📝 05/05/25: Anthropic's API Web Search, Hack Your Inbox, and the Ethics of AI Control
📌 This Week's News Highlight: Anthropic’s New Web Search API
Anthropic just launched Web Search on their API, allowing developers to build Claude-powered apps that pull real-time info from across the web.
This means Claude's API can now:
• Fetch live data like market prices, news updates or regulatory changes for your product in real-time
• Scan industry blogs, product reviews or technical docs to add fresh context
• Run multi-step searches to refine answers and dig deeper into a topic
• Provide source links in all generated responses
Potential use cases across industries:
• Financial services: real-time stock analysis and market insights for fintech startups as a value add for their customers
• Legal research: up-to-date case law and regulatory changes for legaltech businesses to stay up to date with legal changes in realtime
• Augmenting developer tools: access to the latest API documentation, code samples and technical articles for faster troubleshooting and feature updates
• Advanced business intelligence: Real-time competitive analysis, industry reports and trend research for founders tracking market shifts and customer insights
Note that Web Search has also been added to Claude Code (Claude's AI coding assistant), but Claude Code is still in beta.
🔗 Read more here
🕵️ Founder, Anonymous
A founder feeling overwhelmed by a cluttered inbox decides to test Superhuman for faster, stress-free email management.
How it worked:
• Streamlined setup: Connected their Gmail account and cleaned up their inbox view in minutes using Superhuman’s setup wizard
• Faster triage: Used keyboard shortcuts like “E” to archive emails and Command "K" for quick reminders, keeping their inbox under control
• AI-powered replies: Pressed Command "J" to turn bullet points into polished, context-aware emails, saving time without sacrificing tone
Result:
A cleaner inbox, faster response times, and a lot less email anxiety!
🔗 Have a play here
💡 Thought of the Week
Who really controls the future of AI?
OpenAI recently confirmed that its non-profit arm will stay in control, even as the commercial side scales. At first glance, this sounds like a reassuring commitment to keep ethics at the centre.
But it raises a few interesting questions:
• Checks and balances or PR play? The non-profit structure is meant to keep the mission on track, but can it really hold back profit-driven decisions as pressure to monetise grows? And is this just an attempt to measure up to Anthropic's safety-first position in the market?
• Control vs agility: Staying mission-aligned might reduce reckless, profit-first decisions, but will it also slow innovation in a fast-moving space (especially in light of faster moving, cheaper to run alternatives like we've seen with DeepSeek)?
• Power concentration: Even with a non-profit layer, control still sits with a small, elite group. Does this really reduce the risks of runaway AI, or just centralise them?
🔗 Sneaky Links
- Apple is reportedly testing AI-powered search in Safari, a potential move to close the gap with Google and Microsoft in the browser wars
- Figma just dropped a suite of AI tools for faster site builds, app prototypes and marketing assets
- OpenAI launched a free ‘ChatGPT at Work’ training hub to help teams get more from their AI tools
- ChatGPT’s shopping search just got a smarter upgrade, promising more accurate, context-aware product recommendations