we have a year to fix security everywhere
jyn.dev · 2,738 words · saved by 1 readers
consumer-grade hardware can run an LLM that hacks the planet. we can stop it, but we don't have much time.
GLM 5.3-flash released last week, and that means Project Glasswing and Daybreak are running out of time. Cheap models capable of dangerous hacking are now available to anyone, without the normal safeguards for refusing malicious actions. We need to fix vulnerabilities across the industry so that we aren't caught unawares. And for one of the first times in computing history, we have the ability to! We can use frontier LLMs that move faster than a human to find and fix these issues in the time we have left. The hard remaining part is deploying the fixes. This probably sounds like nonsense…
saved by
related reading
- Project Glasswing: Securing critical software for the AI era \ Anthropicanthropic.com
- The End-State Fallacy: Where Is AI Security Headed?endstatefallacy.com
- Guardian Angels: LLM Personalization for Productivity and Security · Gwern.netgwern.net
- Assessing Claude Mythos Preview’s cybersecurity capabilities \ Anthropicred.anthropic.com
- Vulnerability Research Is Cooked - Quarrelsomesockpuppet.org
- Security incident disclosure — July 2026huggingface.co
- My self-sovereign / local / private / secure LLM setup, April 2026vitalik.eth.limo
- What We’ve Learned From A Year of Building with LLMs – Applied LLMsapplied-llms.org
- AI in 2025: gestalt — LessWronglesswrong.com
- llm-security/README.md at main · greshake/llm-securitygithub.com
- Simulated Users & Sad LLMs1a3orn.com
- 2025: The year in LLMssimonwillison.net