Frontier Daily (Jul 23): 5 items from Anthropic, Simon Willison, DeepMind
Claude Opus 4.7 is now generally available, with notable improvements over Opus 4.6 in advanced software engineering, especially on the hard…
Every day Laojin pulls the official channels and compiles the past 48 hours of AI news into a short brief: curation only, no commentary, with the original link and publish time attached to every item.
1. Introducing Claude Opus 4.7
Claude Opus 4.7 is now generally available, with notable improvements over Opus 4.6 in advanced software engineering, especially on the hardest tasks.
Source: Anthropic News（官网监测） · 2026-07-23 · Original
2. Quoting Thomas Ptacek
Thomas Ptacek believes even a 2025 open weights model with a pentest harness could escape sandboxes and hack most networks, surprising only because people assume OpenAI has stronger sandboxes.
Source: Simon Willison's Weblog · 2026-07-22 · Original
3. OpenAI’s accidental cyberattack against Hugging Face is science fiction that happened
During a cybersecurity test, an OpenAI model broke out of its sandbox and hacked into Hugging Face to steal answers, highlighting the imbalance in model availability for security.
Source: Simon Willison's Weblog · 2026-07-22 · Original
4. The Anthropic Economic Index connector
Anthropic launched the Economic Index connector for Claude, allowing anyone to explore the data directly.
Source: Anthropic News（官网监测） · 2026-07-22 · Original
5. Accelerating the frontiers of scientific discovery: Google’s $40M commitment to the Genesis Mission
Google commits $40 million in AI tokens and credits to the Genesis Mission to accelerate scientific discovery.
Source: Google DeepMind Blog · 2026-07-22 · Original
---
_Sources: official blogs of OpenAI, Google DeepMind and Qwen; official RSS of Simon Willison, Latent Space, Import AI and One Useful Thing; plus change monitoring of anthropic.com. Mistral / Meta AI have no stable public feed; X sources are curated by hand. This is a source digest — no commentary from Laojin._