Skip to main content

Posts

Showing posts from January, 2026

New top story on Hacker News: Show HN: LemonSlice – Give your voice agents a face

Show HN: LemonSlice – Give your voice agents a face 11 by lcolucci | 4 comments on Hacker News. Hey HN, we're the co-founders of LemonSlice ( https://lemonslice.com ). We train interactive avatar video models. Our API lets you upload a photo and immediately jump into a FaceTime-style call with that character. Here's a demo: https://ift.tt/p76HPLb Chatbots are everywhere. Voice AI has recently taken off. But we believe video avatars will be the most common form factor for conversational AI. Most people would rather watch something than read it. The problem is that generating video in real-time is hard, and overcoming the uncanny valley is even harder. We haven’t broken the uncanny valley yet. Nobody has. But we’re getting close and our photorealistic avatars are currently best-in-class (judge for yourself: https://ift.tt/Mxw3jNW ). Plus, we're the only avatar model that can do animals and heavily stylized cartoons. Try it: https://ift.tt/0QCiRgq . Warning! Talking to this...

New top story on Hacker News: Show HN: Synesthesia, make noise music with a colorpicker

Show HN: Synesthesia, make noise music with a colorpicker 3 by tevans3 | 1 comments on Hacker News. This is a (silly, little) app which lets you make noise music using a color picker as an instrument. When you click on a specific point in the color picker, a bit of JavaScript maps the binary representation of the clicked-on color's hex-code to a "chord" in the 24 tone-equal-temperament scale. That chord is then played back using a throttled audio generation method which was implemented via Tone.js. NOTE! Turn the volume way down before using the site. It is noise music. :)

New top story on Hacker News: Show HN: See the carbon impact of your cloud as you code

Show HN: See the carbon impact of your cloud as you code 24 by hkh | 2 comments on Hacker News. Hey folks, I’m Hassan, one of the co-founders of Infracost ( https://ift.tt/aCHJIXi ). Infracost helps engineers see and reduce the cloud cost of each infrastructure change before they merge their code. The way Infracost works is we gather pricing data from Amazon Web Services, Microsoft Azure and Google Cloud. What we call a ‘Pricing Service’, which now holds around 9 million live price points (!!). Then we map these prices to infrastructure code. Once the mapping is done, it enables us to show the cost impact of a code change before it is merged, directly in GitHub, GitLab etc. Kind of like a checkout-screen for cloud infrastructure. We’ve been building since 2020 (we were part of YC W21 batch), and iterating on the product, building out a team etc. However, back in 2020 one of our users asked if we can also show the carbon impact alongside costs. It has been itching my brain since then...

New top story on Hacker News: Bypassing Gemma and Qwen safety with raw strings

Bypassing Gemma and Qwen safety with raw strings 16 by teendifferent | 0 comments on Hacker News. OP here. I spent the weekend red-teaming small-scale open weights models (Qwen2.5-1.5B, Qwen3-1.7B, Gemma-3-1b-it, and SmolLM2-1.7B). I found a consistent vulnerability across all of them: Safety alignment relies almost entirely on the presence of the chat template. When I stripped the <|im_start|> / instruction tokens and passed raw strings: Gemma-3 refusal rates dropped from 100% → 60%. Qwen3 refusal rates dropped from 80% → 40%. SmolLM2 showed 0% refusal (pure obedience). Qualitative failures were stark: models that previously refused to generate explosives tutorials or explicit fiction immediately complied when the "Assistant" persona wasn't triggered by the template. It seems we are treating client-side string formatting as a load-bearing safety wall. Full logs, the apply_chat_template ablation code, and heatmaps are in the post. Read the full analysis: https://...

New top story on Hacker News: Show HN: What if your menu bar was a keyboard-controlled command center?

Show HN: What if your menu bar was a keyboard-controlled command center? 20 by pugdogdev | 4 comments on Hacker News. Hey Hacker News The ones that know me here know that I am a productivity geek. After DockFlow to manage my Dock and ExtraDock, which gives me more space to manage my apps and files, I decided to tackle the macOS big boss: the menu bar. I spend ~40% of my day context-switching between apps — Zoom meetings, Slack channels, Code projects, and Figma designs. My macOS menu bar has too many useless icons I almost never use. So I thought to myself, how can I use this area to improve my workflows? Most solutions (Bartender, Ice) require screen recording permissions, and did not really solve my issues. I wanted custom menus in the apps, not the ones that the developers decided for me. After a few iterations and exploring different solutions, ExtraBar was created. Instead of just hiding icons, what if the menu bar became a keyboard-controlled command center that has the action...

