Frontier Daily (Jul 20): 5 items from Simon Willison, Anthropic, Import AI and more

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. Reverse-engineering is cheap now

Simon Willison notes that coding agents have made reverse-engineering home devices much cheaper, reducing the psychological burden of maintenance.

Source: Simon Willison's Weblog · 2026-07-20 · Original

2. Apply for Anthropic’s AI for Science rare disease research grants

Anthropic is offering up to $50,000 in Claude credits for research on rare genetic diseases using AI.

Source: Anthropic News（官网监测） · 2026-07-20 · Original

3. Who’s Afraid of Chinese Models?

Ben Thompson proposes US law to legalize training data as fair use and ban anti-distillation terms, suggesting Alibaba's Qwen 3.8 Max release may be influenced by Xi Jinping's pro-open-source speech.

Source: Simon Willison's Weblog · 2026-07-20 · Original

4. Import AI 465: Open vs closed gaps; Kimi K3; Demis' big policy plan

The UK's AISI reports that the cybersecurity capability gap between open-weight and closed models is narrowing, with some open models performing close to frontier models from months earlier.

Source: Import AI · 2026-07-20 · Original

5. Safety and alignment in an era of long-horizon models

OpenAI shares lessons from deploying long-running AI models, highlighting new safety risks, observed failures, and improved safeguards through iterative deployment.

Source: OpenAI Blog · 2026-07-20 · 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._