its own AI infrastructure program just one day before Microsoft's pledge, signaling that major AI companies are coming to the same realization—that the social license to build AI infrastructure isn't guaranteed.

The bigger picture: AI's infrastructure demands are colliding with local reality. These facilities require enormous amounts of electricity and water, often in communities that didn't ask for them and won't directly benefit from the AI services they power.

Slackbot Just Became an AI Agent

Salesforce has turned Slackbot into a full AI agent, powered by generative AI. The advanced bot is available now to Business+ and Enterprise+ Slack subscribers, and the company's CTO Parker Harris is betting on it becoming as viral as ChatGPT (at least inside enterprises).

What the new Slackbot does:

  • Searches across company tools and data (when given permission)

  • Drafts emails and schedules meetings directly within Slack

  • Pulls info from other apps like Google Drive or Microsoft Teams

The next version? Salesforce envisions Slackbot going beyond text-based interactions to having voice capabilities and being able to browse the internet alongside users. Harris says this is just the beginning of Slackbot's second bloom as an AI agent, and internally, it's been the most adopted tool Salesforce has ever released during employee testing.

The bigger picture: Agentic AI is quickly moving from hype to actual workplace adoption. The advantage here is distribution: Slackbot sits inside a platform millions of workers already use daily, which means AI agents can reach employees through familiar interfaces rather than requiring them to learn new tools.

100+ ChatGPT Prompts to Revolutionize Your Workflow [Free Guide]

Supercharge your productivity with HubSpot's comprehensive "How to Use ChatGPT at Work" guide. This free resource is your fast track to AI mastery:

  • AI Crash Course: Understand ChatGPT's capabilities in under 30 minutes

  • Industry-Specific Use Cases: 15+ real-world applications across various sectors

  • Productivity Guide: 21 best practices to 10x your efficiency with AI

  • Prompt Powerhouse: 100+ ready-to-use prompts for immediate implementation

  • Challenge Buster: Overcome common AI hurdles with expert strategies

Plus, in-depth sections on email composition, content creation, customer support, and data analysis.

Join 10,000+ professionals already leveraging this guide to revolutionize their workflows. From beginners to advanced users, this resource caters to all levels.

AI code reviews

Via Diffray

Diffray uses a coordinated system of specialized AI agents to review GitHub pull requests with full codebase context. 

How you can use it

  • Catch bugs and architectural issues before code hits production

  • Enforce team standards as configurable “culture as code” rules

  • Reduce noisy PR feedback and speed up review cycles

  • Run privacy-safe AI reviews directly via a GitHub app

Pricing: Free trial available

AI-generated email infrastructure for SaaS teams

via Sequenzy

Sequenzy lets you spin up fully wired, brand-matched email sequences from a single prompt. It generates content, then connects everything directly to your product and revenue stack.

How you can use it

  • Launch onboarding, billing, and lifecycle emails in minutes

  • Generate conversion-optimized subject lines and copy automatically

  • Recover revenue with Stripe-integrated dunning flows

  • Plug into developer workflows using APIs instead of manual setup

Pricing: Free and paid plans

Jobs, announcements, and big ideas

  • Lilly and NVIDIA team up on a $1 billion AI lab to accelerate drug discovery.

  • The Senate passes the Defiance Act, giving deepfake victims the right to sue over nonconsensual content on platforms like X.

  • Google rolls out Veo 3.1 with better consistency, creative control, and higher-quality AI video generation.

  • Manus launches an AI Meeting Minutes tool with summaries, speaker recognition, and workflow integrations.

  • Anthropic’s OpenCode misstep backfires, inadvertently strengthening rival OpenAI.

A wave of AI drops: I break down ChatGPT’s move into health and Google’s deeper Gemini integrations, plus new AR wearables and a flood of fresh models.

That’s a wrap! And I hope you’ll join me tomorrow for our 2026 AI predictions event. Last chance to register!

—Matt (FutureTools.io)