prompt
stringlengths 16
15.6k
| completion
stringlengths 4
6
|
|---|---|
Title: Against the trend, just moved to SF from a mid size hub-of-nothing city of the East Coast. My GF got an internship in biotech and I work remote so here we are at least until January.<p>I'm self taught, have a technical high school degree in electronics and computers, and an AA that I meant to use to start a CS degree back in the east coast before we moved.<p>My work is a mix of App support with light software dev. I mostly support and update old apps written using ASP.NET (of which I know almost nothing), Nintex workflows that should've never been, make little scripts for ServiceNow virtual agent, and we have a series of low code apps that I also support, modify, etc. I replaced a guy who was also self taught and had no best-practices checks.<p>I'm solo in this endeavor since my product owner is not a developer and my other peer mostly deals with the red tape and administrative stuff like dealing with InfoSec's recurring paperwork and vendor licenses. Happy to have him, but I can't skill up in my field.<p>I'm now much closer to all the tech brain power of the country, and I'm trying to figure how to leverage my stay to improve my skills. I work 5AM to 1:30PM and so I have the whole evening open to feed my brain.<p>Are there any groups, meetups, schools, apprenticeships, or whatever else I could use to become a better dev? I know I could enroll in a CS degree and I do want that, but I already know of that answer.<p>I'd love to work on tools for people outside of the field that maybe boosts their productivity or makes their lives easier.
A few months ago I worked with my GF to make an app that helps her identify cyanobacteria automating a process they used to do in MS Word and it was super satisfying, but I know it can be much better if I had someone to work with.<p>Anyway, any idea will be very appreciated, even if you are in the same position I'm in and you'd like to meet to work on something.<p>Thanks!
Upvote: | 69 |
Title: Hey everyone,<p>I here is a small open-source project I've been working on lately.
I'd love to hear your thoughts and improvement ideas :)<p>GitHub: [github.com/Vincenius/workout-lol](<a href="https://github.com/Vincenius/workout-lol">https://github.com/Vincenius/workout-lol</a>)
Upvote: | 993 |
Title: I made:<p>First $100 = Freelancing on oDesk.<p>First $1,000 = Freelancing as a Web Developer for a client from referral.<p>How did you earn your first $100 & first $1,000 online?
Upvote: | 51 |
Title: My friend and I have been feeling frustrated at how inefficient it is to find information at work. There are so many tools (Slack, Confluence, GitHub, Jira, Google Drive, etc.) and they provide different (often not great) ways to find information. We thought maybe LLMs could help, so over the last couple months we've been spending a bit of time on the side to build Danswer.<p>It is an open source, self-hosted search tool that allows you to ask questions and get answers across common workspace apps AND your personal documents (via file upload / web scraping)! Full demo here: <a href="https://www.youtube.com/watch?v=geNzY1nbCnU&t=2s">https://www.youtube.com/watch?v=geNzY1nbCnU&t=2s</a>.<p>The code (<a href="https://github.com/danswer-ai/danswer">https://github.com/danswer-ai/danswer</a>) is open source and permissively licensed (MIT). If you want to try it out, you can set it up locally with just a couple of commands (more details in our docs - <a href="https://docs.danswer.dev/introduction" rel="nofollow noreferrer">https://docs.danswer.dev/introduction</a>). We hope that someone out there finds this useful<p>We’d love to hear from you in our Slack (<a href="https://join.slack.com/t/danswer/shared_invite/zt-1u3h3ke3b-VGh1idW19R8oiNRiKBYv2w" rel="nofollow noreferrer">https://join.slack.com/t/danswer/shared_invite/zt-1u3h3ke3b-...</a>) or Discord (<a href="https://discord.gg/TDJ59cGV2X" rel="nofollow noreferrer">https://discord.gg/TDJ59cGV2X</a>). Let us know what other features would be useful for you!
Upvote: | 189 |
Title: Mine currently are:<p><a href="https://www.pragmaticengineer.com/" rel="nofollow noreferrer">https://www.pragmaticengineer.com/</a><p><a href="https://bigtechdigest.substack.com/" rel="nofollow noreferrer">https://bigtechdigest.substack.com/</a><p><a href="https://leaddev.com/" rel="nofollow noreferrer">https://leaddev.com/</a><p><a href="https://stratechery.com/" rel="nofollow noreferrer">https://stratechery.com/</a><p><a href="https://betterdev.link/issues" rel="nofollow noreferrer">https://betterdev.link/issues</a><p><a href="https://hackernewsletter.com/" rel="nofollow noreferrer">https://hackernewsletter.com/</a>
Upvote: | 143 |
Title: I always enjoy reading posts about indie game development and looking at the source code of games other people have built. So for any game I release I naturally I want to do the same.<p><a href="https://github.com/Mknsri/HockeySlam">https://github.com/Mknsri/HockeySlam</a><p>Hockey Slam is a hockey shootout game on Android which I built almost from scratch. Some of the engine features include 3D graphics with PBR materials, a rudimentary physics system, replays, a multithreaded job system and a hot-reload DLL platform for developing on Windows.<p>I detailed the creation of the game in it's own post, but now that time has passed I am releasing the source code and assets under the MIT license. You can still find the making of here: <a href="https://hockeyslam.com/makingof" rel="nofollow noreferrer">https://hockeyslam.com/makingof</a><p>I'm happy to answer any questions you might have.
Upvote: | 128 |
Title: Long-simmering side project that is finally ready to see the light. HAMTs are a cool persistent data structure and implementing one has been a lot of fun. Beyond the code, there is likely some value in the extensive and largely complete implementation docs; basic benchmarks are linked in the README, too.<p>Kind of aiming to be "the libavl for HAMTs". That is obviously a high and aspirational bar but a distinct possibility if it stirs up a little interest and/or contribution.<p>Anyways, it's time for this to go out, collect feedback and maybe even some use outside of toy projects. Let me know how it goes.
Upvote: | 171 |
Title: I would like to know more about how Stripe assesses the risks around future refunds and chargeback.<p>I understand that there needs to be a safety margin and risk hedging but 50% for 9 months is going a bit far.<p>We are a small recreational property manager in the Netherlands. We switched from Mollie to Stripe earlier this year because our new PMS works exclusively with Stripe. Never had any problems with Mollie in the past.<p>We charge 20% commission for our services and pay out our owners on check-out date every month. Currently we have something over €600k tied up until April 2024. This is money from stays that have already been and we have to pay out to our owners.<p>What are your experiences with percentages withheld and the length of it?
How do these decisions come about, what data points do they use?
Are they able to analyze at all on a case-by-case basis?<p>Is there a change of getting this amount payed or are we f*cked?
Upvote: | 119 |
Title: I quit my job last year and worked mostly on this full SwiftUI rewrite of my Japanese reading app, Manabi Reader. The rewrite gave me the opportunity to expand to macOS (and without Catalyst so it feels extra native) and redo the data layer to be offline-first via Realm with iCloud sync.<p>The biggest differentiators compared with other language study apps are that it tracks every word/kanji you read so that you can see how much of a given webpage/article you're already familiar with and other features/analytics built on that foundation; it automatically builds a personal corpus of example sentences; and that it does all the Japanese tokenization/dictionary lookups locally on-device and in a flexible web browser-like UI with readability mode, to be respectful of your privacy and to work offline.<p>I've also added Anki integration. Tap a word, tap another button to save it to Anki with the original source material sentence and URL. I have a Manabi Flashcards app as well if you don't like Anki.<p>Packed with free features. See what percent of each article's vocabulary you're familiar with based on your reading history. Scan paragraphs of text with your camera to look up words. Japanese/English dict. Native Japanese web dicts. Look up kanji by drawing. Expanded JLPT levels. RSS. Web browser UI. Save links from other apps. Works offline. Readability mode. Tap words to look them up. Furigana depending on your familiarity with each word.<p>Future plans: besides more features (ePUB, YouTube, mpv player, WaniKani integration, more languages, etc), I’m also preparing the underlying SwiftUI web browser lib as open source and will launch it as a WebKit-based browser/reader option, which I’m excited to get out alongside other interesting recent entrants to the desktop and mobile browser market.
