Show HN: Marginalia Integrated Site Explorer https://bit.ly/3RtAowh

Show HN: Marginalia Integrated Site Explorer Been working on this redesign for a while, but several pieces fell in place last few days. Can't stop playing with this doodad :D https://bit.ly/3Rq1m89 December 6, 2023 at 01:25AM

Show HN: Frogtab – Private, peaceful task management https://bit.ly/4a5RJTk

Show HN: Frogtab – Private, peaceful task management Hi HN! I'm excited to share the task manager that I've been building and relying on for the past 9 months I started developing Frogtab after a particularly busy period at work, where a large number of ad-hoc tasks would need my attention each day. What I Initially tried was a single (long!) checklist of outstanding tasks, but I found that too overwhelming to look at all day long. Then I tried scheduling tasks based on when I expected to complete them, but that also required too much cognitive load - I only had the capacity to bucket tasks into "today" and "not today" And so Frogtab was born. Originally as a Google doc, then later as a rudimentary web app with some simple automation. The automation was there to nudge me towards good habits, e.g., grouping all my outstanding tasks at the beginning of each day and requiring me to consciously choose which tasks to tackle that day. I subsequently refined the system a little, but the core mechanic has been serving me well ever since! More than the benefits I've gotten from using Frogtab at work, it's been really fun to develop Frogtab into a form that is usable - dare I say enjoyable - by other people too. Thanks to feedback from the awesome community at Bear (bearblog.dev) and also from sharing an early version here on HN, I've been able to make a whole host of improvements to Frogtab Highlights include: - Support for a weekly routine, aka recurring tasks - Data export and import, including auto-backup in supported browsers - Ability to receive tasks sent from any device - Dark theme - Keyboard shortcuts - Full docs at frogtab.com/help As I've been developing Frogtab, one of my primary concerns has been data privacy. Frogtab stores data in your bowser and doesn't require an account. Also, Frogtab does not use cookies, collect PII, serve ads, or track you in any way. It's entirely free to use and fully open source - the code is at github.com/dwilding/frogtab If you try Frogtab, I hope that you have a great experience! I would love to hear any feedback/suggestions, thank you :) https://bit.ly/4a4wcKV December 5, 2023 at 11:57PM

Show HN: SuperDuperDB – Open-source framework for integrating AI with databases https://bit.ly/47Gg8ND

Show HN: SuperDuperDB – Open-source framework for integrating AI with databases Hi everyone, I’m Timo, one of the creators of SuperDuperDB ! Today we are officially launching SuperDuperDB, an open-source framework for integrating AI directly with major databases, including streaming inference, scalable model training, and vector search, with release of v0.1. on GitHub and on ProductHunt. SuperDuperDB is not a database. It transforms your favorite database into an AI development and deployment environment (𝘮𝘢𝘬𝘪𝘯𝘨 𝘪𝘵 𝘴𝘶𝘱𝘦𝘳-𝘥𝘶𝘱𝘦𝘳). SuperDuperDB eliminates complex MLOps pipelines, specialized vector databases - and the need to migrate and duplicate data by integrating AI at the data's source, directly on top of your existing data infrastructure. This massively simplifies building and managing AI applications. SuperDuperDB provides a simple Python interface, but allows experts to drill down to any level of implementation detail such as models weights or training details. Today’s release comes with the full integration of major SQL databases as well as further MongoDB support: PostgreSQL, MySQL, SQLite, DuckDB, Snowflake, BigQuery, ClickHouse, DataFusion, Druid, Impala, MSSQL, Oracle, pandas, Polars, PySpark, and Trino. Currently Supported AI: Any model from PyTorch, Sklearn, HuggingFace as well as AI APIs such as OpenAI, Anthrophic, Cohere. A few useful links: - Our website: https://bit.ly/3NdqdcT - Getting started docs: https://bit.ly/41fANGc - Our repo on Github: https://bit.ly/3R4Yyfr Check the uses-cases that we have already implemented here https://bit.ly/3Nf4RvP as well as apps built by the community here https://bit.ly/41ajBBZ and try all of them with Jupyter your browser https://bit.ly/41bjLsQ For more information about SuperDuperDB and why we believe it is much needed, read the blog post https://bit.ly/3uI9qbw... We are keen to hear your feedback! All the best, Timo https://bit.ly/3R4Yyfr December 5, 2023 at 02:20PM

