While the AI industry chases reinforcement learning, Essential AI made the opposite bet. Their new 8B model embodies a thesis about where machine intelligence originates. The transformer's co-inventor is calling the shots on research.
Tim Cook built Apple's leadership into a monument of stability. In 2025, that monument cracked. Meta poached AI and design chiefs with $25M packages. The chip architect may follow. What broke inside the world's most valuable company?
OpenRouter's 100 trillion token study was supposed to prove AI is transforming everything. The data shows something else: half of open-source usage is roleplay, enterprise adoption is thin, and one account caused a 20-point spike in the metrics.
Claude just got a major upgrade that transforms how it finds and uses information. The AI assistant now searches the web and connects with Google Workspace, making it a more capable research partner.
The new Research feature lets Claude search the web and work documents to answer complex questions. Unlike basic search tools, Claude builds its searches step by step, digging deeper into topics that need more investigation. It backs up its findings with citations, so users can check its work.
The Google Workspace integration connects Claude to Gmail, Calendar, and Google Docs. This means it can pull information from your emails, calendar events, and documents without needing manual uploads. The system maintains security while giving Claude access to your work context.
These tools shine in real-world scenarios. Marketing teams can use Claude to gather competitive research while pulling internal strategy documents. Sales teams can prep for client meetings by having Claude scan past emails and meeting notes while checking the latest company news. Engineers can analyze system specs and documentation to plan technical solutions.
For students, Claude can analyze course materials and search academic resources to create study plans. Parents can use it to track family commitments and check local events that might affect schedules.
Enterprise users get additional features. Administrators can turn on document cataloging, which helps Claude find information buried in long documents or scattered across files. The system uses secure retrieval techniques to search company documents while keeping data private.
The research capability stands out for its speed and depth. Claude can tackle multiple research tasks throughout the workday, delivering thorough answers in minutes. It doesn't just search - it thinks through problems and explores different angles automatically.
Consider a product launch scenario: Claude can check competitor moves, analyze market trends, and pull relevant internal documents about product specs and positioning. Instead of spending hours gathering this information manually, teams get comprehensive insights they can act on quickly.
For meeting prep, Claude saves time by combining calendar data, email threads, and relevant documents. It can identify action items from past meetings and gather context for upcoming discussions. This lets people focus on strategy instead of hunting through old emails.
The system works similarly for technical planning. Engineers can have Claude analyze documentation, check implementation patterns, and review security standards. This speeds up solution design while ensuring nothing gets missed.
These features mark a shift in how AI assistants handle information. Instead of relying on users to provide context, Claude actively gathers what it needs. The step-by-step research approach means it can tackle complex questions that need multiple layers of investigation.
The rollout starts with Max, Team, and Enterprise plans in the US, Japan, and Brazil. Web search is now live in these regions for paid plans. The Google Workspace integration is in beta for all paid users, though admins need to enable it first.
Anthropic, Claude's creator, plans to expand these features. They'll add more content sources and enhance Claude's research abilities in coming weeks. The current release focuses on core business needs, but the potential applications span many fields.
This update reflects a broader trend of AI systems becoming more capable researchers and collaborators. By connecting to both internal and external information sources, they can provide more context-aware and useful assistance.
Why this matters:
Read on, my dear:
Anthropic hired IPO lawyers the same day it announced its first acquisition. The company claims efficiency while burning $2.8B annually. Its safety positioning has won enterprise customers—and alienated Trump's White House. The math is complicated.
Inception Point AI produces 3,000 podcast episodes per week with eight employees, spending roughly $1 per episode and breaking even at 20 listens. The Venice startup doesn't compete on quality. It competes on coverage, treating audio as infrastructure for programmatic ads.
Voize raised $50M for nursing documentation AI. Abridge raised $300M at $5.3B valuation. The 10× gap reveals what healthcare really values—and what happens when efficiency gains hit an industry that already cuts corners on staffing.
Thinking Machines seeks a $50 billion valuation four months after raising $2 billion. The OpenAI spin-out has one API in private beta. Investors aren't pricing the product—they're pricing the fear of missing out on Mira Murati's next move.
Get the 5-minute Silicon Valley AI briefing, every weekday morning — free.
