34,039 results for av

lusob.github.io
Show HN: Neural window manager, neural network moving windows from mouse actionshttps://lusob.github.io/neural-os/I'd been mulling over this crazy idea for a while. Can programs be generated? Inspired by recent advances in world models, I wondered if we could do away with source code and generate pixels directly and interactively.As an experiment to answer this, I set out to create a neural window manager, training a neural network to predict what the screen would look like next.Basically, the idea was to generate the next frame based on the last two frames and the mouse position. That's it: moving windows without programming an event system, just a simple convolutional neural network guessing pixels.To implement the experiment, I used Pygame to simulate a turquoise desktop background, a gray window with a navy blue title bar, a white cursor, and four colors in total. Then, a bot randomly dragged the window, and I recorded everything, processing the frames as color index matrices (not RGB, to avoid complications) and the mouse delta (dx, dy, click) that caused each transition. 8000 frames, a few m
May 13, 2026 5:46 PM
news.ycombinator.com
Thoughts on Claude Code 2.1.139 Agent View and Background Sessionshttps://news.ycombinator.com/item?id=48124057Spent half a day trying Claude Code 2.1.139’s new Agent View and background sessions — useful, but still has quite a few rough edges.The first item in the 2.1.139 changelog released on 2026-05-11 was Added agent view (Research Preview). Interestingly, they even included a dedicated doc link in the changelog this time: https://code.claude.com/docs/en/agent-viewNot sure how many people here have tried it yet. I spent a few hours with it today and found both some really nice ideas and some frustrating issues. Curious to hear other people’s experiences too.1. What is Agent View?Agent View can launch parallel Claude sessions directly, and each session automatically uses a separate git worktree for isolation, so multiple sessions can run in parallel without conflicts.The background sessions are managed by an independent supervisor process. They continue running even if the terminal or Agent View itself is closed. Closing the terminal does NOT stop the background sessions.Checking processes w
May 13, 2026 4:25 PM
news.ycombinator.com
Some Business Ideashttps://news.ycombinator.com/item?id=48123663Infrastructure side:1. A service connector that can be attached to any AI Assistant to authenticate users and connect services easily, similar to MCP but more universal. 2. CDN service connectors that can store assets and retrieve them very fast from anywhere in the world. 3. Database-as-a-service connectors that are highly available and very simple to connect or disconnect from any provider without managing infrastructure. 4. Object storage service connectors that can be attached to personal assistants or AI agents, highly scalable and available, with flexible models like pay-per-use or pay-per-storage. 5. Vector DB service connectors for storing RAG data, designed to be highly available and distributed close to clients for low latency and fast AI responses.You can use AWS or other cloud providers behind these services, but in the future people will not want to manage infrastructure directly. They will need services that are easy to connect and disconnect, simple to scale, and fully A
May 13, 2026 4:00 PM
github.com
Show HN: Rotunda - A browser built for agents with simulated typinghttps://github.com/monkeysee-ai/rotundaHi HN! Pierce here.Rotunda is a firefox fork primarily intended for agent use, which I’ve been hacking on nights/weekends.There was a [lengthy](https://news.ycombinator.com/item?id=48024859) discussion last week on how expensive computer use models are. The cost is going to drop eventually, but I think on some level it's still usually the wrong primitive. The web gives us access to beautiful structured formats, plaintext, etc... why throw that away if we don't have to?I realized at some point that for 99% of automations I just want agents to be able to control my Chrome instance. But that’s easier said that done: CDP (the Chrome automation protocol) leaks a ton of state about being programmatically controlled, either by toggling window attributes or by running `page.evaluate()` commands right in the page context. Plus if you look at an automation running it's pretty obvious what happens: the mouse jumps around, fields are filled instantly, etc.Rotunda tries to fix this. Its standout fe
May 13, 2026 1:44 PM
github.com
Show HN: An open source tool for generating macOS app icons with AIhttps://github.com/TeamDev-IP/MoBrowser-App-Icon-MakerI’m Vladimir. I’m a software engineer. In our company we often develop desktop apps for internal needs and prototyping. And every time I bump into the same problem: how do I make an icon for the macOS app I have just built?I could use the existing icon generators, but they are basically just image converters. You upload an existing image, and the tool generates the required icon sizes and formats from it.But I don’t have an image, and I’m not a designer. Asking designers to create an icon is not always an option.I wanted something that could help me actually create an icon. Something where I can describe an idea, iterate on it over several rounds, experiment with materials, lighting, composition, and gradually arrive at an icon that feels like a real native macOS app icon.Since I’m an engineer, I built a small tool that allows generating a macOS app icon using AI. It’s completely free and open source, so other engineers building desktop apps for macOS can use it too.The app lets you ge
May 13, 2026 1:03 PM
bing.com
The World’s Most Beautiful Countries in 2026, Rankedhttp://www.bing.com/news/apiclick.aspx?ref=FexRss&aid=&tid=6a8cde78949645a8b95cf9048ec92114&url=https%3A%2F%2Fwww.mentalfloss.com%2Fgeography%2Fwanderlust%2Fmost-beautiful-countries-2026-ranked&c=128710348743472094&mkt=en-usTrying to rank the most beautiful countries in the world is a bit like being asked to pick the best pasta shape, the fluffiest cloud, or the prettiest flower: everyone has a favorite, and no one is ...