Show HN: Figma to Lottie Animation via Lottielab https://bit.ly/3uEuteW

Show HN: Figma to Lottie Animation via Lottielab Hello HN! We're super proud to announce the launch of Lottielab's official Figma plugin. The plugin allows nearly 1:1 import of your Figma designs into Lottielab, an easy to use animation tool (still in beta---we launched 3 months ago!). From there, you can use Lottielab's animation powers to animate your designs intuitively and quickly. Then, you can export the animation as a Lottie JSON, or as an animated GIF or MP4. This plugin is the culmination of 3 months of hard work, and we hope you'll like it! Me and a few members of the team will be here to answer all of your questions. https://bit.ly/3t5WaNy December 5, 2023 at 10:20AM

Show HN: Get NYC Bikeshare Availability via SMS so I can give up my smartphone https://bit.ly/414rVDf

Show HN: Get NYC Bikeshare Availability via SMS so I can give up my smartphone https://bit.ly/414S4lj December 5, 2023 at 05:23AM

Show HN: TSDiagram – Diagrams as Code with TypeScript https://bit.ly/3t4Olrf

Show HN: TSDiagram – Diagrams as Code with TypeScript Made an online tool for creating diagrams with TypeScript. I quite like it for thinking about how to shape my data, even if I end up using a different language in the end. https://bit.ly/4a97hWr December 5, 2023 at 03:01AM

Show HN: Zero to Hero – A Haskell Puzzle Game https://bit.ly/47S2XsS

Show HN: Zero to Hero – A Haskell Puzzle Game Hi HN. I am a researcher from Monash University. For one of our projects, we made a Haskell puzzle game called Zero to Hero. I invite you to try out the puzzles and share your thoughts (optional). The game contains 10 unique puzzles; each challenges you to implement a seemingly impossible function. You will seek help from a handful of strange-looking helper functions and your own wits. The game starts easy but quickly escalates into madness. The game requires you to be comfortable with basic Haskell syntax to progress. But no other prior knowledge is needed. You can choose to participate in our study or play for fun without any data collection at all. No stress. More details are explained on the landing page. https://bit.ly/3NgvyA7 Go ahead and enjoy the puzzles! https://bit.ly/482rdIC December 4, 2023 at 11:02AM

Show HN: HNSW-Based Search, in the Browser https://bit.ly/483TS0f

Show HN: HNSW-Based Search, in the Browser https://bit.ly/3uGnNwV December 3, 2023 at 10:36PM

Show HN: Caption – Generate social media posts from a given image using LLaVA https://bit.ly/47IJvPy

Show HN: Caption – Generate social media posts from a given image using LLaVA https://bit.ly/3NbcS4Q December 4, 2023 at 06:07AM

Show HN: A Who is Hiring app with AI filters – December update https://bit.ly/3uGGDUT

Show HN: A Who is Hiring app with AI filters – December update A cureated list of Who is Hiring with posts with smart filters generated with AI https://bit.ly/3OMWt7R December 4, 2023 at 02:21AM

Show HN: Simulate 3D Plants in the Browser https://bit.ly/3uHDi7X

Show HN: Simulate 3D Plants in the Browser https://bit.ly/3MXt8GC December 3, 2023 at 11:55AM

