Hacker News

Show HN: No Trace Chat – secret code, no signup, messages disappear after read

Hacker News - Sat, 02/21/2026 - 2:21pm

No Trace Chat is a private messaging app we built for people who want to chat without leaving a trace: no account, no phone number, no discovery—just a secret code (or QR) you share with someone. Messages are E2E encrypted and disappear after the other person reads them.

How it works: you create or join a “room” with a 16+ character code (or scan a QR). Only people with that code can see the conversation. Keys are derived from the code on device; we never have the key. “Seen & delete” removes the message for everyone after it’s read.

We added optional privacy levels (e.g. don’t save codes on device), screenshot blocking by default, and no push notifications so usage stays as private as possible.

Android and iOS, freemium (free tier + one-time unlock). Would love feedback from anyone who cares about minimal-footprint messaging.

Landing: https://apps.teamzlab.com/no-trace-chat Android: https://play.google.com/store/apps/details?id=com.teamzlab.n... iOS: https://apps.apple.com/us/app/no-trace-chat-private/id675886...

Comments URL: https://news.ycombinator.com/item?id=47103780

Points: 1

# Comments: 0

Categories: Hacker News

Human and AI. How to Modernize Your Consciousness in 10 Minutes

Hacker News - Sat, 02/21/2026 - 1:57pm

For someone let this be simply a joke, and for someone this will become a serious life turning point. Simply answer several questions to yourself. Reading these questions, you must understand that you are not only trying to position yourself, but at the same time you also learn about the opinion of other people through the text.

I repeat, here the main idea is not a questionnaire, this is intended as a means of personal psychological support in positioning oneself. For many questions the answer should/can be "don't know yet." But you will now know that such a dimension exists and will more meaningfully look at all kinds of arguments for the first, for the second, or third. Also you will begin to understand that you are surrounded by a multitude of different-type, different-vector in these senses people.

---

1. What do you see in artificial intelligence (hereinafter - "smart chats")?

- A service? - An exoskeleton? - A partner? - All of the above in turn in an incomprehensible order? - All of the above in turn in a meaningful order?

---

2. How do you see for yourself the main purpose of smart chats?

- Help in qualification educational works? - Help in programming? - Help in self-development? - Help in learning in general? - Literary creativity? - Reports for work? - Help in self-development? - Help in the depths of science? - Survey self-education? - Help with correspondence? - Help with writing articles? - Help with translation? - Help with reading large texts? - Big help with your main subject at work, in hobby? - ... (here the list is large, if something "other," please add in comments) ... 10. To what extent do you accept a person's answer using a "smart chat" to your important question of interest?

- I want the person himself to answer! - For the sake of quality and speed I accept that a person with a smart chat answers! - A correct and understandable answer is important, not who answers! - I clearly understand: where exclusively a person is needed, where a person with chat is possible, and where simply an answer of a smart chat is acceptable.

---

Do you think it's worth developing and extending this topic further? Something interesting, useful?

Full version: https://zenodo.org/records/18726277

Comments URL: https://news.ycombinator.com/item?id=47103582

Points: 1

# Comments: 0

Categories: Hacker News

Show HN: InferShield – open-source security proxy for LLM inference

Hacker News - Sat, 02/21/2026 - 1:53pm

Hey HN! I'm launching InferShield, an open source security proxy that sits between your application and LLM providers (OpenAI, Anthropic, etc.) to detect and block threats in real-time.

I've been talking to CISOs at banks and hospitals who are deploying LLMs without proper security. Their existing tools can't handle LLM-specific threats:

- Prompt injection (attackers manipulate LLM behavior, WAFs can't detect it) - Data exfiltration (sensitive data leaks through LLM responses) - Jailbreak attempts (users bypass safety guardrails using encoding)

Gartner predicts 60% of AI enterprises will face a security incident by 2027.

InferShield is a drop-in proxy with real-time threat detection, multi-encoding detection (Base64, hex, URL, Unicode), complete audit logs, and risk scoring. Self-hosted, provider-agnostic, zero code changes required.

v0.1 MVP is live today with 95%+ detection rate (red team tested). MIT licensed, free forever. Enterprise tier coming for compliance features.

Website: https://infershield.io Quick start: docker pull infershield/proxy:latest

Feedback welcome! What LLM security challenges are you facing?

Comments URL: https://news.ycombinator.com/item?id=47103529

Points: 1

# Comments: 0

Categories: Hacker News

The Luol Deng Law

Hacker News - Sat, 02/21/2026 - 1:43pm
Categories: Hacker News

Our Modern Mistake

Hacker News - Sat, 02/21/2026 - 1:39pm
Categories: Hacker News

Pages