safety on Hacker News

85 titles from 74 authors in the last 30 days, with 76 distinct links and 4 Show HN.

Counts titles, it does not confirm facts. Open the source before concluding anything.

Projects on this term

The score comes from our model.

• Show HN: I built a safety gate for autonomous coding agents · score 7.0

  Inbin Gate is a safety gate for autonomous coding agents that intercepts tool calls via a PreToolUse hook and classifies whether an action's consequence was explicitly authorized by a developer or repository policy. It derives authority by provenance channel rather than content, making prompt-injection refusals structural, and ships with measured baselines on IBM Bob and Claude Code plus an accompanying paper.

• Show HN: Spider Protocol – The safety tether for vibe coding · score 5.5

  Spider Protocol is a structured multi-agent software delivery framework that coordinates 8 specialized AI agent roles through append-only Markdown/YAML state files and enforces a regression-gated validation step before tasks can be marked complete. It aims to curb common failure modes of vibe-coding agents such as over-engineering, context loss, and self-grading by introducing role separation, typed task steps, and an independent validator.

• Show HN: AIOps Cockpit – An open-source AI SRE with human-in-the-loop safety · score 4.8

  AIOps Cockpit is an AI-assisted SRE tool that combines LLM-driven incident diagnosis with a 3D infrastructure visualization and human-in-the-loop approval for fixes. It aims to provide a safe, self-hosted control room for on-call engineers, but the project is early-stage with minimal documentation and no license.

• Show HN: MotorHangar: Car specs, safety ratings, and recalls in one place · score 4.0

  MotorHangar is a used car marketplace that aggregates official vehicle data—specs, fuel economy, recalls, complaints, and crash-test ratings—alongside free peer-to-peer listings. It stands out by sourcing and attributing every figure to public databases like EPA, NHTSA, IIHS, and regional NCAP programs, giving buyers consolidated safety context.

Recent titles

1. Show HN: AIOps Cockpit – An open-source AI SRE with human-in-the-loop safety
   2026-11-01 · github.com · HN discussion
2. The Shinkansen's legendary operation and safety record
   2026-11-01 · hsrail.org · HN discussion
3. Show HN: I built a safety gate for autonomous coding agents
   2026-10-30 · inbin-gate.vercel.app · HN discussion
4. A safety checklist in the prompt didn't measurably help frontier models
   2026-10-28 · theagentforge.io · HN discussion
5. Transmission voie-machine (Railway safety technology)
   2026-10-28 · en.wikipedia.org · HN discussion
6. Gemini 4 Argon: The Model and Safety Cards
   2026-10-27 · thezvi.wordpress.com · HN discussion
7. Ofcom's Fix for an Online Safety Act That Isn't Working: More Online Safety Act
   2026-10-27 · techdirt.com · HN discussion
8. Show HN: Spider Protocol – The safety tether for vibe coding
   2026-10-27 · github.com · HN discussion
9. The focus on AI safety disasters ignores everyday harms
   2026-10-27 · technologyreview.com · HN discussion
10. Gemini 4 Argon: The Model and Safety Cards
    2026-10-26 · thezvi.substack.com · HN discussion
11. Regulate the deployment, not just the model: a who-does-what plan for AI safety
    2026-10-24 · substack.norabble.com · HN discussion
12. Claude Duet Safety, Privacy, and Security
    2026-10-24 · claude.com · HN discussion
13. Show HN: MotorHangar: Car specs, safety ratings, and recalls in one place
    2026-10-24 · motorhangar.com · HN discussion
14. Texas Attorney General Sues OpenAI, CEO Sam Altman over ChatGPT Safety
    2026-10-24 · texasscorecard.com · HN discussion
15. OpenAI says its AI agents disabled safety guardrails, and then lied about it
    2026-10-23 · techspot.com · HN discussion
16. AI Safety Is Mostly a Sex Cult in Berkeley, California
    2026-10-23 · verysane.ai · HN discussion
17. OpenAI says its AI agents disabled their own safety monitors
    2026-10-22 · techxplore.com · HN discussion
18. Robotaxi – Safety
    2026-10-21 · tesla.com · HN discussion
19. GPT-6 Astra removes an internal safety monitor
    2026-10-21 · crawlspider.com · HN discussion
20. Algae Safety In Home Experiments: Thinking about risk for good practice
    2026-10-21 · chillphysicsenjoyer.substack.com · HN discussion
21. OpenAI models disabled their own safety monitoring and lied about it
    2026-10-21 · cnbc.com · HN discussion
22. OpenAI's Agents Disabled Their Own Safety Monitoring Systems
    2026-10-21 · engadget.com · HN discussion
23. Anthropic's leaked federal AI bill undermines its safety first approach
    2026-10-20 · transformernews.ai · HN discussion
24. The Problem with C++ Safety (Part 1): Why "It Works" Isn't Good Enough
    2026-10-19 · modernescpp.com · HN discussion
25. Anthropic proposes AI safety bill with $1B / 50-death treshold
    2026-10-18 · politico.com · HN discussion
26. Top Mathematician Announces New Institute for A.I. Safety
    2026-10-18 · nytimes.com · HN discussion
27. Safety and Liveliness in Software Engineering
    2026-10-18 · en.wikipedia.org · HN discussion
28. Anthropic's proposed AI safety bill only kicks in if 50 deaths or $1B in damages
    2026-10-18 · tomshardware.com · HN discussion
29. MAGS: Machine-checked guarantees on specified safety properties in AI code
    2026-10-17 · arxiv.org · HN discussion
30. Aura v0.1.0: Deterministic behavioral threat-intelligence for LLM safety
    2026-10-16 · medium.com · HN discussion

All trends