The best Hacker News stories from All from the past day
Latest posts:
Göttingen was one of the most productive centers of mathematics (2019)
Göttingen was one of the most productive centers of mathematics (2019)
Reverse-engineering my speakers' API to get reasonable volume control
Free DDNS with Cloudflare and a cronjob
Zen 5's 2-ahead branch predictor: how a 30 year old idea allows for new tricks
In the Beginning Was the Command Line (1999)
SQLite: 35% Faster Than the Filesystem
The New Internet
An experiment in UI density created with Svelte
An experiment in UI density created with Svelte
Launch HN: Undermind (YC S24) – AI agent for discovering scientific papers
Hey HN! We’re Josh and Tom from Undermind (<a href="https://www.undermind.ai/">https://www.undermind.ai/</a>). We’re building a search engine for complex scientific research. There's a demo video at <a href="https://www.loom.com/share/10067c49e4424b949a4b8c9fd8f3b12c?sid=5ee3cd96-2b05-4768-8856-cb1b83e40a70" rel="nofollow">https://www.loom.com/share/10067c49e4424b949a4b8c9fd8f3b12c?...</a>, as well as example search results on our homepage.<p>We’re both physicists, and one of our biggest frustrations during grad school was finding research — There were a lot of times when we had to sit down to scope out new ideas for a project and quickly become a deep expert, or we had to find solutions to really complex technical problems, but the only way to do that was manually dig through papers on Google Scholar for hours. It was very tedious, to the point where we would often just skip the careful research and hope for the best. Sometimes you’d get burned a few months later because someone already solved the problem you thought was novel and important, or you’d waste your time inventing/building a solution for something when one already existed.<p>The problem was there’s just no easy way to figure out what others have done in research, and load it into your brain. It’s one of the biggest bottlenecks for doing truly good, important research.<p>We wanted to fix that. LLMs clearly help, but are mostly limited to general knowledge. Instead, we needed something that would pull in research papers, and give you exactly what you need to know, even for very complex ideas and topics. We realized the way to do this is to mimic the research strategies we already know work, because we do them ourselves, and so we built an agent-like LLM pipeline to carefully search in a way that mimics human research strategies.<p>Our search system works a bit differently from casual search engines. First, we have you chat back and forth with an LLM to make sure we actually understand your really complex research goals up front, like you’re talking to a colleague. Then the system carefully searches for you for ~3 minutes. At a high level, it does something similar to tree search, following citation rabbit holes and adapting based on what it discovers to look for more content over multiple iterations (the same way you would if you decided to spend a few hours). The 3 minute delay is annoying, but we’re optimizing for quality of results rather than latency right now. At the end there’s a report.<p>We’re trying to achieve two things with this careful, systematic agent-like discovery process:<p>1. We want to be very accurate, and only recommend very specific results if you ask for a specific topic. To do this, we carefully read and evaluate content from papers with the highest quality LLMs (we’re just reading abstracts and citations for now, because they’re more widely accessible - but also working on adding full texts).<p>2. We want to find everything relevant to your search, because in research it’s crucial to know if something exists or not. The key to being exhaustive is the adaptive algorithms we’ve developed (following citations, changing strategy based on what we find, etc). However, one cool feature of the automated pipeline is we can track the discovery process as the search proceeds. Early on, we find many good results, and later on they get more sparse, until all the good leads are exhausted and we stop finding anything helpful. We can statistically model that process, and figure out when we’ve found everything (it actually has an interesting exponential saturation behavior, which you can read a bit more about in our whitepaper (<a href="https://www.undermind.ai/static/Undermind_whitepaper.pdf">https://www.undermind.ai/static/Undermind_whitepaper.pdf</a>), which we wrote for a previous prototype.)<p>You can try searching yourself here:
<a href="https://www.undermind.ai/query_app/promotion/">https://www.undermind.ai/query_app/promotion/</a>. This is a special HN link where, for today, we’ve dropped the signup gate for your first few searches. Usually we require login so you can save searches.<p>We’re excited to share this with you! We’d love to hear about your experiences searching, what’s clear or not, and any feedback. We’ll be here to answer any questions or comments.
Generating sudokus for fun and no profit
Defense of Lisp macros: The automotive field as a case in point
Jacek Karpińśki, the computer genius the communists couldn't stand (2017)
My Favorite Algorithm: Linear Time Median Finding (2018)
A Swiss town banned billboards. Zurich, Bern may soon follow
Show HN: Wat – Deep inspection of Python objects
Reverse Engineering for Everyone
Reverse Engineering for Everyone
Show HN: Haystack – an IDE for exploring and editing code on an infinite canvas
Hi HN, we’re building Haystack Editor (<a href="https://haystackeditor.com/" rel="nofollow">https://haystackeditor.com/</a>), a canvas-based IDE that automates the boring stuff (plumbing, refactoring, and finding code) so that you can focus on the exciting parts of software development! You can see a quick overview of Haystack at <a href="https://www.youtube.com/watch?v=c2uZnR5D_cc" rel="nofollow">https://www.youtube.com/watch?v=c2uZnR5D_cc</a>!<p>(It's currently only on Mac OS but we're working on Linux and Windows. Edit: just added a Linux download!)<p>Haystack was born out of our frustrations with working in large and mature codebases, specifically with navigating and editing functional flows (e.g. the code flow for adding an item to the Amazon shopping cart).<p>Oftentimes dealing with such flows would involve navigating a maze of files and functions, and making any edits would involve a lengthy process of doing corresponding downstream/upstream plumbing.<p>Haystack attempts to address this in the following ways:<p><pre><code> 1. It allows you to explore your codebase as a directed graph of functions, classes, etc on the canvas. We feel like this better fits how your mind understands your codebase and helps you find and alter functional flows more intuitively. We especially want to utilize this for pull request reviews!
2. It has a navigational copilot that makes edits across files or functions much easier. After you make some changes, Haystack will try to predict your next action and create functions/methods or refactor upstream/downstream code for you. Haystack will surface these speculative edits on the canvas in a way that you can easily dismiss or incorporate them, allowing you to make large changes with a few clicks or keystrokes.
3. Haystack will utilize natural language search so you don’t have to play “Where’s Waldo” to find a functional flow in your codebase. This is coming soon!
</code></pre>
We’re still pretty early in development and we really want to perfect the experience of navigating and editing code on a canvas. Any feedback would be much appreciated!<p>PSA: Since Haystack is a VS Code fork, you should be able to move your extensions and keyboard shortcuts. Please let us know if you have any issues with this!