Issue #1 | DailyPrompt

01 announcement

Anthropic Unveils Claude Mythos and Project Glasswing to Harden Critical Software

Anthropic previewed Claude Mythos — its most powerful model ever, a new tier above Opus — exclusively to more than 50 cybersecurity organisations through Project Glasswing. Mythos has already autonomously discovered thousands of zero-day vulnerabilities across every major OS and browser, including a 17-year-old remote code execution flaw in FreeBSD. Founding partners including AWS, Apple, Google, and Microsoft are receiving over $100M in combined usage credits to help secure the world's most critical software.

TechCrunch →

product

Claude Code Gets Automated Routines and a Desktop Makeover

A major Claude Code update landed on April 14, introducing "routines" — saved configurations pairing a prompt, one or more repositories, and connectors that can run automatically on Anthropic's cloud infrastructure. The release also ships a refreshed desktop interface, stronger team onboarding guides, improved MCP handling, and reliability fixes across auth, settings, and editor workflows.

SiliconAngle

Anthropic

@AnthropicAI

New Anthropic Fellows research: developing an Automated Alignment Researcher. We ran an experiment to learn whether Claude Opus 4.6 could accelerate research on a key alignment problem: using a weak AI model to supervise the training of a stronger one.

Apr 14 View on X →

community

$ cat story_03.md

Power Users Revolt Over Undisclosed Drop in Claude's Performance

Heavy Claude users are pushing back after the AI began consistently failing to follow complex instructions and producing lower-quality outputs on demanding workflows. The complaints appear tied to Anthropic quietly reducing Claude's default "effort" level to economise on token usage — a change made without public notice. Users have accused Anthropic of prioritising compute cost savings over the quality that originally won them over.

$ open fortune →

announcement No. 04

CoreWeave Shares Jump 11% on Multi-Year Deal to Power Claude

CoreWeave announced a multi-year infrastructure agreement with Anthropic to power the Claude AI model family, sending its stock up 11% on the news. The deal highlights the enormous and ongoing compute buildout required to keep Claude competitive at scale as demand for AI inference continues to grow.

Source: CNBC →

Claude

@claudeai

We've redesigned Claude Code on desktop. You can now run multiple Claude sessions side by side from one window, with a new sidebar to manage them all.

Apr 14 View on X →

community

Claude Was the Only AI Everyone at HumanX Wanted to Talk About

At the HumanX AI conference in San Francisco — attended by thousands of practitioners focused on agentic AI — Claude consistently emerged as the most-mentioned product in hallway conversations, panels, and demos. TechCrunch reported that Anthropic stood out as the clear star of the event, with enterprise attendees repeatedly naming Claude as their tool of choice for complex, multi-step workflows.

TechCrunch →

product

Anthropic Bars Claude Subscriptions From Covering Third-Party Tools

From April 4, Claude Code subscribers can no longer use their subscription token allowance with third-party tools such as OpenClaw, and must instead pay separately at pay-as-you-go rates. Anthropic confirmed the policy extends to all third-party harnesses and will roll out more broadly, drawing criticism from developers who had built cost-effective workflows around the subscription's economics.

Read on TechCrunch →

Breaking

Claude Outage Hits More Than 7,000 Users on April 15

Claude AI went down for thousands of users on the morning of April 15, with Downdetector tracking over 7,000 reports at peak. The outage compounds a difficult stretch for Anthropic, which has recently faced user backlash over performance regressions and contentious changes to its subscription and third-party access policies.

GV Wire

Alex Albert

@alexalbert__

New version of Claude Code in the desktop app dropped today with tons of new features and performance improvements. Between Cowork and Code, I've found I don't really need to open other apps (or even my terminal) to do most of my work now.

Apr 14 View on X →

08 announcement

Vas Narasimhan Joins Anthropic's Long-Term Benefit Trust

Anthropic announced on April 14 that Vas Narasimhan, CEO of Novartis, has been appointed to its Long-Term Benefit Trust. The Trust is a key governance structure designed to hold Anthropic accountable to its public benefit mission as the company scales, and Narasimhan's appointment expands the oversight body with an outside perspective from the life sciences industry.

Anthropic →