Show HN: Gobble Bot https://bit.ly/3NcR6xE

Show HN: Gobble Bot All your content digested into one text file You have a website, PDF or a Youtube video and want to make a GPT chatbot on ChatGPT. Simply use Gobble Bot to turn it all into one text file, ready to use for training. see demo video: https://bit.ly/3t4CUjl -> What is this? Gobble Bot is an all-in-1 scraper designed to convert all different kinds of content into one text file. -> Why do I need it? You need it if you want to create a GPT (a custom ChatGPT chatbot) but your content is a website, YouTube video or spans more than 20 files (the current upload limit for GPTs). You can crawl the whole internet and turn it into a chatbot, I guess. -> What data does it handle? Youtube videos with transciptions, public websites and files with extensions: .TXT, .MD, .XSL, .PDF, .DOC, .DOCX, .ODT, .OTT, .RTF, .HTML, .HTM, .ATOM, .RSS, .XML, .XLS, .XLSX, .XLSB, .XLSM, .XLTX, .CSV, .ODS, .OTS, .PPTX, .POTX, .ODP, .OTP, .ODG, .OTG https://bit.ly/3T8Rzo9 December 3, 2023 at 01:57PM

Show HN: Fluvio – Distributed stream processing system written in Rust and WASM https://bit.ly/47K9XIz

Show HN: Fluvio – Distributed stream processing system written in Rust and WASM https://bit.ly/46DRxI5 December 3, 2023 at 01:35AM

Show HN: Concurrent Agents Showcase in React https://bit.ly/3uD3PmN

Show HN: Concurrent Agents Showcase in React I wrote a small "game" to showcase concurrent agents (via a library) working in a React app. It demonstrates how a React program can be written with DOD-inspired pattern to make it easy to reason about, straightforward in re-render instruction semantic. Thus as the side effect the program is performant. You're invited to view the source to examine the neat design pattern underneath as well as the agent management library that powers the app. Caveat: it doesn't work well in mobile yet, so switch to desktop site mode. https://bit.ly/47FDD9B December 2, 2023 at 01:52AM

Show HN: Python decorator to log local variables with each line of code run https://bit.ly/4a3WnS0

Show HN: Python decorator to log local variables with each line of code run https://bit.ly/3t67YPB December 2, 2023 at 02:43AM

Show HN: 80% faster, 50% less memory, 0% loss of accuracy Llama finetuning https://bit.ly/46KgNMO

Show HN: 80% faster, 50% less memory, 0% loss of accuracy Llama finetuning Hi HN! I'm just sharing a project I've been working on during the LLM Efficiency Challenge - you can now finetune Llama with QLoRA 5x faster than Huggingface's original implementation on your own local GPU. Some highlights: 1. Manual autograd engine - hand derived backprop steps. 2. QLoRA / LoRA 80% faster, 50% less memory. 3. All kernels written in OpenAI's Triton language. 4. 0% loss in accuracy - no approximation methods - all exact. 5. No change of hardware necessary. Supports NVIDIA GPUs since 2018+. CUDA 7.5+. 6. Flash Attention support via Xformers. 7. Supports 4bit and 16bit LoRA finetuning. 8. Train Slim Orca fully locally in 260 hours from 1301 hours (5x faster). 9. Open source version trains 5x faster or you can check out Unsloth Pro and Max codepaths for 30x faster training! https://bit.ly/46G3tcj... has more info about Unsloth! Hopefully you can try it out! Wrote a blog post at https://bit.ly/3uNphFz if you want to learn more about our manual hand derived backprop or Triton kernels and stuff! Thanks once again! https://bit.ly/46VObR7 December 1, 2023 at 03:42PM

Show HN: Generate a video to animate stars of any GitHub repository https://bit.ly/47XTxvP

Show HN: Generate a video to animate stars of any GitHub repository I made this as a fun tiny project to experiment with Remotion [1] to generate videos. You can find the source code on GitHub [2]. [1] https://bit.ly/3sOfTRO [2] https://bit.ly/3T16Nf3 https://bit.ly/46FvMHK December 1, 2023 at 04:09PM