Upvote: | 105 |
Title: Laser is a turn based game similar to chess, with different piece movement rules, starting position, and win conditions. It's named after the <i>laser</i> piece, which can shoot diagonally through every piece on the board except for the <i>wall</i>, which blocks it. The detailed rules are on the website.<p>I made the website as a super minimal way to play online against friends. Nobody really knows about it so it might be hard to find a game, but maybe you will against someone on HN if people read this. It uses the lichess.org chessboard UI which is super pretty and allows you to draw on the board, make premoves, etc. The code is public if you are interested: github.com/melgrove/laser<p>If you are playing, good luck, and don't get lasered!
Upvote: | 218 |
Title: TLDR: Inbound email routes don't have SPF/DKIM/DMARC protection, meaning any inbound email sent to a webhook can be trivially spoofed / phished.
Mailgun describe inbound routes as:<p>"complex incoming messages are simplified and parsed into all of the data you need with Inbound Routes." [1]<p>Mailgun refers to DKIM/SPF/DMARC with:<p>"Mailgun leads the charge in email authentication by requiring DKIM and SPF records by default before you can begin sending." [2]<p>However, nowhere does it say that Mailgun does not provide SPF or DKIM validation checks on inbound emails. Confusingly, these headers appear only for a subset of inbound emails.<p>Attack scenario:<p>1) Pick any domain using mailgun for inbound email processing (trivially found via MX records).<p>2) Spoof an email from any address you want to impersonate (eg. [email protected]).<p>If that address is handled via an inbound route to a webhook (eg. CRM system / other), then it will appear as from the spoofed sender without any DKIM/SPF flags. This makes phishing via Inbound Routes trivial.<p>Requested fix:<p>Include SPF (X-Mailgun-Spf), DKIM (X-Mailgun-Dkim-Check-Result), and Spam Assassin headers (eg. DMARC_QUAR / DMARC_REJECT / DMARC_NONE / DMARC_MISSING) headers for ALL inbound routes to a webhook. The application can then decide what to do.<p>Mailgun Response:<p>"Our security team has confirmed that our routes act as an open relay and that this is not considered a security vulnerability with Mailgun."<p>Nowhere in the Mailgun documentation / sales pages are inbound routes described like this, and in fact, they pretend to have protection as headers are sometimes present!<p>Note A: Using throwaway to not give information away for our own exposed systems.<p>Note B: The spoofed email must not trigger above a certain spam assassin threshold (this is fairly trivial to do and openly testable by any attacker), otherwise it might get blocked by higher level Mailgun spam handling.<p>[1] - <a href="https://www.mailgun.com/products/send/inbound-routing/" rel="nofollow noreferrer">https://www.mailgun.com/products/send/inbound-routing/</a><p>[2] - <a href="https://www.mailgun.com/blog/deliverability/implement-dmarc/#chapter-5" rel="nofollow noreferrer">https://www.mailgun.com/blog/deliverability/implement-dmarc/...</a>
Upvote: | 88 |
Title: The way I think about Haskell programs is locked-in statements that tightly sit next to each other (i.e., function compositions) and nothing "leaks", as in no mutations are allowed (except in monads, like IO).<p>But I wonder if this is mostly a matter of taste. In small programs, the end result of a Haskell program is the same as a Python program. Is there a threshold after which Haskell's purely functional paradigm shines the most?
Upvote: | 56 |
Title: Clickvote takes the hassle of building your own reaction components around your content.<p>Showing real-time updates of likes, upvotes, and reviews between clients.<p>Learn about your members through deep analytics.<p>Deal with an unlimited amount of clicks per second.<p>You can read the full article here:<p><a href="https://links.github20k.com/click-vote" rel="nofollow noreferrer">https://links.github20k.com/click-vote</a><p>The open-source Library is here:<p><a href="https://github.com/clickvote/clickvote">https://github.com/clickvote/clickvote</a><p>Please let me know your thoughts!<p>Am I building something useless?
Upvote: | 123 |
Title: I've been on LinkedIn for a while with zero issues with my normal cartoon avatar (which I use absolutely everywhere). Then yesterday, in the middle of my new startup's fundraise, they hit me with an ID verification and banned my account + deleted my startup's page.<p>My profile is missing. I have unanswered messages with people whose emails I don't have. Potential investors etc. Friends reaching out asking me why my profile is missing.<p>They didn't tell me why. Still this morning I didn't know. Now they just responded with an email telling me my "appeal has been denied", and telling me that they removed my avatar for not complying with the TOS.<p>Zero warning, just straight to ban. This is the avatar: https://leclan.ch (it's used with the author's permission)
Upvote: | 130 |
Title: How the hell does a company that make 300m$ in revenue a year has it's main application be this garbage.<p>I'm not even talking about the "review-wall" requirement. The whole thing is garbage, the cities filter for example, it lists all the cities in the world!! you can't filter by country easily, and the filter don't stick from tab to tab, and you can barely search for sh*t, it's so clunky and bad it looks like it's a side project done by an intern.<p>I'm just so confused as to why it's the still by far the most popular website of its genre, I mean look at Twitter, it's a way more sophisticated enginereing product and with so much higher UX investments and it's getting destroyed by Threads soo fast the moment they started slipping.<p>Excuse the rant, I just don't get how there are no real alternatives to Glassdoor, I know levelsfyi exist and stuff similar to that, but nothing at the scale and breadth that Glassdoor covers
Upvote: | 48 |
Title: I purchased a domain name and OneDrive started sending generic "you haven't logged in in awhile" messages and even image thumbnails to the admin@$domain$ address. No way to make it stop, other than blocking onedrive.com. Very creepy.<p>https://answers.microsoft.com/en-us/windows/forum/all/i-received-a-suspicious-email-purporting-to-be/8168982b-bf65-44b7-8561-0f9f64b634b5
Upvote: | 88 |
Title: Been a dev for about 10 years. I’ve had no issues advancing in my career or leveling up my skillset and I’ve generally gotten great feedback from my boss and peers.<p>But I’ve worked for several companies and every single one burned me out, i feel like I don’t have the capacity to grind the way my peers do. Things are usually good for a few months and then management asks for some huge delivery on a tight deadline and it completely burns me out. This has happened so many times in my career. Does anyone else struggle with this consistently?
Upvote: | 72 |
Title: I have two old macbook pros with 8vcpu's and 16 GB of ram each. A comparable computer on AWS would be approximately $100 a month. Why isn't there an easy way to utilize my unused hardware for my production servers (if not for mission-critical stuff, perhaps just for background jobs)?<p>Any counter / pro arguments are welcome.
Upvote: | 64 |
Title: There are many Indians living abroad and are the biggest segment when it comes to cross-border remittances. However, getting started and taking care of their Indian finances while abroad has been extremely manual and tedious.<p>We solve that problem through a wealth management platform only for Indian expats (<a href="https://goinri.com/">https://goinri.com/</a>)<p>Our Launch HN post for more context - <a href="https://news.ycombinator.com/item?id=35389110">https://news.ycombinator.com/item?id=35389110</a><p>One of the trickiest parts for investors in this process has been KYC - Know Your Customer requirements by Indian financial institutions. For the longest time, Indians abroad were either required to submit physical paperwork or required to be in India to complete KYC to get started.<p>We have made this 100% online - You submit your docs in less than 5 mins and your KYC gets done in 3-4 working days.<p>The other problem for Indians abroad who have wanted to invest in India is not knowing where to start. For that, we curate investment options based on cross-border compliances (e.g. Not all mutual funds are open for US / Canada residents) and make the entire investment journey simple and personalized to your investment goals, like Wealthfront. Additionally, the platform also offers support for cross-border tax reporting and repatriation to solve for the full journey in a self-serve manner.<p>Finally, cross-border tax compliance becomes a pain. US and a few other countries have special reporting requirements for foreign investments, and we provide these tax documents online. Just like Robinhood provides a detailed tax statement to attach with your 1099.<p>To educate people on tax and investments, we also conduct webinars (The next webinar on India taxes is on July 15 - <a href="https://lu.ma/inri" rel="nofollow noreferrer">https://lu.ma/inri</a>) and have built Search powered by Chat GPT (<a href="https://goinri.com/blog">https://goinri.com/blog</a>) to answer questions related to India investing, taxation, compliance, and more related topics.<p>We would love to get your feedback! Thank you!
