github.com
Show HN: Girlfriend in CLI – AI-native persona harness for social traininghttps://github.com/NomaDamas/girlfriend-in-cliBeing good at coding alone is no longer enough.In the AI-native era, the ability to talk to people, empathize, and convey your charm is also part of your stack.Feeling lonely while vibe coding? Do you want to find a real-world girlfriend or boyfriend? Do you want to measure your dating combat power and see whether you could actually date IU?With a single brew install click, talk in your terminal while vibe coding with the person you want to become lovers with. Just enter a name or a link, and it will do the deep research you want, create a person-based persona, and let you immediately talk to the kind of person you want. While you talk, it will help you discover your charm, and at the end, you can receive a dating combat power evaluation!girlfriend in cli will equip you with a harness inside the terminal that helps you become an attractive person!!Wake up, nerds! Let’s escape being single!!! Stars are always welcome hahaApr 13, 2026 3:55 PM
news.ycombinator.com
Show HN: Zero-identity messaging app with physics-based post-quantum encryptionhttps://news.ycombinator.com/item?id=47751692Show HN: Zero-identity messaging app with physics-based post-quantum encryption (Layer 2 from my own paper)Hey HN,I'm building a privacy-first messaging app in Flutter/Dart, developed with AI assistance (Gemini 2.5 Pro + Claude Opus 4.6). No phone number, no email, no account — no identity, nothing to track. Currently focused on Android.*Six-layer defense architecture*Layer 0 - Identity layer: Zero identifier. No account, no identity, nothing to track.Layer 1 - Key exchange: X25519 + ML-KEM-768 hybrid. X25519 is Bernstein-designed, outside the NIST ecosystem. Even if ML-KEM has an NSA backdoor, X25519 still holds. The probability of both being broken simultaneously approaches zero.Layer 2 - Key hardening (my original contribution): session_key = HKDF(shared_secret || Y), where Y = PQ_C(W(shared_secret)). Even if Layer 1 is fully compromised, the attacker still faces the Landauer thermodynamic lower bound + Margolus-Levitin operation limit. Physical law, not a mathematical assumption.LaApr 13, 2026 1:29 PM
esa.int
Artemis II: around the Moon in 10 dayshttps://www.esa.int/ESA_Multimedia/Videos/2026/04/Artemis_II_around_the_Moon_in_10_daysArtemis II completed a 10-day journey around the Moon, carrying humanity farther into space than it has gone in over 50 years.Apr 13, 2026 11:30 AM
news.ycombinator.com
Artemis II: Christina Koch's PCD Failurehttps://news.ycombinator.com/item?id=47747307Early in the Artemis 2 Mission, Christina Koch's PCD (personal computing device, essentially a Surface tablet) completely failed to power on, and Houston had her attempt to perform a hard reset (hold down the power button, and the volume up button) several times, later having her remove the SD card and try again, at one point referring to last ditch efforts to get it running as a "hail-mary."These efforts were ultimately unsuccessful, and this equipment failure led to the crew having to share three working PCDs between the four of them for all of their work for the entire remaining duration of the mission. They were able to make this configuration work, albeit not without complications. Their PCDs were their direct means of tracking their jobs, documenting their work, running checklists, looking up procedures, etc. During the lunar fly-by in particular, while they were doing photography of their various lunar targets, they said that they had to swap out SD cards as they rotated betweenApr 13, 2026 3:40 AM
bing.com
Microsoft wants you to know Copilot AI is not just for entertainmenthttp://www.bing.com/news/apiclick.aspx?ref=FexRss&aid=&tid=6a8cf3fca5fe4e7ea5ea1a3b94a86096&url=https%3A%2F%2Fwww.msn.com%2Fen-us%2Fmoney%2Fother%2Fmicrosoft-wants-you-to-know-copilot-ai-is-not-just-for-entertainment%2Far-AA20KCUL&c=11828545017293087845&mkt=en-usMicrosoft appears to be trying to clear up an awkward contradiction around its Copilot AI. After one of its own documents made the AI sound a lot less useful than the company’s marketing would suggest ...Apr 13, 2026 1:25 AM
