frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Get found where people search today

https://kleonotus.com/
1•makenotesfast•2m ago•1 comments

Show HN: An early-warning system for SaaS churn (not another dashboard)

https://firstdistro.com
1•Jide_Lambo•2m ago•0 comments

Tell HN: Musk has never *tweeted* a guess for real identity of Satoshi Nakamoto

1•tokenmemory•3m ago•0 comments

A Practical Approach to Verifying Code at Scale

https://alignment.openai.com/scaling-code-verification/
1•gmays•5m ago•0 comments

Show HN: macOS tool to restore window layouts

https://github.com/zembutsu/tsubame
1•zembutsu•7m ago•0 comments

30 Years of <Br> Tags

https://www.artmann.co/articles/30-years-of-br-tags
1•FragrantRiver•14m ago•0 comments

Kyoto

https://github.com/stevepeak/kyoto
2•handfuloflight•15m ago•0 comments

Decision Support System for Wind Farm Maintenance Using Robotic Agents

https://www.mdpi.com/2571-5577/8/6/190
1•PaulHoule•15m ago•0 comments

Show HN: X-AnyLabeling – An open-source multimodal annotation ecosystem for CV

https://github.com/CVHub520/X-AnyLabeling
1•CVHub520•18m ago•0 comments

Penpot Docker Extension

https://www.ajeetraina.com/introducing-the-penpot-docker-extension-one-click-deployment-for-self-...
1•rainasajeet•19m ago•0 comments

Company Thinks It Can Power AI Data Centers with Supersonic Jet Engines

https://www.extremetech.com/science/this-company-thinks-it-can-power-ai-data-centers-with-superso...
1•vanburen•22m ago•0 comments

If AIs can feel pain, what is our responsibility towards them?

https://aeon.co/essays/if-ais-can-feel-pain-what-is-our-responsibility-towards-them
3•rwmj•26m ago•4 comments

Elon Musk's xAI Sues Apple and OpenAI over App Store Drama

https://mashable.com/article/elon-musk-xai-lawsuit-apple-openai
1•paulatreides•29m ago•1 comments

Ask HN: Build it yourself SWE blogs?

1•bawis•29m ago•1 comments

Original Apollo 11 Guidance Computer source code

https://github.com/chrislgarry/Apollo-11
3•Fiveplus•35m ago•0 comments

How Did the CIA Lose Nuclear Device?

https://www.nytimes.com/interactive/2025/12/13/world/asia/cia-nuclear-device-himalayas-nanda-devi...
1•Wonnk13•35m ago•0 comments

Is vibe coding the new gateway to technical debt?

https://www.infoworld.com/article/4098925/is-vibe-coding-the-new-gateway-to-technical-debt.html
1•birdculture•39m ago•1 comments

Why Rust for Embedded Systems? (and Why I'm Teaching Robotics with It)

https://blog.ravven.dev/blog/why-rust-for-embedded-systems/
2•aeyonblack•40m ago•0 comments

EU: Protecting children without the privacy nightmare of Digital IDs

https://democrats.eu/en/protecting-minors-online-without-violating-privacy-is-possible/
3•valkrieco•41m ago•0 comments

Using E2E Tests as Documentation

https://www.vaslabs.io/post/using-e2e-tests-as-documentation
1•lihaoyi•41m ago•0 comments

Apple Welcome Screen: iWeb

https://www.apple.com/welcomescreen/ilife/iweb-3/
1•hackerbeat•43m ago•1 comments

Accessible Perceptual Contrast Algorithm (APCA) in a Nutshell

https://git.apcacontrast.com/documentation/APCA_in_a_Nutshell.html
1•Kerrick•44m ago•0 comments

AI agent finds more security flaws than human hackers at Stanford

https://scienceclock.com/ai-agent-beats-human-hackers-in-stanford-cybersecurity-experiment/
3•ashishgupta2209•45m ago•2 comments

Nano banana prompts, updates everyday

https://github.com/fionalee1412/bestnanobananaprompt-github
4•AI_kid1412•49m ago•0 comments

Skills vs. Dynamic MCP Loadouts

https://lucumr.pocoo.org/2025/12/13/skills-vs-mcp/
3•cube2222•53m ago•0 comments

Top validated AI-SaaS Ideas are available here

1•peterbricks•57m ago•0 comments

UnmaskIP: A Clean, Ad-Free IP and Deep Packet Leak Checker

https://unmaskip.net
1•kfwkwefwef•1h ago•0 comments

PydanticAI-DeepAgents – AI Agent Framework planning, filesystem, and subagents

https://github.com/vstorm-co/pydantic-deepagents
1•kacper-vstorm•1h ago•1 comments

DeepCSIM – Detect duplicate and similar code using AST analysis

https://github.com/whm04/deepcsim
1•whm04•1h ago•1 comments

Chip‐8 Technical Reference

https://github.com/mattmikolay/chip-8/wiki/CHIP%E2%80%908-Technical-Reference
1•AlexeyBrin•1h ago•0 comments
Open in hackernews

I analyzed 10k LinkedIn posts to understand what drives engagement

https://www.hyperclapper.com
1•neerajnegi1003•1h ago

Comments

neerajnegi1003•1h ago
After spending way too much time manually engaging on LinkedIn to grow my startup's presence, I got curious: what actually works? The Problem Most devs and founders hate LinkedIn's engagement game. You post something technical or meaningful, get 3 likes from bots, while someone posting "5 lessons from my cat about product management " gets 10,000 impressions. But here's what the data actually showed after analyzing top-performing posts: Timing matters more than we think:

Posts between 7-9 AM and 5-7 PM (local time) get 3x more initial engagement The first 60 minutes are critical for algorithmic distribution Early engagement signals quality to LinkedIn's feed algorithm

The cold start problem is real:

Without initial engagement, even great content dies in obscurity LinkedIn's algorithm uses the first 10-20 interactions to determine if your post is "worthy" Most technical posts never escape this cold start because our networks are small

What I Built I created Hyperclapper - a simple tool that helps you coordinate genuine engagement within your network. Not bots, not fake engagement, just helping real people find and support each other's content. How it works:

Connect your LinkedIn (read-only OAuth) Join engagement pods with people in similar industries Get notified when someone posts (you choose how often) Engage authentically with content you actually find valuable

Why this isn't spammy:

You control what you engage with - no forced likes All engagement is from real accounts You're helping surface good content that would otherwise be buried Think of it as a RSS feed for your network's best content

The Results After 3 months of testing with 500 beta users:

Average post impressions increased 4.2x Meaningful comments (not just "Great post!") up 6x Several users got inbound leads directly from increased visibility

The trickiest part was building the engagement pod matching algorithm. We use a combination of:

Industry tags Posting frequency compatibility Historical engagement patterns Mutual connection depth

Why I'm Sharing This LinkedIn engagement is a grind, especially for technical folks who just want to share knowledge. This tool helps solve the cold start problem without resorting to engagement bait or algorithmic hacks.

I'm launching the beta publicly this week. Would love feedback from the HN community since you folks are typically skeptical of social media growth tools (rightfully so).

Try it: hyperclapper.com

Genuine question for HN: How do you balance the need for visibility with the distaste for "engagement optimization"? Is there a way to play the game without losing your soul?