Top all-time · #5281
Lakera: The AI-Native Security Platform to Accelerate GenAI
Company blog · lakera.ai/blogaisecurity
Lakera: The AI-Native Security Platform to Accelerate GenAI is the company blog at lakera.ai, writing about ai and security. Since 2021, 15 different Hacker News readers have linked it 27 times, most often “The Beginner's Guide to Visual Prompt Injections (2023)”. No feed was found, so read it on the site itself.
Too few sampled comments (7) to say how often HN recommends it.?
Why HN links it
HN commenters link Lakera's posts on AI security, such as Claude 4 weaknesses, training-data poisoning and visual prompt injection, as support for their own points. No comment praises the writing directly.
Summarised by Claude from a sample of 7 HN comments that link here. Each comment is marked recommending, critical or neutral (citing without judging); how this works.
The AI-native security platform to accelerate GenAI initiatives—trusted by Fortune 500s, backed by the world’s largest AI red team.
Most-linked posts?
- The Beginner's Guide to Visual Prompt Injections (2023)2 people · 3 mentions · first linked 2023-11-01
- /2 people · 2 mentions · first linked 2024-03-05
- The computer vision bias trilogy: Data representativity2 people · 2 mentions · first linked 2022-03-09
What Hacker News says?
https://www.lakera.ai/blog/claude-4-sonnet-a-new-standard-fo... These LLMs still fall short on a bunch of pretty simple tasks. Attackers can get Claude 4 to deny legitimate requests easily by manipulating third party data sources for example.
Comment linking /blog/claude-4-sonnet-a-new-standard-for-secure-enterprise-llms
on “Claude 4 System Card”Submitted The Beginner's Guide to Visual Prompt Injections (2023) as a story · 183 points
HN discussionLakera AI ( https://lakera.ai ) | Full-time | Zurich, Switzerland | ONSITE We're looking for Software, Infrastructure, Platform, and ML Engineers (Senior-Staff level) in our office in Zurich. Our team works on AI security (our flagship product is a layer for GenAI threat detection). Our current customers include Dropbox, Twilio and other large F500 companies. We've created the game Gandalf, which helped us collect 2…
Comment linking /careers
on “Ask HN: Who is hiring? (February 2025)”Submitted The computer vision bias trilogy: Data representativity as a story · 6 points
HN discussionLakera | San Francisco | Staff SW Engineer I'm looking for a few Staff-level software engineers to kick-start our US engineer team. We're building guardrails and safety for anyone and everyone deploying LLMs. Lakera is on a mission to secure the era of intelligent computing. We are heading towards a future where AI-powered applications take center stage. Here at Lakera, we're not just dreaming about the future; we'r…
Comment linking /careers?ashby_jid=c269a2d5-6a6d-432b-848f-e7bfc25817c0
on “Ask HN: Who is hiring? (June 2024)”Submitted The computer vision bias trilogy: Shortcut learning as a story · 5 points
HN discussionit's highly similar to what https://www.lakera.ai/ does. I'm wondering if lakera.ai will change their strategy now that a big player is offering the same product
Comment linking /
on “Cloudflare Announces Firewall for AI”Submitted Detecting Data Bugs as a story · 4 points
HN discussionYou have to assume all publicly available data is used for training. If you don’t want people using your images, consider data poisoning techniques to punish systems that train on them. https://www.lakera.ai/blog/training-data-poisoning https://www.technologyreview.com/2023/10/23/1082189/data-poi...
Comment linking /blog/training-data-poisoning
on “AI Has Created a Battle over Web Crawling”Submitted How did software get so reliable without proof? as a story · 4 points
HN discussionMight want to check out https://www.lakera.ai/ you send every prompt to their API first and they deal with checking it for problematic content etc. I'm not affiliated but I've enjoyed hacking their game https://gandalf.lakera.ai/ [0] They also have a sandbox/free API if you want to test it out (they used to at least) [0] previous hn discussion https://news.ycombinator.com/item?id=35905876
Comment linking /
on “Ask HN: Best practices to control LLM responses with user queries?”Submitted Free of bias? We need to change how we build ML systems as a story · 4 points
HN discussionSeems like a way to use a sledgehammer to hammer in screws, and inviting nondeterminism in important systems. Besides being way larger and more complex than what most specialized industrial processes need, they are also vulnerable to adversarial attacks. https://www.lakera.ai/blog/visual-prompt-injections https://www.theverge.com/2021/3/8/22319173/openai-machine-vi...
Comment linking /blog/visual-prompt-injections
on “Qwen3.5 Fine-Tuning Guide – Unsloth Documentation”Submitted Zero-Click Remote Code Execution: Exploiting MCP and Agentic IDEs as a story · 3 points
HN discussionSubmitted Selecting Machine Learning Models as a story · 3 points
HN discussionSubmitted Drift Monitoring for Computer Vision as a story · 3 points
HN discussionSubmitted How to get better test data with metamorphic relations as a story · 3 points
HN discussionSubmitted Algorithms Are the New Drugs as a story · 3 points
HN discussionSubmitted The Security Company of the Future as a story · 2 points
HN discussionSubmitted The Beginner's Guide to Visual Prompt Injections as a story · 2 points
HN discussionSubmitted You shall not pass: the spells behind Gandalf as a story · 2 points
HN discussionSubmitted The computer vision bias trilogy: Data representativity as a story · 2 points
HN discussionSubmitted The Backbone Breaker Benchmark: Testing the Real Security of AI Agents as a story · 1 points
HN discussionSubmitted The Rise of the Internet of Agents: A New Era of Cybersecurity as a story · 1 points
HN discussionSubmitted Prompt Injection Handbook as a story · 1 points
HN discussionSubmitted Detecting Data Bugs as a story · 1 points
HN discussion