bing.com
We can’t activate Windows on this device as we can’t connect to your organizations serverhttp://www.bing.com/news/apiclick.aspx?ref=FexRss&aid=&tid=6a8cc30230774628b5e870deb04a99bf&url=https%3A%2F%2Fwww.thewindowsclub.com%2Fwe-cant-activate-windows-on-this-device&c=340901163106550959&mkt=en-usWindows Activation errors are one of the most common errors users face post-upgrade, or even switching to a new Windows 11/10 PC. This is, of course, a problem with the Windows Activation server, ...Apr 12, 2026 8:32 PM
github.com
Show HN: Revdiff – TUI diff reviewer with inline annotations for AI agentshttps://github.com/umputun/revdiffI built a terminal diff viewer for a workflow I couldn't do comfortably with existing tools: reviewing AI-generated code changes without leaving the terminal session where the agent runs, annotating what needs to change, and feeding those annotations straight back to the agent. Plenty of diff viewers exist, and some can even feed notes back to an agent, but none of them fit that flow for me - they pulled me out of the terminal into a separate app, or the round-trip back to the agent was clunky.Revdiff is how I solved it. From a Claude Code session I type `/revdiff main` or just say "review diff for last 3 commits" - it opens as a terminal overlay on top of the agent session. I read the diff, drop annotations on any line, hunk, or file, then quit. The annotations go straight back to the agent, which picks them up and starts working on them. When the agent is done, revdiff reopens on the new changes so I can annotate again - and this loop continues until I quit without leaving any annotaApr 12, 2026 5:52 PM
specsight.app
Show HN: Specsight – Living product specs generated from your codebasehttps://specsight.appHeyy HN, I'm OlaI'm an engineer myself, and everywhere I've worked there was a similar dynamic: someone from the non-technical side of the team (PM, CS, stakeholders) needs to understand what changed recently or how something works today. They rely on stale Confluence pages or ping engineers in Slack and wait. engineers get interrupted daily answering "how does this actually work". it's a mess that gets worse as the team growsor when you join a new company and try to get a sense of the product. turns out the documentation doesn't exist or is months old. you ask 5 people and get 5 different answers. I've seen it and experienced it sooo many timesI decided to start exploring this space and built Specsight for those non-technical teams. it connects to your GitHub repos, analyses the codebase with Claude (Agent SDK), extracts business-level features of the product, and generates plain-language specifications in Context/Action/Outcome format for every feature. it also builds a visual graph Apr 12, 2026 5:39 PM
bing.com
Meeting their new baby sister for the first timehttp://www.bing.com/news/apiclick.aspx?ref=FexRss&aid=&tid=6a8d24872fe64470979efcfbd9a5fdea&url=https%3A%2F%2Fwww.msn.com%2Fen-us%2Fvideo%2Fpeopleandplaces%2Fmeeting-their-new-baby-sister-for-the-first-time%2Fvi-AA20LPcn&c=16353725414712088268&mkt=en-usThe family introduces and meets their new baby sister together. Woman posts picture of daughter online, one detail has everyone worried Trump scores Supreme Court win on mail-in voting restrictions ...Apr 12, 2026 5:00 PM
news.ycombinator.com
Show HN: Android AI agent-assistant operating your apps (no adb,PC,root,etc.)https://news.ycombinator.com/item?id=47738583Hi HN,We built Sova AI https://ayconic.io/sova, an Android assistant agent that actually controls and operates your apps. It's not a chat and not another LLM wrapper.We were incredibly frustrated with the current state of mobile AI. Built-in assistants like Gemini are deeply integrated into the OS, yet if you ask them to "Order an Uber to the airport", they mostly just give you web search results or a button to open the app yourself. They don't do the work. (The Perplexity "assistant" is just a browser agent :/ )So, we built an agent that does operate your phone. (NO root, NO adb, NO PC, NO appium/whatever, NO usb, NO browser)How it works: You give Sova a prompt - either voice or text, you can make it a default assistant if you like. Instead of relying on non-existent official app APIs, Sova acts as a virtual human - clicks, scrolls, types etc. It uses the Android Accessibility API to read the screen's UI node tree. About AI models - currently we support main AI cloud providers (OpenAIApr 12, 2026 11:48 AM
