We Are Still Unable to Secure LLMs from Malicious Inputs

Comments (1)

simonw · 3h ago

Bruce Schneier:"We simply don’t know to defend against these attacks. We have zero agentic AI systems that are secure against these attacks. Any AI that is working in an adversarial environment—and by this I mean that it may encounter untrusted training data or input—is vulnerable to prompt injection. It’s an existential problem that, near as I can tell, most people developing these technologies are just pretending isn’t there."

Ask HN: Why hasn't x86 caught up with Apple M series?

Anthropick.com Redirects to ChatGPT

Ask HN: How can I recover and run my old mobile game from the 2010s?

Ask HN: GitHub Copilot down?

Ask HN: Did modern AI's coding abilities make you lose interest in programming?

Ask HN: How to Learn to Build Agentic AI Systems (Like Claude Code)

Ask HN: What are the best Google alternatives in 2025?

Ask HN: What measures are you taking to stop AI crawlers?

Ask HN: Is backlink trading still a problem worth solving?

Ask HN: What to Do with Old iPads?

Units of Economics of LLMs. Reply to Ed Zitron's "AI Is a Money Trap"

Ask HN: Does sentience put stress on the brain?

Ask HN: Is there a temp phone number like temp email?

Out of curiosity: what kind of people use this "forum" (I mean Hacker News)?

Ask HN: Why are so many services rejecting Google Voice numbers for signups?

Ask HN: What should I use to run React Native tests on a device?

Ask HN: Best codebases to study to learn software design?

Stop squashing your commits. You're squashing your AI too

Ask HN: Any experienced devs who use AI extensively in their work?

Why is AI search still bad at trust and context?

Ask HN: Are AI filters becoming stricter than society itself?

Ask HN: How are you attributing your AI usage when developing software?

Patient Lisp Hacker Seeks Same for Long Walks Through IPL-V Code

Ask HN: Has anyone else used online communities that are archetypically "savvy"?

High rate of LLM (GPT5) hallucinations in dense stats domains (cricket)

Ask HN: Why can't ChatGPT count to a million?

Ask HN: What are you working on (August 2025)?

Ask HN: I just abandoned my PyCharm subscription, what should I use now?

429 Too Many Requests from registry.npmjs.org

How can a mutex in Wine be faster than a native one on Linux

Ask HN: What is wrong with modern software development

Ask HN: How do you find early stage startups to join

Ask HN: Recommandation for an Ergonomic Keyboard?

Ask HN: Someone has committed 20K+ LoC to a PR, exhausting my CI & AI workflows

Ask HN: Best Marketplaces for Used Servers?

Ask HN: Does using public transportation make you more creative than driving?

Ask HN: How can I trace what user queries make AI bots crawl my site?

Problem with Payment Gateways

ASK HN: AI in high school. Will teachers and schools have to compensate?

Ask HN: What is your source for answers?

Ask HN: Why do people hate on Sabine Hossenfelder so much?

Ask HN: Devices to allow children to listen to podcasts on my local network?

DSPy GEPA Example: Listwise Reranker

Ask HN: How is Nvidia streaming 4k games while remote control sw is low q lag

Ask HN: What is the biggest problem LLMs solved in your life/work?

Ask HN: No easy way for tvOS to display long documents (e.g., terms of service)?

We Are Still Unable to Secure LLMs from Malicious Inputs

Comments (1)