Upvote: | 106 |
Title: I just sold three tickets on Stubhub for a concert I won't be able to make this weekend.<p>When I went to transfer them on Ticketmaster I noticed that the domain the email address was at for all of them appeared to be the names of businesses, but none of them had websites or any presence in google when I searched for them.<p>Being the curious person I am, I then went and looked up the WHOIS for the domain. They all were registered in the last six months, supposedly by Stubhub itself (nobody actually verifies the accuracy of WHOIS data though, its all an honor system)<p>It seems to me like there is probably a hustle going on, where someone is buying tickets under fake names and flipping them. Is this something Stubhub is known to do (SeatGeek pretty openly does it with their return program). Or is lying on the whois part of a larger scheme someone runs?<p>P.S.<p>If anyone is curious, the email domains were crimsonhillpartners.com , oneclassic.org , and ambercovecapital.com
Upvote: | 265 |
Title: I am an aspiring technologist who is interested in understanding how the tech and techniques in High frequency trading systems. Can you recommend me some reading resources?
Upvote: | 100 |
Title: Floneum is a graph editor for local AI workflows. The project focuses on ease of use and community made plugins. The application is a single executable with no dependencies you need to install. You can extend Floneum with WASI plugins that run in a environment that is issolated from your system so that you can use untrusted community plugins without worrying about malware. It uses llamma.cpp to run large language models locally, and wasmtime to run issolated plugins.<p>If you are interested in the project, consider joining the discord, or building a plugin for Floneum in rust using WASI<p>Download: <a href="https://github.com/floneum/floneum/releases/tag/v0.1.0">https://github.com/floneum/floneum/releases/tag/v0.1.0</a><p>Join the discord: <a href="https://discord.gg/dQdmhuB8q5" rel="nofollow noreferrer">https://discord.gg/dQdmhuB8q5</a><p>Build a plugin: <a href="https://floneum.com/docs/developer/index" rel="nofollow noreferrer">https://floneum.com/docs/developer/index</a>
Upvote: | 61 |
Title: App has typical requirements plus some interactions with OS native APIs.<p>This is a startup so the returns of using a single development platform to target iOS/Android seems unquestionable.
Upvote: | 66 |
Title: As per this paper:<p>“Low-Resource” Text Classification: A Parameter-Free Classification Method with Compressors<p>Zhiying Jiang, Matthew Yang, Mikhail Tsirlin, Raphael Tang, Yiqin Dai, Jimmy Lin<p>https://aclanthology.org/2023.findings-acl.426/<p>via: twitter.com/goodside/status/1679358632431853568
Upvote: | 52 |
Title: Hi, do you think the market for job seekers, Startup funding, is recovering? if not could you estimate when it will recover?
Upvote: | 71 |
Title: Hi all,<p>I've spent some time working on music demixing or music source separation algorithms, which take in a mixed song and output estimates of isolated components (e.g. vocals, drums, bass, other).<p>I took a popular PyTorch model with good performance (Open-Unmix, UMX-L weights), reimplemented the inference steps in C++, and compiled it to WebAssembly for a free client-side music demixer.
Upvote: | 190 |
Title: Is there any skill that you know that you are good at, but, were really bad at at some point in your life?<p>What is that skill? How did you become good at it?<p>If there are books or tools or other resources that helped you in achieving that please share your story with us?
Upvote: | 40 |
Title: It appears that Teddit has been effectively shut down via extreme rate limits by Reddit rendering the app unusable. HTTP Error 429 is now returned nearly 100% of the time.
Upvote: | 54 |
Title: Ok, this is going to sound a little weird, but bear with me.<p>I have an idea that has the potential to reduce fentanyl deaths, reduce drug-related violence, and improve operating margins for drug cartels. I assume the cartel leaders are essentially profit motivated, and would be interested in something that would prevent their customers from dying, cut down on the violence that brings law enforcement attention, and improve their profit margins. I realize it sounds a little crazy, but, hypothetically, how would I safely go about setting up meetings with cartels? I'm not even sure where to start or what kind of lawyer to check with for advice.<p>Email in my bio if you'd rather reach out directly.<p>Thanks!
Upvote: | 107 |
Title: We are a small SaaS company with a recently launched product. Currently we have mostly smaller customers (50+ users is "large" for us at the moment).<p>We now have, somewhat unexpectedly this early, the interest of a large enterprise that wants a deal to onboard 5000+ users. The problem is that we have no idea how to price this, as we have not yet had time to scale up our pricing enough to gain the necessary experience.<p>Our current pricing model, for smaller customers, is a simple, linear $49 per-seat plan.<p>- How do pricing for enterprise SaaS usually work?
- What kind of volume discounts are they expecting?
- Should we offer a "flat" price for 1-3 year contract, or per-seat model?
- How did you handle your first large-scale customers?<p>We honestly feel a bit thrown into the deep end here, and we don't want to miss this opportunity just because we are inexperienced.
Upvote: | 163 |
Title: I find existing BitTorrent clients offer a rather old user interface, and it's hard to set up remote access, involving setting up a web server, a TLS certificate, domain name, configuring the router, etc...<p>PikaTorrent tries to offer a modern and easy UI. It is available on desktop (Linux & Windows for now), mobile (Android for now), as CLI as an npm package, and even on the web (as a frontend for remote control).<p>Thanks to WebRTC, one can link the mobile or web app with the desktop or CLI app. This way, it's possible to remotely manage torrents securely and without any complicated setup.<p>From a technical point of view, the app is using Expo and Tamagui to target the web, desktop, and mobile (native), meanwhile libtransmission is used as the torrent engine. So we should expect the same level of performance as the Transmission client.<p>PikaTorrent is open source, and I just released v0.3.0 to let potential users try it out and contribute to the bug/features list on GitHub: <a href="https://github.com/G-Ray/pikatorrent">https://github.com/G-Ray/pikatorrent</a>.
Thank you for your feedback.
Upvote: | 65 |
Title: Last night I learned that Twitter has changed its default Direct Message settings so that by default, you can only receive DMs from people that don't follow you ONLY if they are verified.<p>This is obviously really bad for people who use Twitter for business, journalism, etc. and want to be contacted by anyone. Also, a really shady move since apparently it hasn't been publicly announced.<p>More importantly, switching the setting back to "allow all users" seems not to work that well for me, since I changed the settings last night and it's back to "only from verified accounts" as of this morning, I hope this is an unintentional bug, however I've seen reports of that happening to other users as well.<p>Some sources:<p>https://twitter.com/GergelyOrosz/status/1680156192876756992<p>https://twitter.com/snafuqd/status/1680037620041039873<p>https://twitter.com/halomancer1/status/1680039297637040129
Upvote: | 44 |
Title: Built a tool for transforming unstructured data into structured outputs using language models (with 100% adherence).<p>If you're facing problems getting GPT to adhere to a schema (JSON, XML, etc.) or regex, need to bulk process some unstructured data, or generate synthetic data, check it out.<p>We run our own tuned model (you can self-host if you want), so, we're able to have incredibly fine grained control over text generation.<p>Repository: <a href="https://github.com/automorphic-ai/trex">https://github.com/automorphic-ai/trex</a><p>Playground: <a href="https://automorphic.ai/playground" rel="nofollow noreferrer">https://automorphic.ai/playground</a>
Upvote: | 174 |
Title: dig +short TXT youpay.govorenefekt.com @1.1.1.1 | fold -s<p>You can base64 encode an image, split to TXT records and send over Internet. Useful in certain circumstances. Like when one of the communicating parties is under severe censorship.