brasilia-phi.vercel.app
Show HN: A living Vancouver. Connor is walking dogs at the SPCA this morninghttps://brasilia-phi.vercel.appI've spent most of my career in marketing, which for the last few years has meant building consumer personas for campaigns. I wanted to see if I could make these real, living in real neighborhoods, had real weather, real budgets, real Saturday lunches. I always wanted to build a world, not a segment.This is that. 140 people so far, split across Vancouver (100), San Francisco (20), and Tokyo (20). Each one is about 1,000 lines of profile — family, finances, daily schedule, health, worldview, media diet, the channels you'd actually reach them through and the ones that will explicitly never work on them. Demographics are census-grounded income, age, ethnicity, household composition follow normal distributions against StatsCan, ACS, and Japanese e-Stat data, so the panel is roughly representative of the city instead of representative of whatever's overrepresented in an LLM's training corpus. The specific details come from real stories.They live in real local time on a live map. Right now iApr 11, 2026 6:42 PM
nasa.gov
Artemis II Splashes Downhttps://www.nasa.gov/image-article/artemis-ii-splashes-down/This image from April 10, 2026, captures NASA’s Orion spacecraft, with its parachutes deployed, seconds before splashdown in the Pacific Ocean. The Artemis II crew accomplished many milestones on their nearly 10-day mission, surpassing the Apollo 13 record for farthest crewed spaceflight and capturing views of the far side of the Moon. Under Artemis, NASA […]Apr 11, 2026 11:05 AM
news.ycombinator.com
Ask HN: Would you take your engineering team to Buenos Aires for an offsite?https://news.ycombinator.com/item?id=47726018Hey HN,I’m doing some market research on the logistics and viability of hosting corporate tech offsites in Buenos Aires, Argentina.With the shift to remote/distributed work, annual or bi-annual company retreats have become the main way teams build rapport. We see a lot of teams going to Mexico City, Lisbon, or Costa Rica, but I'm looking closely at Buenos Aires (as I live there).Timezone: Favorable for the US (EST +1 or +2) and easier for EU overlap.Cost: Highly favorable exchange rate makes luxury/premium experiences much more affordable.Culture: World-class food, infrastructure, and a strong local tech ecosystem.However, the flight time from the US West Coast or parts of Europe is significantly longer than going to Mexico or Portugal.I’d love to hear from founders, CTOs, and engineering managers who have planned or attended company retreats:1. Has your team done an offsite in LATAM? If so, where, and how did you choose the location?2. Is flight duration the ultimate dealbreaker? WoulApr 11, 2026 12:59 AM
news.ycombinator.com
Show HN: BNNR – a closed-loop pipeline for improving vision modelshttps://news.ycombinator.com/item?id=47719542Hi HN,we’ve been working on computer vision models for a while, and one thing kept coming up: improving them is surprisingly unstructured.You train a model, try a few augmentations, tweak some hyperparameters, run it again - and if the metric goes up, you keep it. But it’s often unclear why it improved, or whether the model is actually learning something better.We kept running into questions like:- Did this change actually help, or is it just noise?- Is the model focusing on the right features?- Are we improving generalization, or just overfitting differently?So we built BNNR (Bulletproof Neural Network Recipe) - an open-source PyTorch toolkit that turns model improvement into a closed loop:- Train a controlled baseline- Explain what the model actually learned (via XAI methods like OptiCAM / GradCAM)- Improve by testing candidate strategies in parallel- Prove the result with structured comparisonsOne thing we focused on is avoiding “blind” changes. Instead of committing to a single ideApr 10, 2026 3:24 PM
