最新

Hello Hacker News,<p>A friend and I recently finished building Stridewars, which is step competition with the addition of power-ups.<p>Power-ups can increase your steps and reduce your opponent&#x27;s steps, adding a bit of fun and strategy to a standard step competition.<p>Power-ups are earned by syncing your steps each day in the Stridewars app and hitting a daily 7,000-step target. The further your team falls behind the leaders, the higher your chances of drawing stronger&#x2F;rarer power-ups.<p>Example Power-ups: - Duvet Day (Rare) – play 10,000 steps for the day - Roller Skates (Epic) – increase your step count by 50% for the day - Shield (Rare) – protect your team from being targeted - Thief (Rare) – steal an uncommon power-up from another team - Sludge (Common) – reduce another team’s steps by 4%<p>A full list of power-ups can be viewed here:<p>https:&#x2F;&#x2F;www.stridewars.com&#x2F;dashboard&#x2F;powerups<p>There’s no power-up equivalent of a Blue Shell. We thought that might be a bit too much.<p>We originally tried using the Fitbit app to sync steps, but the setup was clunky and the data wasn&#x27;t always accurate. So we reluctantly built the Stridewars app. It reads steps directly from your phone (via Health Connect or Apple Health), includes a simple team chat, and links to a web dashboard showing the leaderboard, recent news feed, and your available power-ups.<p>We’re running a one-week competition starting Monday 12th January (local time), open to anyone. If you would like to participate, please register by Saturday 10th January and complete a quick test sync from the Stridewars app. It&#x27;s free to play.<p>Here is the link to join the competition:<p>https:&#x2F;&#x2F;www.stridewars.com&#x2F;join-competition?competitionId=2fec3e01-8921-433c-b99e-2a27acd4020a<p>It does require a quick registration but you can use a fake name and email if you prefer. Please use an address ending in @deleteme.com so the system doesn’t try to send an email. After registering, download the Stridewars app and complete one test sync so you’re ready when the competition starts.<p>We only use your email address for sending communications like reminders and joining information. Your email will be deleted shortly after the competition too.<p>It&#x27;s 5 people per team and you will be randomly assigned to a team. If you want to be on the same team as a friend, send me the email addresses you both used and I’ll place you together. If team sizes are uneven, we have a few “bot walkers” that sync steps daily and donate their power-ups to teammates.<p>Privacy All data is deleted within six weeks of the competition ending, including your account. The only data we collect from the Stridewars app is your daily step count. We don’t send any data to third parties.<p>It&#x27;s all a bit of fun while being a bit more active than you normally would.<p>Happy to answer questions.

Hello HN,<p>I built Archivist because I needed a better way to manage and organize my local video collection. There are plenty of tools for sorting your library, but none that fit my use case. I wanted a tool that could go deeper than just file names—something that could filter by technical properties like bitrate, codecs, and resolution, while also handling metadata.<p>It is a cross-platform desktop application (Windows, macOS, Linux) that packages FFmpeg for analysis.<p>Key Features: Filtering: Filter library by resolution, codec, audio tracks, bitrate, etc.<p>Metadata &amp; Ratings: Fetches ratings&#x2F;metadata from TMDB&#x2F;OMDB and allows editing&#x2F;embedding metadata directly into files.<p>FFmpeg Integration: Built-in support for media analysis, metadata imprinting and file conversions.<p>Internationalization (currently EN, DE, SV).<p>The Tech Stack: Frontend: Angular (which is my daily driver) Runtime: Electron Package Manager&#x2F;Bundler: Bun Language: TypeScript<p>I’m looking for feedback on the usability and the feature set. If you have a large local media library, I’d love to know if this solves any specific pain points for you. Also interested if the ffmpeg file, subtitle editor is useful or if you would anyways use a tool like handbrake. This feature could be extended.<p>Repository: <a href="https:&#x2F;&#x2F;github.com&#x2F;blackfan23&#x2F;archivist" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;blackfan23&#x2F;archivist</a><p>Thanks!

With rings, bands and sunglasses interface are popping up from various companies. Anyone trying to build software for it?

嗨，HN！ 我是正在构建ONYX数据室的团队成员之一。 我们在自己的融资过程中开始着手这个项目，并注意到许多数据室主要关注后期融资流程和预算，而忽视了早期和成长阶段创始人的需求。 让我们最感到突出的问题不是数据的缺乏，而是缺乏清晰度。作为创始人，我们能够看到文件被打开，但很难理解： - 哪些投资者是真正参与的，哪些只是随便浏览 - 哪些文件受到关注，哪些被跳过 - 尽职调查在哪些方面进展缓慢或产生了问题 ONYX专注于让这些变得更加清晰： - 无限的数据室和用户 - 分析功能，突出哪些投资者活跃，哪些文件被阅读 - 内置问答功能，使问题与相关文件保持关联 我们的目标不是增加更多的指标，而是帮助创始人优先处理后续事项，并了解在融资过程中应该花费时间的地方。 如果你想了解更多，可以访问： [https://onyxdataroom.com](https://onyxdataroom.com) 欢迎提出问题或分享其他人今天如何处理尽职调查和投资者信号的经验。 谢谢！