Upvote: | 143 |
Title: Hi HN, my name is Ian. My co-founder Edward and I started Grai (<a href="https://grai.io">https://grai.io</a>), an open-source data observability platform. It helps prevent production data outages by evaluating changes to your data pipelines in CI, rather than at runtime.<p>Ever experienced a production outage due to changes in upstream data sources? That's a problem we regularly encountered whether deploying machine learning or keeping a datawarehouse operational and it led us to create Grai.<p>Systematically testing the impact of data changes on the rest of your stack turns out to be quite difficult when the same data is copied and used across many different services and applications. Simple changes like renaming a column in a database can result in broken BI dashboards, incorrect training data for ML models, and data pipeline failure. For example, business users regularly deal with questions like "why does revenue look different in different dashboards".<p>These sort of problems are commonly dealt with by passively monitoring application execution logs for anomalies that might indicate an outage. Our goal was to move that task out of runtime where an outage has already occurred back into testing.<p>At its core, Grai is a graph of the relationships between the data in your organization, from columns in a database to JSON fields in an API. This graph allows Grai to analyze the downstream impact of proposed changes during CI and before they go live.<p>It includes a variety of pre-built integrations with common data tools such as PostgreSQL, Snowflake, dbt, and Fivetran, which automatically extract metadata and synchronize the state of your graph. It's built on a flexible data model backed by REST and GraphQL APIs and a Python client library. This way, users can directly build on top of Grai as they see fit. For example, because every object in Grai serializes to a yaml definition file, sort of like a CRD in Kubernetes, even if a pre-built integration doesn't exist it's fairly easy to manually create or script a custom solution.<p>We made the decision to build open-source from the beginning in part because we believe lineage is underutilized both organizationally and technologically. We hope to provide a foundation for the community to build cool concepts on top and have already had companies come to us with amazing ideas, like optimizing their real-time query pipelines to take advantage of spot price arbitrage between cloud and on-prem.<p>We try not to be overly opinionated about how organizations work, so whether you maintain a development database or run service containers in GitHub Actions it doesn't really matter. When your tests are triggered we evaluate the new state of the environment and check for any impacts, before reporting back as a comment in the pull request.<p>Data observability can have unexpected benefits. One of our customers uses us because we make on-boarding new engineers easier. Because we render an infinitely zoomable Figma-like graph of the entire data stack it's possible for them to visually explore end-to-end data flows and application dependencies.<p>You can find a quick demo here: <a href="https://vimeo.com/824026569" rel="nofollow noreferrer">https://vimeo.com/824026569</a>, we've also put together an example getting started guide if you want to try things out yourself: <a href="https://docs.grai.io/examples/enhanced-dbt">https://docs.grai.io/examples/enhanced-dbt</a>. Since everything is open source, you can always explore the code (<a href="https://github.com/grai-io/grai-core">https://github.com/grai-io/grai-core</a>) and docs (<a href="https://docs.grai.io">https://docs.grai.io</a>), where we have example deployment configurations for docker-compose and Kubernetes.<p>We would love to hear your feedback. If there's a feature we're missing, we'll build it. If you have a UX or developer experience suggestion, we'll fix it. If it's something else, we want to hear about it. We can't wait to hear your feedback and thank you in advance!
Upvote: | 101 |
Title: Hey folks! Shahar and Tal from Keep (<a href="https://www.keephq.dev">https://www.keephq.dev</a>) here!<p>For the last few weeks we’ve been building Peeng and can now share our beta with you: <a href="https://www.peeng.sh" rel="nofollow noreferrer">https://www.peeng.sh</a>. Peeng is the easiest and quickest “heartbeat” architecture we could think of. Just pick a subdomain (e.g. x.peeng.sh), configure an interval, an endpoint, and a payload, and hit that subdomain every <X (interval) seconds — If you won’t, Peeng will send an HTTP POST request to your configured endpoint.<p>It’s Pingdom/Cronitor/heartbeat.sh free alternative (but the other way around and A LOT simpler, with a lot more capabilities), suitable for developers, system administrators, DevOps, and individuals with complex networking situations (think “onprem” or K8s clusters with no inbound).
Instead of inbound heartbeat checks — Peeng presents outbound heartbeat checks!<p>Quick demo: <a href="https://youtu.be/ZX5mrnMRCwU" rel="nofollow noreferrer">https://youtu.be/ZX5mrnMRCwU</a><p>Why we built this:<p>- We needed an easy way to let Keep (<a href="https://github.com/keephq/keep">https://github.com/keephq/keep</a>) customers behind closed networks monitor their Keep instance
- We needed an easy & quick way to setup monitoring for our cronjobs
- We wanted to give people with complex networking situations (e.g. behind a firewall) an easy way to monitor their services/processes<p>The beta version lets you:<p>- Create 5 endpoints for free
- Configure the endpoint and the payload to be sent when the subdomain is not hit
- See the visits (every HTTP GET request to your subdomain) and requests (every HTTP POST sent to your configured endpoint)
- Secret header (x-peeng-secret) that confirms requests are made by you<p>What’s next:<p>- A status page that displays your subdomains and their health together with embeddable status blocks that allow you to display the status of an endpoint in your web page (you can also send query params when sending the GET requests that will be included)
- Rest API (for subdomain creation, beats retrieval, etc., imagine curl -X POST peeng.sh/subdomain -H API_KEY —json {”subdomain”: “hn”, “endpoint”: “<a href="https://..”" rel="nofollow noreferrer">https://xn--ivg</a>, “payload”: {…}})
- Hierarchy-based subdomains that allow you to create a nested heartbeat solution (i.e. dynamically create a heartbeat subdomain under x.peeng.sh → y.x.peeng.sh, z.x.peeng.sh)<p>This is still very early, so we’d love to hear your feedback and opinions. We’re open to any feature request, so just reach out via Intercom :)
Upvote: | 84 |
Title: Over the last decade, firefox's market share has dropped from ~25% to less than 10% (numbers vary depending on where you look). At the same time, the browser only got better in my opinion. What is going on?
Upvote: | 100 |
Title: I feel like a few years ago all the rage was that Amazon and other retailers were going to deliver things via drones, straight to your doorstep<p>Food delivery companies as well seemed to be testing out robots that would bring your food to you albeit in a more sidewalk bound way<p>Its been a few years and I have yet to have a single thing drone delivered to my house<p>What happened to that alternate future? Are companies still working on it? Or did we move on from that idea for some reasons we discovered?
Upvote: | 157 |
Title: I worked at tech-team-of-one startups for years. Now I'm working at a big organisation.<p>In the last sprint, I chalked up 41 story points. The rest of the team (3 other devs and a mostly hands-off tech lead) collectively achieved 9 story points. That's 1 medium-sized ticket each over the space of a fortnight. Everyone has a good excuse, but they ALWAYS have a good excuse.<p>It's driving me nuts. I'm a good coder, but this isn't about me being good, this is about them all being, on paper, completely ineffective. No one inside or outside the team seems to be batting an eyelid about this. I feel resentful and... confused.<p>It's hard to stay motivated when the bar is so low, but I would feel guilty slacking off all day. What do I do?
Upvote: | 51 |
Title: Hi Hacker News! We’re Jay and Vadim from Highlight.io (<a href="https://highlight.io">https://highlight.io</a>). We’re building a truly open source [1] observability platform for modern web applications. We posted some of our tools to HN in recent months [2][3]. Today, we’re excited to formally launch the project, share more about where we’re going, and of course, poll the community for some feedback.