news.ycombinator.com
Ask HN: Will we ever do something about patent trolls?https://news.ycombinator.com/item?id=47718079Some of you may have seen a project like Mycroft AI (nothing to do with LLMs!) from back in around 2017 or so, it was an open source Alexa alternative, even had skills and everything. I kept wondering why it quietly disappeared, and the repo was archived. Turns out a patent troll hit them hard, drained them of all their funds in court, till they shut down.https://patentprogress.org/2024/05/another-startup-bites-the-dust-courtesy-of-patent-trolls/https://www.theregister.com/2023/02/13/linux_ai_assistant_killed_off/Then there's many others, and I'm sure I'm missing some, but here's a list of companies / projects hit by patent trolls with various outcomes:That time some company shut down a free / open source site with a patent that came 10 years after the technology already existed.https://www.techdirt.com/2014/11/19/patent-troll-kills-open-source-project-speeding-up-computation-erasure-codes/https://x.com/JamesBessen/status/532906754364149760Here's one where the courts did the right thinApr 10, 2026 1:45 PM
bing.com
The ‘Emotional Support Daughter’: How Family Roles Shape Women’s Mental Healthhttp://www.bing.com/news/apiclick.aspx?ref=FexRss&aid=&tid=6a8cef36705e4d8ca24e5958899033f3&url=https%3A%2F%2Fwww.forbes.com%2Fsites%2Fmaiahoskin%2F2026%2F04%2F10%2Fthe-emotional-support-daughter-how-family-roles-shape-womens-mental-health%2F&c=17239547127636139698&mkt=en-usThis voice experience is generated by AI. Learn more. This voice experience is generated by AI. Learn more. She knows when something is wrong before anyone says a word. She can feel the shift in tone, ...Apr 10, 2026 8:56 AM
nasa.gov
Starstruckhttps://www.nasa.gov/image-article/starstruck/The Artemis II crew captured this photo of our galaxy, the Milky Way, on April 7, 2026. The Milky Way’s elegant spiral structure is dominated by just two arms wrapping off the ends of a central bar of stars. Spanning more than 100,000 light-years, Earth is located along one of the galaxy’s spiral arms, about […]Apr 9, 2026 8:29 PM
news.ycombinator.com
Show HN: Jsonl Viewer – An offline JSON Lines viewer in a single HTML filehttps://news.ycombinator.com/item?id=47699179I build AI agents and got tired of squinting at raw JSONL chat logs. Existing online viewers either choke on JSONL or upload your data. So I made one that runs offline in a single HTML file. Paste JSONL → auto-formatted with syntax highlighting. Auto-expands nested JSON strings (common in tool-call args). Toggle to render \n as real line breaks for reading embedded markdown/code. It's just one HTML file — open DevTools to inspect, Ctrl+S to save your own copy and modify however you like. https://tasks.hamsterbase.com/jsonl-viewer/Apr 9, 2026 4:11 AM
bing.com
Forbes 250: The Greatest Self-Made Americanshttp://www.bing.com/news/apiclick.aspx?ref=FexRss&aid=&tid=6a89d049c21f4412805ebe28b58caee2&url=https%3A%2F%2Fwww.forbes.com%2Fsites%2Falexknapp%2F2026%2F04%2F09%2Fforbes-self-made-250-the-greatest-living-self-made-americans%2F&c=13806491560566846467&mkt=en-usGrit. Hustle. Resilience. The American Dream is built on the audacious belief that anyone can make it to the top. Every elementary school kid is imbued with the belief that anyone can become president ...Apr 9, 2026 3:29 AM
bing.com
From prompts to partnership: How LTM’s Rajesh Kumar collaborates with Microsoft 365 Copilothttp://www.bing.com/news/apiclick.aspx?ref=FexRss&aid=&tid=6a8cc1e1322c448f9c2e58448d9361d1&url=https%3A%2F%2Fnews.microsoft.com%2Fsource%2Fasia%2Ffeatures%2Ffrom-prompts-to-partnership-how-ltms-rajesh-kumar-collaborates-with-microsoft-365-copilot%2F&c=1546774424808218929&mkt=en-usRajesh Kumar’s prompts on Microsoft 365 Copilot have grown longer since he began using the AI assistant two years ago. Then, he was posing one-liners like: “What is this email about?” and “Write a ...Apr 8, 2026 8:29 PM
