Thursday, 23 July 2026 Sign in

Safety Research

15 summarised stories about Safety Research, each linking back to the original source. Browse all topics →

Tuesday, 20 February 2018

Preparing for malicious uses of AI

OpenAI Blog 8 years ago

Researchers from multiple institutions including the Future of Humanity Institute and Centre for the Study of Existential Risk have published a paper analyzing how malicious actors could misuse AI and proposing prevention strategies. The work represents nearly a year of collaborative research across at least five organizations. The paper aims to inform policymakers and technologists about threat scenarios and mitigation approaches to reduce risks from intentional AI misuse.

Read at OpenAI Blog

Safety Research

The daily briefing

Every AI story that matters,
in your inbox by 8am.

TLDRocket reads 60+ sources, removes duplicate coverage, and summarises the day in two minutes. Free, no spam, unsubscribe anytime.

TLDRocket 23 July 2026

The AI industry's infrastructure binge is hitting a hard limit. Google burned cash for the first time in its history, spending more on AI infrastructure than it generates quarterly—a stark inversion of the company that built a search-driven money machine. Sequoia's $300 million bet on Etched, valuing the chip startup at $10.3 billion, reflects the same calculus: if you can't afford the compute, build the hardware. Etched's inference-optimized silicon and competitors like Cursor's model router are solving a real problem—frontier models like Claude Opus cost too much for routine work. Cursor users report saving 30–50% by routing simple tasks to cheaper models, while Runway and Ramp built similar routers for their respective domains. The proliferation suggests the all-frontier-all-the-time era is ending.

Read the full briefing →

Safety Research

Tuesday, 20 February 2018

Preparing for malicious uses of AI

Every AI story that matters, in your inbox by 8am.

Every AI story that matters,
in your inbox by 8am.