A bit of background: Vadim and I have worked at quite a few startups at this point, and a recurring challenge we’ve faced was tracing usability issues on the frontend to downstream errors and logs on the server. Understanding the real reason behind customer issues was always a chaotic juggling of multiple tools. With the rise of "frontend-forward" frameworks such as NextJS, which blur the boundary between the client and server, the complexity of tracing these issues is only growing.<p>This is where Highlight.io comes in: our product bridges the gap between client and server to give you a holistic view of your entire application.<p>At its core, Highlight.io has three main “products”: Session Replay, Error Monitoring, and Logging. The novelty here is not in each product but in how they are connected. For example, in Highlight.io it’s very easy to click from a given error to the associated user session where it is thrown [4], and from a given error, you can easily inspect all of the logs that fired leading up to it. Ensuring that all of our products work together seamlessly with little to no effort is a core principle of our product strategy. If you’re using a common framework [5], for example, we’ll automatically link your frontend sessions with backend errors and logs. No agents, configuring facets, or anything else, It just works.<p>We depend on several open source projects that help us move quickly. OpenTelemetry (OTEL) [6] is one of them, which helps us with maintainability, i.e. for every language that we support, we only maintain a thin wrapper around its respective OTEL SDK. OTEL is also a great way to enable the community to contribute, and we’re already seeing traction in this space (ie. an open source contributor built a wrapper for a Java SDK [7]).<p>rrweb [8] is another project we leverage heavily for our session replay product. It drives our ability to record and replay the DOM to visualize user flows in the frontend. We’ve had the privilege to work closely with the rrweb team to ship improvements, and we’re now actively sponsoring the project [9].<p>ClickHouse [10] has recently become a loved database on our team, as we historically used Opensearch for search-heavy workloads and started to hit growing pains with ingest throughput. We recently rolled it out for our logging product [3] and plan to replace our sessions and errors (and upcoming tracing work) with the database as well.<p>From a business perspective, Highlight.io is open source under the Apache 2.0 license, and we make money with our hosted product [11]. For the hosted product, you can set billing caps for each offering and we don’t charge for seats. At this point, we have 100+ companies paying for our product (some of which are large enterprises), and thousands of sole developers use Highlight.io every week.<p>On our roadmap [12] for the future includes metrics, tracing, release management and more. We also are launching several updates this week on our launch week page [13].<p>Overall, we’re excited to be sharing Highlight.io with the world, and Vadim and I are particularly excited to get some feedback from the HN community. Please give us a test-drive at <a href="https://app.highlight.io">https://app.highlight.io</a> and let us know what you think. We would love to learn about what you wish you had in an observability product as well as any other experiences and ideas in this space. We look forward to hearing from you!<p>1: <a href="https://github.com/highlight/highlight/blob/main/LICENSE">https://github.com/highlight/highlight/blob/main/LICENSE</a><p>2: <a href="https://news.ycombinator.com/item?id=34897645">https://news.ycombinator.com/item?id=34897645</a><p>3: <a href="https://news.ycombinator.com/item?id=35643255">https://news.ycombinator.com/item?id=35643255</a><p>4: <a href="https://youtu.be/EvGsmbt0F7s?t=65" rel="nofollow noreferrer">https://youtu.be/EvGsmbt0F7s?t=65</a><p>5: <a href="https://highlight.io/frameworks">https://highlight.io/frameworks</a><p>6: <a href="https://opentelemetry.io/" rel="nofollow noreferrer">https://opentelemetry.io/</a><p>7: <a href="https://github.com/highlight/highlight/pull/4812">https://github.com/highlight/highlight/pull/4812</a><p>8: <a href="https://github.com/rrweb-io/rrweb">https://github.com/rrweb-io/rrweb</a><p>9: <a href="https://github.com/rrweb-io/rrweb#gold-sponsors-">https://github.com/rrweb-io/rrweb#gold-sponsors-</a><p>10: <a href="https://clickhouse.com/blog/overview-of-highlightio" rel="nofollow noreferrer">https://clickhouse.com/blog/overview-of-highlightio</a><p>11: <a href="https://highlight.io/pricing">https://highlight.io/pricing</a><p>12: <a href="https://www.highlight.io/docs/general/roadmap">https://www.highlight.io/docs/general/roadmap</a><p>13: <a href="https://highlight.io/launch-week-2">https://highlight.io/launch-week-2</a>
Upvote: | 261 |
Title: Sites hosted via Cloudfront<p>(e.g https://app.circleci.com/) are having an outage and returning 421s and 500s currently
Upvote: | 53 |
Title: I'm looking for resources that could help me write better, more professional code in Python.<p>I've worked extensively as a software developer in Java before, so I am aware of the many things that go into writing "real" code (as opposed to what one writes for Leetcode or even an academic project), from little things like using the right annotations, to bigger ones like dependency injection frameworks.<p>I've used Python for Leetcode, scripting, isolated ML work, writing little games, etc., and I've read and practiced much of PEP8, so it's not that I write bad Python code. But I do feel that I could be doing so much better (since I've done better in Java) and I'd like to get to that level of proficiency as soon as I can.<p>In that vein, I'm looking for any resources that have worked for any of you and you think would be suited for me. Thanks!
Upvote: | 63 |
Title: Does this mean we are slowing down and things will mature? Hopefully.
Upvote: | 126 |
Title: Based on the framerate-independent momentum simulation[0] that I used in my TPMouse script[1]<p>If you've ever used a mouse with Infinite-scrollwheel such as Logitech, this utility for Windows basically recreates that functionality for any generic mouse.<p>Actually, it's even better than that: this allows for simultaneous horizontal and vertical scrolling, so essentially it combines two of the best features of the Logitech MX Master -- horizontal wheel, and unlocked momentum scrolling -- into one intuitive control scheme.<p>To enable horizontal scrolling, set the X-sensitivity to a value you prefer.<p>[0] <a href="https://old.reddit.com/r/Trackballs/comments/ym9q2t/tpmouse_a_virtual_trackball_for_windows/ivl0sr8/" rel="nofollow noreferrer">https://old.reddit.com/r/Trackballs/comments/ym9q2t/tpmouse_...</a><p>[1] <a href="https://github.com/EsportToys/TPMouse">https://github.com/EsportToys/TPMouse</a>
Upvote: | 43 |
Title: Gemini protocol documentation claims that it is possible to write basic web client in 100 lines of code proving protocol simplicity. Easy in modern scripting language but can it be done in ANSI C? Let the source code decide.<p>Someone suggested to share this silly project of mine with HN community so here it is. Enjoy
Upvote: | 91 |
Title: 5 months ago, I launched jimdesigns.co, my one-person subscription-based design studio for SaaS startups, and by doing so, made my first internet $.<p>Today, I passed $10,000 in cumulative revenue. This is not a significant amount whatsoever, but it means a lot to me and feels pretty surreal tbh.<p>This is neither an overnight success as I previously tried for years to make money online by building online tools, without any success.<p>Therefore, by no means do I pretend to have the secret sauce on how to make money online. I just thought I’d share a few thoughts on how I believe that happened for me:<p>1. Focus on what you’re really good at, and what you know best instead of trying to solve problems that you don’t have for an audience that you don’t know. In my case, as a product designer with +10y of experience designing SaaS products for startups, I didn’t have to look further.<p>2. Don’t fall into the build trap. In my case, a design studio only requires building a landing page, which means that the rest of the time can be focused on getting traction (+ client work).<p>3. Deliver outstanding value (product or service). That’s the only thing that matters to your customers. There is no shortcut here, your product or service got to provide significant value. In my case, getting high-quality designs for your SaaS at a fraction of the cost of hiring a world-class designer is a no-brainer.<p>4. Make your pricing a no-brainer when you’re getting started as the initial goal is to learn and get traction, not to get rich.
The $10,000 is total revenue, not MRR but I’m still pumped to have generated my first 5 figures of internet dollar!<p>This post is not about bragging about my little win, but rather to inspire some folks out there to get started. In fact, my only wish is that I’d started earlier.<p>You can follow my journey on Twitter (@JimDesignsCo) as I continue to share my journey.<p>Jim
Upvote: | 97 |
Title: My 3 year old kid was diagnosed with sensory processing disorder. She has trouble with transitions, has meltdowns and gets overwhelmed in public places like stores. Grabs and wants to cling on us and needs constant attention. Also very picky eater and has trouble falling asleep.<p>What can I do ? Should I try to minimize clutter and reduce screen time? Has anyone experienced it themselves or have anyone close they dealt with? What are some workarounds to make parenting easier and keeping my sanity while working a high pressure job? I feel very tired and can barely get any work done ( work from home) .<p>I recently received a coveted job offer as an engineering manager but I am so exhausted that I am planning to turn it down and stay an IC to concentrate helping my kid.
Upvote: | 44 |
Title: Lots of companies building with LLMs and seemingly infinite startups showing off flashy demos with a waitlist. Who's deployed to production, released to everyone, and iterating on them? How are you evaluating them over time and making sure you don't regress stuff that's already working?