May 13, 2026 12:59 PM
bing.com
The World’s Most Beautiful Countries in 2026, Rankedhttp://www.bing.com/news/apiclick.aspx?ref=FexRss&aid=&tid=6a8d97a4a69043239bf21a8be58180b8&url=https%3A%2F%2Fwww.mentalfloss.com%2Fgeography%2Fwanderlust%2Fmost-beautiful-countries-2026-ranked&c=128710348743472094&mkt=en-usTrying to rank the most beautiful countries in the world is a bit like being asked to pick the best pasta shape, the fluffiest cloud, or the prettiest flower: everyone has a favorite, and no one is ...
May 13, 2026 12:59 PM
news.ycombinator.com
What is your AI agent constantly doing wrong?https://news.ycombinator.com/item?id=48120148My colleague and I have found AI is consistently falling short on the following things. I'd be interested to know whether this is a prompting issue or a model issue (we use claude 4.7):- comments being too verbose, and making reference to the previous code. Also just using phrases that no human would ever use (e.g. saying 'gate', 'belt-and-braces' a lot)- over-using default arguments in method definitions- failing to see the bigger picture in terms of DRYness, code organisation- way too much defensive coding like handling edge cases that would never happen, adding fallbacks for everything
May 13, 2026 10:36 AM
mccormick.cx
Show HN: Embedded Spritesheet PNG Standardhttps://mccormick.cx/games/embedded-spritesheet-png-spec/Hi HN!2d spritesheets for game development are often distributed as a PNG image, sometimes with a separate JSON or XML metadata document describing the layout, including any animation sequences within the sheet. Sometimes there is no metadata document and you have to work it out and manually write custom code to chop up the spritesheet and reference the sprites you want.I think it would be good if it became standard practice to distribute spritesheet PNGs with the metadata inline inside a PNG "chunk" so the whole thing is self-contained. That way artists can just distribute one file, and tools and libraries can read the metadata directly from it. That's what this spec/standard is about.The linked page proposes a standard, describes the technique's details, and provides an online tool for inlining metadata into the sheet. Hope it's useful!
May 13, 2026 9:04 AM
esa.int
Waterworn chaos on Marshttps://www.esa.int/Science_Exploration/Space_Science/Mars_Express/Waterworn_chaos_on_MarsThis month, ESA’s Mars Express takes us to Shalbatana Vallis: a fascinating martian valley surrounded by signs of water, lava, craters and chaos.
May 13, 2026 9:00 AM
science.nasa.gov
America’s Emerald Islehttps://science.nasa.gov/earth/earth-observatory/americas-emerald-isle/Beaver Island is one in a string of verdant and scenic jewels in a northern Lake Michigan archipelago.