I built Catelingo, a small constraint-based checker that flags semantically impossible LLM outputs, independent of likelihood, retrieval, or chain-of-thought. Many failures (temporal inconsistencies, numeric impossibilities, semantic type clashes) are fluent and high-likelihood. Catelingo reframes “semantic validity” as constraint satisfiability, not plausibility. Intentionally minimal &amp; deterministic: - small sense-level lexicon - explicit constraint propagation (dependency-local) - verdict: SAT &#x2F; UNSAT &#x2F; UNKNOWN - optional degeneration rules for metaphor &#x2F; domain adaptation<p>Paper (Zenodo): <a href="https:&#x2F;&#x2F;doi.org&#x2F;10.5281&#x2F;zenodo.18148498" rel="nofollow">https:&#x2F;&#x2F;doi.org&#x2F;10.5281&#x2F;zenodo.18148498</a>

I built LightningProx - access Claude&#x2F;GPT-4 without API keys or accounts.<p>How it works:<p>1. Send request, get Lightning invoice (~5 sats) 2. Pay with any Lightning wallet 3. Get AI response<p>Payment = authentication. No keys to leak.<p>Python: pip install langchain-lightningprox<p>Site: <a href="https:&#x2F;&#x2F;lightningprox.com" rel="nofollow">https:&#x2F;&#x2F;lightningprox.com</a> Docs: <a href="https:&#x2F;&#x2F;lightningprox.com&#x2F;docs" rel="nofollow">https:&#x2F;&#x2F;lightningprox.com&#x2F;docs</a> GitHub: <a href="https:&#x2F;&#x2F;github.com&#x2F;unixlamadev-spec&#x2F;langchain-lightningprox" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;unixlamadev-spec&#x2F;langchain-lightningprox</a> Full writeup: <a href="https:&#x2F;&#x2F;medium.com&#x2F;@unixlamadev&#x2F;i-built-the-payment-layer-for-ai-agents-5fb2545c5272" rel="nofollow">https:&#x2F;&#x2F;medium.com&#x2F;@unixlamadev&#x2F;i-built-the-payment-layer-fo...</a>

A few years ago I thought Facebook Marketplace was a goldmine.<p>The messages never stopped. “Is this still available?” “Can I pick it up today?” “Very interested.”<p>On paper, demand looked insane.<p>In reality, most of it was fake.<p>I was selling higher ticket items and services to everyday people. Many of them had bad credit, unstable income, or financial stress. Facebook Marketplace put me directly in front of that audience, but it did not filter for ability to buy.<p>I confused attention with intent.<p>I spent hours responding, explaining, following up, and holding items. Deals would feel close, then fall apart at the last moment. Financing applications would fail. Cash buyers would disappear. People would ghost after long conversations.<p>At first I blamed the platform. Then I blamed the buyers.<p>Eventually I realized the problem was me.<p>I had built a system optimized for volume instead of qualification. Facebook Marketplace is incredible at creating conversations, but terrible at signaling seriousness. Bad credit was not the real issue. Unclear expectations were.<p>Many people reached out because it was easy, free, and low commitment. Not because they were ready or able to buy.<p>The failure taught me something important. You cannot scale sales on hope.<p>I had to change everything.<p>I tightened qualification early. I stopped treating every message as a lead. I asked uncomfortable questions sooner. I stopped chasing people who wanted possibilities instead of decisions.<p>That reduced my message count but increased my close rate.<p>The biggest lesson was this. Bad credit customers are not bad customers. But unqualified demand will burn you out faster than no demand at all.<p>Facebook Marketplace still works, but only if you respect what it is. It is a conversation engine, not a closing engine.<p>Once I stopped expecting it to be something it was not, my results improved and my stress dropped.<p>That failure probably saved me years of wasted effort.

created Crystalline Protocol. It treats the blockchain state not just as a key-value store, but as a mathematical set governed by the Axioms of Zermelo-Fraenkel set theory.<p>The PoC (written in Rust) demonstrates an &quot;Axiomatic Engine&quot; that enforces the Axiom of Regularity. It physically prevents circular dependencies (logical loops) at the VM level. If a transaction creates a cycle, the &quot;Logic Firewall&quot; rejects it before it can ever be executed.

I have been cranking out apps for the past few years and loving it. Then one morning a week or 2 ago I got a little ambitious and decided to build a desktop email client because outlook was so-so and superhuman was ridiculously expensive.<p>Is this a big mistake? Am I wasting my time ?<p>So far I have Microsoft and Google working and I can do the normal email interaction read, reply, archive etc. I also got local AI working and training on your emails (all local).<p>Let me know your thoughts.<p>PS: my target audience is developers not the general public