Upvote: | 144 |
Title: Hi HN, we’re the founders of Infisical, the open source secret management platform – it provides an end-to-end set of tools to manage your secrets across your team and infrastructure (<a href="https://infisical.com/">https://infisical.com/</a>).<p>Excited to show you all the progress that we’ve made in the past few months after our Launch HN in February (<a href="https://news.ycombinator.com/item?id=34955699">https://news.ycombinator.com/item?id=34955699</a>) and Show HN in December (<a href="https://news.ycombinator.com/item?id=34055132">https://news.ycombinator.com/item?id=34055132</a>).<p>During the previous Show HN and Launch HN, we received a ton of feedback which helped us improve Infisical. We’ve since released:<p>- Secret scanning: a new toolset to block commits with hardcoded secrets and continuously monitor your code.<p>- Folders: Deeper organizational structure within projects to accommodate for microservice architectures and storage of more secret types like user API keys and OAuth tokens.<p>- Node and Python SDKs, Webhooks: More ways to integrate and start syncing secrets with Infisical across your infrastructure.<p>- Integrations with Terraform, Supabase, Railway, Checkly, Cloudflare Pages, Azure Key Vault, Laravel Forge, and more.<p>- Secret Referencing and Importing: to create a proper single source of truth.<p>- 1-click deployments to AWS EC2, Digital Ocean, Render, Fly.io: More ways to self-host Infisical on your own infrastructure.<p>In addition, the platform has become more stable and undergone a full-coverage penetration test; we’ve also begun the SOC 2 (Type II) certification process.<p>Overall, we’re really lucky to have support of the developer community, and, in fact, Infisical has gathered over 7k GitHub stars, and now processes over 200 million secrets per month for everyone from solo developers to public enterprises.<p>Our repo is published under the MIT license so any developer can use Infisical. Again, the goal is to not charge individual developers. We make money by charging a license fee for some enterprise features as well as providing a hosted version and support.<p>Check out Infisical Cloud (<a href="https://infisical.com/">https://infisical.com/</a>) or self-host Infisical on your own infrastructure (<a href="https://github.com/Infisical/infisical">https://github.com/Infisical/infisical</a>). We’d love to hear what you think!<p>We’re excited to continue building Infisical, and keep shipping features for you. Please let us know if you have any thoughts, feedback, or feature suggestions!
Upvote: | 131 |
Title: I posted a few months ago how our Stripe account is a victim of card testing attacks (https://news.ycombinator.com/item?id=35682534)<p>At that point, I was frustrated that they wouldn't address the situation. I had contacted customer support numerous times only to be put on the back burner. It was only after that post that Stripe reached out to me and the problem seemed to damp down for a little bit. Now, less than 3 months later we are experiencing the same issue yet again. Lots of new "customers" that never visit our website but end up signing up for a subscription. Every new customer that signs up we have to manually check to see if they are "real". If not, we refund and block them immediately in Stripe. We missed one the other day and ended up getting a charge back. We reached out to Stripe and they advised us to accept the dispute, which we did. However, Stripe refuses to reimburse us for the charge back fee of $15. In light of Stripe's refusal to correct the problem and reimburse us for their mistakes, we are exploring alternatives for our subscription service. We already moved away from PayPal because of their hostility towards other businesses (though we never experienced any such issues directly).<p>What alternatives to Stripe exist? Stripe's own data suggest that card testing attacks on their network will only increase (see https://images.ctfassets.net/fzn2n1nzq965/Sfz8gaQMU7lsp0Ubd3w1X/84a4254625d05e1fbfe0117de34c1d62/Card_testing_sent_to_Stripe_vs_baseline_v4b_OL.svg?w=900&q=80<p>Thanks!
Upvote: | 42 |
Title: Hello HN,<p>Seven years ago, I complained about Salesforce on HN. Somebody said: "one day, someone will do better". That stuck and today we're trying to be that “someone” with my co-founders Thomas (design) and Charles (eng like me). Our company is called Twenty and our repo is here: <a href="https://github.com/twentyhq/twenty">https://github.com/twentyhq/twenty</a><p>We want to fix two issues: most CRMs aren't enjoyable to use and they often clash with engineering teams.<p>YC encouraged us to launch early. What you see now is about two months' worth of feature development. Our tool only does a small part of what big CRM players offer, but we focused on providing a great user experience on the basics, instead of spreading ourselves thin across a vast range of features and delivering them half-heartedly. Plus, we've found that many small companies like the product as it is because they don't need all the complex stuff.<p>Once we have covered the basics, we’ll soon be working on three big features: - Moving to a robust metadata-driven architecture; - Providing innovating ways to extend the CRM with Typescript; - Making it easy to connect data sources, and fetch data in real-time like in BI tools.<p>The startup world is littered with ghosts of so-called "Salesforce killers", so we know it sounds naive to pitch ourselves in a similar way. But we think that if someone ends up changing this market, it will most likely be through a community-led effort. And there hasn’t really been any serious attempt to start a great new open source CRM in the last decade.<p>Twenty is built with Typescript, React, and NestJS with GraphQL, and licensed under AGPL. We plan to make money by offering a hosted version. Our docs are here: <a href="https://docs.twenty.com" rel="nofollow noreferrer">https://docs.twenty.com</a>. Try on cloud: <a href="https://app.twenty.com" rel="nofollow noreferrer">https://app.twenty.com</a>.<p>Dev setup and demo: <a href="https://www.loom.com/share/7b20b44d8d5146fea8923183511bb818" rel="nofollow noreferrer">https://www.loom.com/share/7b20b44d8d5146fea8923183511bb818</a> (Loom said they couldn’t provide a transcript because they don’t support “language other than english” haha... apologies in advance!)<p>We’re very eager to get your feedback as we haven’t launched anywhere before this post. What's your CRM story? What should we prioritize next?
Upvote: | 415 |
Title: Obviously interested in all of it, but here is what I am rally trying to solve:<p>I live in an apartment, so unfortunately I cannot run full fat rack mounted equipment for space, heat, and noise reasons. Also, a partner who isn't keen on the sight of it<p>I was thinking maybe some small thinkcenter or dell micro PCs + a NAS (Thinking QNAP because they are cheaper). Lower TDP would be great too. Power costs are a concern.<p>Thinking about Nextcloud, plex, jellyfin, home assistant, local backups, etc... I know that it evolves quickly
Upvote: | 64 |
Title: Bad news for me, I have lots of accounts I set up with forwarding back in the way, as a way to keep control of spam. I want to keep those accounts as they’re important to me even though I didn’t sign into them.<p>So I was bummed to receive this message about deletion of “inactive” accounts:<p>https://pastebin.com/fgLUN45V<p>It makes me sad as I consider all my accounts active.
Upvote: | 76 |
Title: Popular among my group seems to be:<p>- Google Flights<p>- Kayak<p>- ITA Matrix (which shockingly still seems to be online)<p>Is there any new great options out there for finding cheap flights?
Upvote: | 48 |
Title: I have a lot of experience with side-projects and startups. I will briefly explain my experience. I'm coming from small eastern European country, so my perspective on everything might be different.<p>1. My country is too small and market is small. You can do something here, but everything is so limited that it's terrible. 3.5M people and then when you target certain people who wants your product or service or app or website, you get super small amount of people<p>2. The Balkans (group of countries around my country where people speak the same language) is a better story, but every nation is so isolated. I asked yourself, how often do I use products/app/websites from the other countries and vice versa? Very rarely. Every country is in its own bubble.<p>3. European market. Every country has its own language, culture, mentality, etc. it's hard to get into those countries especially if you don't know the lang.<p>4. The US market is great because it's unified, it's easier to go viral, the language is the same, purchasing power is great, but it's a bit of a problem because me as an European don't know their habits and thinking, I have no connections or friends there, I don't know where and how to start. The market is saturated, something is happening there all the time, and now you have to break through with something new.<p>5. The UK is ok market, they are more related and similar to the USA. For example OnlyFans is from the UK, MillionDollarHomePage the same etc.<p>I had quite a lot of partners, it didn't really work out. But I think it's me, I think I function better alone. Everything I tried failed, and I tried exactly in those countries that I mentioned. Worst of all, I wasn't persistent enough.<p>I have the feeling like I'm trying impossible. I had no money or capital, everything I did I relied on word of mouth, but that's shit. Without good capital that will support the marketing, it's all a shot in a dark.<p>Sounds like I'm complaining here or trying to find excuses but I'm not, I'm still trying to make it. I'm just confused, I don't know what to pursue, where to try, which market to try
Upvote: | 41 |
Title: Hi HN<p>A few folks and I have been working on this project for a couple weeks now. After previously working on the Docker project for a number of years (both on the container runtime and image registry side), the recent rise in open source language models made us think something similar needed to exist for large language models too.<p>While not exactly the same as running linux containers, running LLMs shares quite a few of the same challenges. There are "base layers" (e.g. models like Llama 2), specific configuration to run correctly (parameters, temperature, context window sizes etc). There's also embeddings that a model can use at runtime to look up data – we don't support this yet but it's something we're looking at doing soon.<p>It's an early project, and there's still lots to do!