May 13, 2026 4:00 AM
display.dev
Show HN: Display.dev, agent-native way to publish HTML or MD behind company authhttps://display.dev/Last week's post from Thariq "The Unreasonable Effectiveness of HTML" went viral in X. He made a good point about HTML being the more expressive and human friendly alternative to markdown files. I completely buy in where he's coming from, I've been asking Claude Code to create me HTML files to better understand what we're planning and building together.What Thariq didn't address was the part of sharing and collaborating on these HTML artifacts (he proposed uploading to S3). While I've been heavy user of HTML artifacts to share quarterly plans, implementation specs, technical guides and more, it has been a challenge of sharing them with my teammates securely. Not all people in the company have GitHub access to see the markdown file. Slack and Google Drive don't render HTML.There's been no good solution today on the market, until now, which gives you tools for your agent to publish the HTML and MD files, gate the artifacts behind company SSO or OTP and collaborate with your teammates on
May 12, 2026 11:36 PM
gifpile.com
Show HN: GIF Pile. a site to make piles of GIFshttps://gifpile.com/I'm quite fond of obnoxious looking gifs in a post-ironic way as a manner of shitposting and or injecting humor into a chat. The issue with this however is that, for no real good reason at all, the simple usecase of "Have image/gif background, bombard with garbage" had no real good tooling.There's gif editors out there, EZgif my beloved is probably my most used non-search-indexing-slash-social-media-site, but they're kinda clunky for my specific usecase of making digital eye-sandpaper bombastic garbage. Other options are bleak and gave me the mark of the beast via shitty watermarks. I just wanted a pile of gifs on top of each other, and thus far the "easiest" way was to bust open a video editor, muck around with it, mess up exporting as a gif directly, get mad, export it as a 4 second mp4, and then use ffmpeg to get it working. is this probably moronic? yes. am I likely to have missed a decent tool? yes. Did I give up looking after sending 4 dollars to some Indian guy for "No watermark
May 12, 2026 9:11 PM
candle.fi
Show HN: Fast, reliable MCP for LinkedIn, Uber, Venmo (r/w)https://candle.fiMy friend and I have built a tool that can replicate mobile app API traffic precisely. We then wrap the app’s functionality in a standardized API and route traffic through a local proxy running on your computer or phone, so your traffic can’t be fingerprinted.We’re building an open platform around these connections, starting with a macOS app y’all can try for free. It lets you securely log in to your accounts and then exposes an MCP with one-click install for Claude. We’re most excited about enabling “write” actions, for example asking Claude to:- automate your biz ops on LinkedIn - schedule a ride to the airport with Uber - settle up with your friends on Venmo
May 12, 2026 8:56 PM
bing.com
Will You Look at Me - Where to Watchhttp://www.bing.com/news/apiclick.aspx?ref=FexRss&aid=&tid=6a8cce229ab74d3ab3e1f030728c0191&url=https%3A%2F%2Fwww.moviefone.com%2Fmovie%2Fwillyoulookatme%2FPwS8t6HH6q2MzHSf5IIOc2%2Fwhere-to-watch%2F&c=10458007576697313711&mkt=en-usStreaming platforms for Will You Look at Me haven’t been announced yet. Check back soon for updates on where you can watch it online.
May 12, 2026 4:24 PM
bing.com
Will You Look at Me - Where to Watchhttp://www.bing.com/news/apiclick.aspx?ref=FexRss&aid=&tid=6a8cb9f5eb794e03baa8f17bba9cc4ba&url=https%3A%2F%2Fwww.moviefone.com%2Fmovie%2Fwillyoulookatme%2FPwS8t6HH6q2MzHSf5IIOc2%2Fwhere-to-watch%2F&c=10458007576697313711&mkt=en-usStreaming platforms for Will You Look at Me haven’t been announced yet. Check back soon for updates on where you can watch it online.
May 12, 2026 4:24 PM
bing.com
Will You Look at Me - Where to Watchhttp://www.bing.com/news/apiclick.aspx?ref=FexRss&aid=&tid=6a8cb221553547a6b513ab484a324e3f&url=https%3A%2F%2Fwww.moviefone.com%2Fmovie%2Fwillyoulookatme%2FPwS8t6HH6q2MzHSf5IIOc2%2Fwhere-to-watch%2F&c=10458007576697313711&mkt=en-usStreaming platforms for Will You Look at Me haven’t been announced yet. Check back soon for updates on where you can watch it online.
May 12, 2026 4:24 PM
voker.ai
Launch HN: Voker (YC S24) – Analytics for AI Agentshttps://voker.aiHey HN, we're Alex and Tyler, co-founders of Voker.ai (https://voker.ai/), an agent analytics platform for AI product teams. Voker gives full visibility into what users are asking of your agents, and whether your agents are delivering, without having to dig through logs. Our main product is a lightweight SDK that is LLM stack agnostic and purpose-built for agent products. (https://app.voker.ai/docs)Agent Engineers and AI product teams don’t have the right level of visibility into agent performance in production, which results in bad user experiences, churn, and hundreds of hours wasted with spot checks to find and debug issues with agent configurations.Demo: https://www.tella.tv/video/vid_cmoukcsk1000i07jgb4j65u67/vie...We recently conducted a survey of YC Founders and 90%+ of respondents said that the only way they know if their Agents are failing users in production is by hearing complaints from customers. They push a prompt change hoping that it fixes the problem and doesn’t break s
May 12, 2026 3:45 PM
mosaicqr.com
Show HN: MosaicQR – QR codes that double as pixel art of any simple imagehttps://mosaicqr.comI wanted to make a QR code for a separate business (a software company for MEP engineers) where the QR code itself looked like the logo. I found several tools online, but none that produced the result I was looking for.I made a tool that powers MosaicQR.com to achieve the balance of aesthetics and functional scanning that I wanted. This website generates QR codes where the module layout looks like the inputted image. The free version auto-fetches the favicon from any URL you enter.Curious what people think.
May 12, 2026 2:21 PM