New top story on Hacker News: Show HN: 1Code – Open-source Cursor-like UI for Claude Code

Show HN: 1Code – Open-source Cursor-like UI for Claude Code 17 by Bunas | 5 comments on Hacker News. Hi, we're Sergey and Serafim. We've been building dev tools at 21st.dev and recently open-sourced 1Code ( https://1code.dev ), a local UI for Claude Code. Here's a video of the product: https://www.youtube.com/watch?v=Sgk9Z-nAjC0 Claude Code has been our go-to for 4 months. When Opus 4.5 dropped, parallel agents stopped needing so much babysitting. We started trusting it with more: building features end to end, adding tests, refactors. Stuff you'd normally hand off to a developer. We started running 3-4 at once. Then the CLI became annoying: too many terminals, hard to track what's where, diffs scattered everywhere. So we built 1Code.dev, an app to run your Claude Code agents in parallel that works on Mac and Web. On Mac: run locally, with or without worktrees. On Web: run in remote sandboxes with live previews of your app, mobile included, so you can check on age...

New top story on Hacker News: Show HN: OpenWork – an open-source alternative to Claude Cowork

Show HN: OpenWork – an open-source alternative to Claude Cowork 7 by ben_talent | 1 comments on Hacker News. hi hn, i built openwork, an open-source, local-first system inspired by claude cowork. it’s a native desktop app that runs on top of opencode (opencode.ai). it’s basically an alternative gui for opencode, which (at least until now) has been more focused on technical folks. the original seed for openwork was simple: i have a home server, and i wanted my wife and i to be able to run privileged workflows. things like controlling home assistant, or deploying custom web apps (e.g. our customs recipe app recipes.benjaminshafii.com), legal torrents, without living in a terminal. our initial setup was running the opencode web server directly and sharing credentials to it. that worked, but i found the web ui unreliable and very unfriendly for non-technical users. the goal with openwork is to bring the kind of workflows i’m used to running in the cli into a gui, while keeping a very de...

New top story on Hacker News: Show HN: FastScheduler – Decorator-first Python task scheduler, async support

Show HN: FastScheduler – Decorator-first Python task scheduler, async support 8 by michielme | 2 comments on Hacker News. Hi! I've built this because I kept reaching for Celery for simple scheduled tasks and it felt like overkill. I just needed "run this function every hour" or "daily at 9am", not distributed workers. So it's decorators for scheduling (@scheduler.every(5).minutes, @scheduler.daily.at("09:00")), state saves to JSON so jobs survive restarts, and there's an optional FastAPI dashboard if you want to see what's running. No Redis, no message broker, runs in-process with your app. Trade-off is it's single process only — if you need distributed workers, stick with Celery.

New top story on Hacker News: Show HN: Hover – IDE style hover documentation on any webpage

Show HN: Hover – IDE style hover documentation on any webpage 9 by sampsonj | 1 comments on Hacker News. I thought it would be interesting to have ID style hover docs outside the IDE. Hover is a Chrome extension that gives you IDE style hover tooltips on any webpage: documentation sites, ChatGPT, Claude, etc. How it works: - When a code block comes into view, the extension detects tokens and sends the code to an LLM (via OpenRouter or custom endpoint) - The LLM generates documentation for tokens worth documenting, which gets cached - On hover, the cached documentation is displayed instantly A few things I wanted to get right: - Website permissions are granular and use Chrome's permission system, so the extension only runs where you allow it - Custom endpoints let you skip OpenRouter entirely – if you're at a company with its own infra, you can point it at AWS Bedrock, Google AI Studio, or whatever you have Built with TypeScript, Vite, and the Chrome extension APIs. Coming to...

New top story on Hacker News: Show HN: FP-pack – Functional pipelines in TypeScript without monads

Show HN: FP-pack – Functional pipelines in TypeScript without monads 3 by superlucky84 | 1 comments on Hacker News. Hi HN, I built fp-pack, a small TypeScript functional utility library focused on pipe-first composition. The goal is to keep pipelines simple and readable, while still supporting early exits and side effects — without introducing monads like Option or Either. Most code uses plain pipe/pipeAsync. For the few cases that need early termination, fp-pack provides a SideEffect-based pipeline that short-circuits safely. I also wrote an “AI agent skills” document to help LLMs generate consistent fp-pack-style code. Feedback, criticism, or questions are very welcome.

New top story on Hacker News: Show HN: Feature detection exploration in Lidar DEMs via differential decomp

Show HN: Feature detection exploration in Lidar DEMs via differential decomp 4 by DarkForestery | 0 comments on Hacker News. I'm not a geospatial expert — I work in AI/ML. This started when I was exploring LiDAR data with agentic assitince and noticed that different signal decomposition methods revealed different terrain features. The core idea: if you systematically combine decomposition methods (Gaussian, bilateral, wavelet, morphological, etc.) with different upsampling techniques, each combination has characteristic "failure modes" that selectively preserve or eliminate certain features. The differences between outputs become feature-specific filters. The framework tests 25 decomposition × 19 upsampling methods across parameter ranges — about 40,000 combinations total. The visualization grid makes it easy to compare which methods work for what. Built in Cursor with Opus 4.5, NumPy, SciPy, scikit-image, PyWavelets, and OpenCV. Apache 2.0 licensed. I'd appreciate...