Upvote: | 284 |
Title: I'm a self-taught programmer but I recently realized I need to really understand how to write programs elegantly. GPT-4 was a huge motivation for this because whenever I asked it to rewrite the code like a professional Python programmer, it would come up with amazing things. But after GPT-4 became nerfed by OpenAI, I felt a void: how can I keep writing elegant programs? My answer is: I should learn how to do that myself.<p>SICP uses Scheme, which I don't mind. My main concern is that it's an old book. Are there ideas and concepts not discussed in the book which are crucial in today's programming landscape? Will I be better off reading a book that uses Python in the first place?
Upvote: | 120 |
Title: Hey hacker news,<p>We’re the cofounders at Psychic.dev (<a href="http://psychic.dev">http://psychic.dev</a>) where we help companies connect LLMs to private data. With the launch of Llama 2, we think it’s finally viable to self-host an internal application that’s on-par with ChatGPT, so we did exactly that and made it an open source project.<p>We also included a vector DB and API server so you can upload files and connect Llama 2 to your own data.<p>The RAG in RAGstack stands for Retrieval Augmented Generation, a technique where the capabilities of a large language model (LLM) are augmented by retrieving information from other systems and inserting them into the LLM’s context window via a prompt. This gives LLMs information beyond what was provided in their training data, which is necessary for almost every enterprise application. Examples include data from current web pages, data from SaaS apps like Confluence or Salesforce, and data from documents like sales contracts and PDFs.<p>RAG works better than fine-tuning the model because it’s cheaper, it’s faster, and it’s more reliable since the provenance of information is attached to each response.<p>While there are quite quite a few “chat with your data” apps at this point, most have external dependencies to APIs like OpenAI or Pinecone. RAGstack, on the other hand, only has open-source dependencies and lets you run the entire stack locally or on your cloud provider. This includes:<p>- Containerizing LLMs like Falcon, Llama2, and GPT4all with Truss
- Vector search with Qdrant.
- File parsing and ingestion with Langchain, PyMuPDF, and Unstructured.io
- Cloud deployment with Terraform<p>If you want to dive into it yourself, we also published a couple of tutorials on how to deploy open source LLMs for your organization, and optionally give it access to internal documents without any data ever leaving your VPC.<p>- How to deploy Llama 2 to Google Cloud (GCP): <a href="https://www.psychic.dev/post/how-to-deploy-llama-2-to-google-cloud-gcp">https://www.psychic.dev/post/how-to-deploy-llama-2-to-google...</a>
- How to connect Llama 2 to your own data using RAGstack: <a href="https://www.psychic.dev/post/how-to-self-host-llama-2-and-connect-it-to-your-private-data">https://www.psychic.dev/post/how-to-self-host-llama-2-and-co...</a><p>Let a thousand private corporate oracles bloom!
Upvote: | 84 |
Title: I've been a software engineer at my (large) company for a long time, in a very highly technical job at what you would call "staff level".<p>So far everything went well, but factors outside of my control start to turn things sour. It's not an issue pertaining to my individual role, just general developments in our particular part of the company.<p>I'm thinking about seeing what else is there. Before I go into disappointment, what is the market like when you are 45 or older?
Upvote: | 56 |
Title: I'm looking to process ~20k transactions a month with an average ticket size of $15 per transaction. Would ISOs(Independent sales organizations) with sponsor banks be good options to consider?
Upvote: | 155 |
Title: I've been revisiting Agile principles through "The Art of Agile Development" lately. While it's a valuable refresher, I'm having difficulty reconciling the concept of 'sprints' in Agile after 11 years in the industry. My observation is that significant time and energy go into sprint planning, discussions, and subsequent introspection when things don't pan out. Agile sprints, to me, seem reminiscent of calorie counting - good in theory but more often failing in practice.<p>To clarify my intent:<p>1. I'm seeking input primarily from software developers rather than project managers, executives, customers, etc.
2. This is not an indictment of Agile overall. I endorse limiting work-in-progress, decomposing larger tasks, continuous delivery, and scheduling improvement time - all achievable without sprints.
3. When I refer to "Agile", I'm talking about methodologies adhering to the Agile manifesto.<p>To be clear, I don't believe sprints inherently benefit developers, and it feels as if the Agile community hasn't adequately consulted engineers on this topic. When I've asked developers about their feelings on sprints (excluding WIP limits or task decomposition), the responses lean towards indifference or opposition. Is the effort of sprint planning worthwhile when faced with such apathy?<p>Common pro-sprint arguments I've encountered include:<p>1. Incremental improvement/Faster feedback loop. However, sprint planning seems overly complex, with most of the book dedicated to explaining a convoluted system and potential pitfalls.
2. Early value delivery. But we're in the era of CI/CD, where deployments can happen multiple times a day. Are sprint deadlines just relics of a time when physical software delivery was necessary?
3. Autonomy. I'm not sold on this either. Autonomy for developers is often overstated. We're not pursuing hobbies or creating art; we want a drama-free, predictable, and coordinated work environment where we're respected. Autonomy, especially technical, is important, but it's not the only thing that matters on a software team, and two-week Jira ticket batches don't necessarily provide it.<p>My reservations include:<p>* Overcomplication, with an entire industry of trainers, books, and seminars evolving around it. Over time, sprint discipline often deteriorates or transforms into something less Agile.<p>* Lack of relevance for teams practicing CI/CD.<p>* The creation of artificial deadlines leading to overtime, burnout, and, eventually, high staff turnover.<p>* Inability to handle reality. Interruptions and unforeseen tasks will occur. The database may malfunction, revealing an architectural issue requiring immediate attention. While frustrating, these could be managed with stricter WIP limits.<p>* Contradiction of the Agile manifesto's "Responding to change over following a plan". Sprint planning can degenerate into "micro-waterfall".<p>After substantial reflection, I'm left with these thoughts:<p>1. Most benefits attributed to sprints could be addressed through strict WIP limits. "The Art of Agile" disappointingly provides only cursory coverage of continuous planning methodologies (like Kanban, Scrumban, etc.), recommending most organizations adopt a sprint-based approach without further elaboration.
2. If sprints offer any advantages, they aren't for the developers. I've seldom encountered developers who genuinely enjoy sprints or would miss them in a sprint-less model.<p>Do any fellow developers here <i>actually enjoy</i> conducting sprints? If so, why and how do you make it work?<p>And if any of the book's authors frequent HN, your insights would be appreciated. The book was excellent, barring the points mentioned above.
Upvote: | 73 |
Title: I'm trying to get an understanding on what quality means in terms of publishing research and I learn best through examples. Any recommendation is highly appreciated ^^.<p>I'm also interested in papers from many disciplines, so the wider the range of domains, the higher the value!
Upvote: | 237 |
Title: Could've sworn there were 1 or 12 startups in the recent batch doing this...but can't find any off the top of my google search
Upvote: | 188 |
Title: I'm building an app, and starting to realize that it makes sense to pay someone to help me. I don't have the budget for full time, or even part time, really. I want to hire a person or agency to help me with tasks as I need them and as I can afford. My ideal situation is to hire a consistent person who I can give access to my repo and have them work directly in GitHub on assigned tasks. I've used Upwork for other tasks in the past, but I'm wondering if Upwork is the best place to find affordable and good dev help. Are there any other platforms you would recommend for finding consistent devs who are willing to work on occasional small or medium sized tasks via GitHub?
