Hacker News
LLMs Refuse High-Cost Attacks but Stay Vulnerable to Cheap, Real-World Harm
Article URL: https://expectedharm.github.io/
Comments URL: https://news.ycombinator.com/item?id=46955705
Points: 2
# Comments: 0
Report on the subject of Manufacturers (1791) [pdf]
Article URL: https://constitution.org/2-Authors/ah/rpt_manufactures.pdf
Comments URL: https://news.ycombinator.com/item?id=46955546
Points: 1
# Comments: 0
Explaining the PeV neutrino fluxes with quasiextremal primordial black holes
Article URL: https://journals.aps.org/prl/accepted/10.1103/r793-p7ct
Comments URL: https://news.ycombinator.com/item?id=46955545
Points: 1
# Comments: 0
Internet Background Noise
Article URL: https://en.wikipedia.org/wiki/Internet_background_noise
Comments URL: https://news.ycombinator.com/item?id=46955543
Points: 1
# Comments: 0
Show HN: Secure managed hosting for OpenClaw (free and BYOK)
It’s managed hosting for OpenClaw aimed at people who want an always-on personal AI assistant but don’t want to self-host / debug infra.
What it does
- 1 minute setup: sign in with Google → connect Telegram → connect your LLM provider - Isolated runtime: your OpenClaw instance runs isolated (smaller blast radius vs “run on your laptop/VPS”) - Access control: allowlist who can message your bot (DMs + groups) - Keys handled safely: credentials stored encrypted and kept separate from the OpenClaw instance - Usage + cost visibility: dashboard for tokens/cost so you can spot “token burn” - Tutorials: step-by-step guides for OpenAI / Anthropic / Gemini / OpenRouter / Kimi keys + setup credentials - Multiple instances: you can spin up several OpenClaw instances (a "personal AI team" with roles)
Current status / limitations
- Telegram is supported today; more channels later. - You bring your own model/provider keys (BYOK). - Early-stage project — I’d love feedback on onboarding, security assumptions, and what features matter most.
Comments URL: https://news.ycombinator.com/item?id=46955517
Points: 1
# Comments: 0
Ask HN: How do you interpret P99 latency without being misled?
I’ve seen many teams rely heavily on P50/P95/P99 latency numbers, but still miss real user pain or misdiagnose incidents.
Recently I tried to write down a more systematic way to reason about latency distributions in production: how different distribution shapes behave, why aggregation and sampling often lie to us, and why segmentation (by endpoint, tenant, region, workload) usually matters more than adding more percentiles.
I’m curious how others here approach this in practice:
Do you have a mental model for interpreting P99 during incidents?
What charts or breakdowns have actually helped you debug latency issues?
Have you been burned by “good-looking” percentiles that hid real problems?
I wrote up my notes here for reference: https://optyxstack.com/performance/latency-distributions-in-practice-reading-p50-p95-p99-without-fooling-yourself
Would love to hear how people handle this in real systems.
Comments URL: https://news.ycombinator.com/item?id=46955498
Points: 1
# Comments: 0
Trapped Between Pitch, Disclaimer, and Confession
Article URL: https://gilpignol.substack.com/p/trapped-between-pitch-disclaimer
Comments URL: https://news.ycombinator.com/item?id=46955497
Points: 1
# Comments: 0
The Vocabulary Priming Confound in LLM Evaluation [pdf]
Article URL: https://github.com/Palmerschallon/Dharma_Code/blob/main/paper/vocab_priming_confound.pdf
Comments URL: https://news.ycombinator.com/item?id=46955474
Points: 1
# Comments: 0
The committee problem: why B2B demos die after the form
Article URL: https://blog.skipup.ai/buying-committee-demo-scheduling-problem/
Comments URL: https://news.ycombinator.com/item?id=46955464
Points: 1
# Comments: 0
Show HN: Seedance2.live – One place to try many AI image and video models
I've been playing around with a bunch of AI image and video models and got tired of jumping between different sites, so I put them all in one place.
Right now there are 45+ models you can try — text-to-image, text-to-video, image-to-video, effects, etc. You can pick a model yourself or just let it auto-select one for you. No signup needed to try it out.
Honestly I'm really looking forward to Seedance 2.0 — it can do multi-shot video in one go and you can feed it images, video, and audio as references instead of just text. That's huge. The moment the API is available I'm plugging it in.
Would love to hear what you think.
Comments URL: https://news.ycombinator.com/item?id=46955441
Points: 1
# Comments: 1
Show HN: Grok Video 10s – Grok AI video generation and creator contest
I’m the creator of https://grok-video.org, a web platform built for high-quality AI video generation.
What’s currently available:
Grok Video Integration: Full support for Text-to-Video and Image-to-Video using Grok’s video architecture.
Fast Rendering: Optimized for speed to help creators iterate quickly.
The Roadmap: The current version is powered by Grok, but I am actively working on supporting additional open-weights models like Wan 2.1 and Tencent Hunyuan in the coming updates to provide more creative variety.
Join our Contest: We just launched the Grok Video Contest to encourage creators to share their best AI-generated clips. Check it out here: https://grok-video.org/grok-contest
I’d love to hear your thoughts on the user experience and any features you'd like to see added.
website: https://grok-video.org/
Comments URL: https://news.ycombinator.com/item?id=46955437
Points: 2
# Comments: 0
Grumpy Julio plays with CLI coding agents
Article URL: https://jmmv.dev/2026/02/one-week-with-claude-code.html
Comments URL: https://news.ycombinator.com/item?id=46955436
Points: 1
# Comments: 0
The Software Business
Article URL: https://ivanbercovich.com/2026/the-software-business
Comments URL: https://news.ycombinator.com/item?id=46955419
Points: 1
# Comments: 0
Monopoly Round-Up: The $2T Collapse of Terrible Software Companies
Article URL: https://www.thebignewsletter.com/p/monopoly-round-up-the-2-trillion
Comments URL: https://news.ycombinator.com/item?id=46955415
Points: 1
# Comments: 0
Show HN: HelixNotes – Local-first Markdown notes app built with Rust and Tauri
Article URL: https://helixnotes.com
Comments URL: https://news.ycombinator.com/item?id=46955400
Points: 4
# Comments: 0
Security audit of Browser Use: prompt injection, credential exfil, domain bypass
Article URL: https://arxiv.org/abs/2505.13076
Comments URL: https://news.ycombinator.com/item?id=46955396
Points: 2
# Comments: 1
Infinite Terrain
Article URL: https://mesq.me/infinite-terrain/
Comments URL: https://news.ycombinator.com/item?id=46955391
Points: 1
# Comments: 1
"What Questions Do You Have for Me?": Acing the Reverse Interview
Article URL: https://robbygrodin.substack.com/p/what-questions-do-you-have-for-me
Comments URL: https://news.ycombinator.com/item?id=46955390
Points: 1
# Comments: 0
Show HN: SplitFXM – Multi-Dimensional Computational Library for Physics-Aware AI
Article URL: https://splitfxm.com
Comments URL: https://news.ycombinator.com/item?id=46955375
Points: 1
# Comments: 0
Tired of sharing small files via Google Drive or Dropbox just to manage access
Article URL: https://www.styloshare.com/
Comments URL: https://news.ycombinator.com/item?id=46955373
Points: 1
# Comments: 1