Show HN: Australian Acoustic Observatory Search https://bit.ly/3T2ibqT

Show HN: Australian Acoustic Observatory Search The Australian Acoustic Observatory ( https://bit.ly/3TnEwzF ) has 360 microphones across the continent, and over 2 million hours of audio. However, none of it is labeled: We want to make this enormous repository useful to researchers. We have found that researchers are often looking for 'hard' signals - specific call-types, birds with very little available training data, and so on. So we built an acoustic-similarity search tool, allowing researchers to provide an example of what they're looking for, which we then match against embeddings from the A2O dataset. Here's some fun examples! Laughing Kookaburra: < https://search.acousticobservatory.org/search/index.html?q=h... > Pacific Koel: < https://search.acousticobservatory.org/search/index.html?q=h... > Chiming Wedgebill: < https://search.acousticobservatory.org/search/index.html?q=h... > How it works, in a nutshell: We use audio source separation (< https://blog.research.google/2022/01/separating-birdsong-in-... >) to pull apart the A2O data, and then run an embedding model (< https://arxiv.org/abs/2307.06292 >) on each channel of the separated audio to produce a 'fingerprint' of the sound. All of this is put in a vector database with a link back to the original audio. When someone performs a search, we embed their audio, and then match against all of the embeddings in the vector database. Right now, about 1% of the A2O data is indexed (the first minute of every recording, evenly sampled across the day). We're looking to get initial feedback and will then continue to iterate and expand coverage. (Oh, and here's a bit of further reading: https://bit.ly/3T3ivWs... ) https://bit.ly/3Nab4ZY December 1, 2023 at 06:57AM

Show HN: Why an Open Source JSON-LD Database https://bit.ly/4a3Egf4

Show HN: Why an Open Source JSON-LD Database We are pleased to launch a public preview of an open source JSON-LD database which combines the simplicity of a document db with the analytical power of semantic graph. We hope it is approachable for most any developer while having the capabilities, when needed, for data interoperability, embedded cell-level security (ReBAC style), rich shape/schema restrictions, data provenance (time travel), and fact inferencing. Why did our team at Fluree build this? Data is increasingly critical for great decision making, AI, and more. The way we typically store and manage data, mainly as an application state store, makes it challenging to leverage data strategically. The popular approaches to leveraging data outside of an app are difficult to scale and have limited effectiveness (e.g. copy/paste of stale, un-permissioned data into a lake/ warehouse, custom python scripts for feature enrichment, ETL tools). We believe a shift in view for data management from application-centric to data-centric can address the root issue and reduce the needs of the additional layers. If all data was semantically interoperable (composable), available in a network where data was current, and data has more native capabilities around security and inferencing, we can address the foundational challenges – and is what we set out to build. Semantically interoperable data is not a new thing, semantic web/knowledge graph users enjoy many great database products, but the focus has been analytics and inferencing, and often run on a laptop or isolated server for an analyst – plus, they are considered to have a steep learning curve. Having a platform that is optimized for operational data and minimizing the learning curve, while retaining the power, are the largest obstacles we had to overcome. Beyond semantics, data that can be dynamically shared and collaborated on in a network is where we think real data leverage emerges, but data owners need programmatic policy control to agree to participate. Controlling time also becomes critical, as reproducing results across a federated network is impossible without it. We built programmability into the data layer to address all of these needs, and shift our thinking from a “database” being a mutable thing, to an immutable thing - where every update to a database creates a new immutable database, much like every “git” commit creates a new immutable version of your source code. We are still early on our journey, but would love any feedback, thoughts, help (we are open source). We hope our vision aligns with many and the product resonates. Feel free to register for an account on our cloud version, and reserve your data namespace in what we’d like to become a valuable and trusted data network. https://bit.ly/3uGVoXJ November 30, 2023 at 10:02PM

Show HN: Qdorks.com – Advanced Google search query composer https://bit.ly/47WfjQD

Show HN: Qdorks.com – Advanced Google search query composer Hi HN, qdorks.com is an advanced Google search query composer I have been working on lately. It makes it super easy to write complex Google search queries and save them for later usage. Main features: * Query composer with complex grouping, logical and exclusion operators per clause. * Share your queries with others so they can easily copy and modify the query. * No need to register; only when you want to save the queries. * AI support for PRO users. * More to come... I am pretty happy with how it works and would love your feedback! Happy to answer any questions. Thank you! https://bit.ly/47ZFyWz November 30, 2023 at 03:50PM