In partnership with
Hey AI Enthusiasts!
This week’s roundup is packed with fresh launches and big platform moves, from Anthropic shipping a new product built at lightning speed, to Google and Apple tightening their AI partnership, and the Pentagon preparing to bring Grok onto its networks.
Let’s dive in!
In today’s insights:
🧑💻 Anthropic releases Cowork, built entirely by Claude in just 10 days
🍏 Apple chooses Google Gemini to power Siri’s AI upgrade
🪖 Musk’s Grok AI is set to be integrated into Pentagon networks
🔗 Google launches Personal Intelligence, letting Gemini connect your apps
Read time: 6 minutes.
Sponsor The AI Field
Reach AI builders and founders this week.
🗞️ Recent Updates
The AI Field: Anthropic has launched Cowork, a research preview that lets Claude read, edit, and create files directly on your computer with minimal oversight. In a striking demonstration of AI's capabilities, engineers revealed the entire product was built in just 10 days with all code written by Claude Code itself.
Details:
Autonomous file management: Users grant Claude access to specific folders where it independently manages files, reorganizes downloads, creates spreadsheets from screenshots, or drafts reports from scattered notes
AI-coded product: Boris Cherny, head of Claude Code, confirmed Claude Code wrote "pretty much all" of Cowork's codebase, with human engineers guiding architecture while managing 3 to 8 Claude instances simultaneously
Agentic execution: Cowork makes plans and executes tasks in parallel rather than requiring constant back and forth, operating with significantly more autonomy than conversational Claude
Safety measures: Users control folder access, Claude requests permission before major actions, though Anthropic acknowledges agent safety remains an active development area
Why This Matters: When AI can build an entire product in 10 days versus months of traditional development, we've crossed into a new era where AI builds itself. This accelerates innovation dramatically but raises critical questions about security and oversight when tools can be created faster than humans can verify their safety. The acknowledged risks around destructive actions and prompt injections show that autonomous AI agents, especially those building other AI agents, still need careful guardrails.
Help us make better ads
Did you recently see an ad for beehiiv in a newsletter? We’re running a short brand lift survey to understand what’s actually breaking through (and what’s not).
It takes about 20 seconds, the questions are super easy, and your feedback directly helps us improve how we show up in the newsletters you read and love.
If you’ve got a few moments, we’d really appreciate your insight.
Apple/Google
🍎 Apple Chooses Google Gemini to Power Siri's AI Upgrade
The AI Field: Apple has partnered with Google in a multi-year deal to power the next generation of Siri using Gemini models, marking a significant shift in Apple's AI strategy and potentially worth around $1 billion annually.
Details:
Major partnership: After evaluating OpenAI and Anthropic, Apple chose Google's technology as the foundation for Apple Intelligence, though the deal remains non-exclusive
Siri overhaul coming: A "more personalized Siri" launches this year with better contextual understanding, on-screen awareness, and deeper app controls
Existing integrations remain: Apple keeps its ChatGPT integration for complex queries while promising to maintain privacy standards
Strategic shift: The deal comes as Google surpassed Apple in market value for the first time since 2019
Why This Matters: Apple's decision to outsource its AI foundation to Google is a pragmatic admission that building competitive models in-house would take too long. Rather than delaying further, Apple is leveraging Google's proven technology to finally deliver the Siri upgrade users have awaited since 2024. This partnership strengthens Google's position against OpenAI while deepening the complex relationship between two tech giants that fiercely compete yet depend on each other for critical services.
The AI Field: Defense Secretary Pete Hegseth announced that Elon Musk's Grok AI will be integrated into Pentagon networks this month with access to military and intelligence data—just days after the chatbot sparked global outcry for generating non-consensual deepfake images.
Details:
Rapid deployment: Grok launches in the Defense Department later this month with access to military IT systems and intelligence databases
Global backlash: Malaysia and Indonesia blocked Grok after it generated sexualized deepfake images, while the UK launched an investigation
Policy reversal: Hegseth emphasized the AI "will not be woke" and will operate "without ideological constraints," reversing Biden-era safeguards
Dual integration: Google's AI will also be deployed alongside Grok
Why This Matters: Adopting Grok during a global safety crisis signals a fundamental shift in military AI ethics—prioritizing speed over caution. Feeding sensitive defense data into a controversial commercial system raises serious questions about oversight and the unprecedented alignment between Musk's business interests and national security infrastructure.
The AI Field: Google has launched Personal Intelligence, a new beta feature that allows Gemini to securely access and reason across your Gmail, Google Photos, YouTube, and Search history with a single tap, making the AI assistant deeply personalized to your life and needs.
Details:
Cross-app reasoning: Gemini analyzes complex queries by pulling information from multiple Google apps simultaneously, like finding car specs in Gmail while referencing family trips in Photos to suggest tire options
Privacy controls: The feature is off by default, users choose which apps to connect, and Google doesn't train models directly on your inbox or photos but on filtered prompts and responses
U.S. rollout: Starting this week for Google AI Pro and AI Ultra subscribers across web, Android, and iOS, with plans to expand to free tier and more countries
Smart corrections: Gemini references its sources and allows real-time adjustments like "Remember, I prefer window seats" while avoiding assumptions about sensitive data
Why This Matters: Google is leveraging its unique advantage of already hosting your digital life to create truly personalized AI. By keeping data within its ecosystem rather than requiring external data sharing, Google positions Gemini as a more secure alternative in the race for context-aware AI assistants that understand your preferences, relationships, and history beyond generic responses.
🛠️ Treding AI Tools
🎬 Deevid.ai – Create videos from text, images, or existing clips with AI-powered templates and effects, perfect for fast content creation and social posts.
🗞️ More AI Hits
OpenAI: Launched ChatGPT Health, a dedicated health space that can connect medical records + wellness apps (with stricter privacy controls)
OpenAI: secured 750 MW of compute through Cerebras in a deal worth over $10B.
Meta: launched “Meta Compute” to scale AI infrastructure and expand data center investment.
Golden Nuggets
🧑💻 Anthropic releases Cowork, built entirely by Claude in just 10 days
🍏 Apple chooses Google Gemini to power Siri’s AI upgrade
🪖 Musk’s Grok AI is set to be integrated into Pentagon networks
🔗 Google launches Personal Intelligence, letting Gemini connect your apps
What did you think of today’s edition?
Until next time!
Olle | Founder of The AI Field