Upvote: | 61 |
Title: Hello HN!<p>So early this year I decided to build an iOS browser offering a large range of customization options, to accommodate anyone’s usage and visual preferences. Been using it as my default browser for 6 months and not looking back.<p>On first launch it looks and behaves a bit like Safari, but to give you an idea, here are the various ways I customized it on my phone.<p>- More screen real estate for the webpage, less for the toolbar.<p>- A compact toolbar containing only the buttons for actions I use the most: new tab, close tab, open tab list. Since I use them hundred times a day, I need them available with button taps instead of swipe gestures.<p>- A toolbar button layout adapted to my left-handedness.<p>- A toolbar that disappears on scroll to allow full-screen reading.<p>- The toolbar and address bar at the bottom, as they should be.<p>- A popup menu showing the full URL and the buttons I use less frequently: back/forward (already available as screen edge gestures), share, reload, settings, etc.<p>- Showing the page title in the toolbar.<p>- Read time estimation for each tab.<p>- Opening the keyboard automatically when I open a new tab.<p>- Sorting tabs by read time, so that I can decide what to read based on how much time and focus I have.<p>- Grouping tabs by domain.<p>- A flat, condensed tab list, without snapshots.<p>- A full-black toolbar in dark mode to read at night.<p>It’s still early days but things like content blockers, reader mode, iPad support, and more should arrive soon enough.<p>And of course, no analytics/monitoring/telemetry, no account creation, no backend. It’s not open source, but you can also inspect the app’s web views with Safari developer tools to see what’s going on under the hood.<p>Would love to hear if the level of personalization my app provides resonates with like-minded people.<p>Have a great day!
Upvote: | 226 |
Title: To their credit, Microsoft puts the cursor in the text field ready for me to type in my authenticator code. Most web sites I've found, do not. So I look up with my shiny new code, ready to start typing, realize I need to move my hand to the mouse, place the cursor into the field then there's a 50% change I've forgotten the code and have to look back at my phone.
Upvote: | 80 |
Title: Hi all! Excited to share cc.dev after months of work and refinement.<p>The idea for this product came from trying to do email marketing for my side project, CubeDesk, a site where Rubik's Cube enthusiasts can time themselves, race with one another, train algorithms — it's a fun niche!<p>With over 40k users, sending even a single campaign was becoming expensive with MailChimp. I knew AWS SES would be much cheaper, but it’s just an API with none of the other necessities you need for a robust email marketing platform.<p>Beyond cost, I was also frustrated with having to make sure my database was always in sync with MailChimp and the audience schema they enforced. If I wanted to email every user who had completed 10 solves, that would be a whole ordeal and eat up hours of my day.<p>So, I started (and am now launching):<p><a href="https://cc.dev" rel="nofollow noreferrer">https://cc.dev</a><p>cc.dev connects directly to your database and lets you write SQL queries to target your audience. It's backed by AWS SES, so the cost to send emails is significantly less than what you're used to seeing. Combined with a template builder, media management, and campaign monitoring, cc.dev is meant to be your final destination whenever you need to send marketing emails to your users.<p>Would love to hear your feedback on this! If you're interested in trying out cc.dev as your email marketing platform, shoot me an email and let's have a chat: kash at cc.dev
Upvote: | 302 |
Title: Hi everyone,<p>Great Books Homeschool has just released this free tool for generating high school transcripts using the standard American unweighted GPA system. It's available to the public at no cost, and no account creation is required.<p>These are both resources that would have saved me time as a new homeschooling parent, and I hope they are helpful to others.<p>Comments and feedback are welcome!
Upvote: | 68 |
Title: My wife, a doctor who knows virtually nothing about software development, made an interesting observation when I told her how I was using ChatGPT to write code.<p>She noted that if a new programming language is created, it will have far less training data for ChatGPT than old programming languages do. So, to the degree that ChatGPT or similar AIs become an important part of writing code, the new language will be at a disadvantage even if it's intrinsically superior. So people won't use it, so it will never catch up to the existing ones in ChatGPT support due to lack of training data. So people won't use it, so...<p>In response, I noted that something like SudoLang is rooted in ChatGPT and so doesn't seem to have that problem. But I also said that her point seemed valid and worth thinking about. It seems like the outcome may be that the next truly popular programming language may be intrinsically tied to an AI. It won't necessarily be SudoLang or much like it, but that's not necessarily the only way it could be done.
Upvote: | 45 |
Title: Basically I would like to have an extension where I could say "click publish" and it will find the button with the text publish and click it.<p>Is there anything like this for FF, I ask because the FF voice repo got shut down a couple years ago and haven't found a competitor.
Upvote: | 60 |
Title: My country (Israel) is going through a coup d'état, which would result in a regime that can use surveillance without going through the courts.<p>Israel has one the most advanced surveillance capabilities, that has long been used to target other peoples and nations, but as far as is known, not its own people. This might now change.<p>I am trying to prepare in advance with encryption software, and optimally a way to communicate if traditional networks go down. Unfortunately other nations have gone through similar situations recently, so I'm wondering if there's a known guideline for these situations.
Upvote: | 83 |
Title: I've started using HN pretty recently, and already impressed by some of the not-so-common and interesting topics and discussions here. Curious to know your favorites!
Upvote: | 56 |
Title: .
Upvote: | 347 |
Title: Hey, I’m Robin and I’m the founder/CTO at Artie (<a href="https://www.artie.so/">https://www.artie.so/</a>). We solve the problem of stale production data in the data warehouse. We’re open source, and you can try it for free here: <a href="https://github.com/artie-labs/transfer">https://github.com/artie-labs/transfer</a>.<p>Specifically, we do real time data replication from databases to data warehouses. We leverage change data capture (CDC) and stream processing to perform data transfers efficiently, enabling sub-minute latency and significant cost savings. Here’s a quick demo (jumping straight to the important part): <a href="https://www.youtube.com/watch?v=uAi1tm4gd9U#t=81s">https://www.youtube.com/watch?v=uAi1tm4gd9U#t=81s</a>.<p>I encountered this problem when I was a heavy data warehouse user at prior jobs. The data in our data warehouse was super lagged and analytics were always stale. Imagine a fintech company performing anti-fraud/transaction monitoring with batched ETLs and finding out that fraudulent transactions occurred 24 hours ago. This was very frustrating to me! Since my background was in distributed systems and database architecture, I knew that there was a better way to perform data transfers.<p>The more teams I spoke with, the more I realized this was a real pain point. People wanted real time data for analytics, fraud alerting, transaction monitoring, and training AI/ML models; but there wasn’t an easy out-of-the-box solution. Companies were either constrained on latency or schema integrity/data usability or data accuracy. Companies started telling me that if I built a tool that is robust, near real time, but also maintained schema integrity and data accuracy, they would very happily pay for it.<p>So I built Artie, a real time, open-source data streaming solution to transfer data from databases to data warehouses and handle schemas automatically in-flight (DMLs and DDLs).<p>Typical ETL solutions leverage batched processes that run on a schedule (DAGs, Airflow), which cannot achieve real time data syncs. This means that when companies aggregate production data into their data warehouse, the underlying data is always stale.<p>Streaming change data capture (CDC) logs is a more efficient way to transfer data, and helps lower networking/data ingestion costs as well. However, building data pipelines with CDC streaming is complicated. I wanted Artie to be the tool that abstracts away that complexity so that any company can benefit from having real time data.<p>A big challenge is implementing CDC streaming for stateful data (i.e. row updates/deletes) and schema changes (DDLs), which most streaming solutions just don’t address, meaning that complexity is passed down to the customer. A lot of in-house streaming solutions leverage some combination of Debezium + Kafka/Kinesis + Apache Flink and are able to achieve near real time syncs, but they only handle append-only events (inserts) and don't handle schema changes/schema evolution like DMLs and DDLs. Not handling thse means the data at the destination doesn't look exactly like the production database, which obscures the source of truth. You end up having to do additional work to make the data warehouse tables match the source DB.<p>So how do we offer a robust CDC streaming solution? We grab CDC logs using Debezium and/or our custom connector (which solves for certain edge cases that Debezium doesn’t handle) and push them into Kafka (or Google Pub/Sub). Kafka helps ensure ordering and ease of recovery upon an outage - we use one table per topic, and the partition key is the primary key(s) to ensure no out of order writes. Artie then consumes these events from Kafka and we have an in-memory DB with our typing library that can infer schemas (DML and DDL), performs optimizations like deduplications, and then flushes data to the data warehouse. When the data warehouse confirms a successful merge, we then commit the offset within Kafka. This all happens with sub-minute data latency even at high volumes (several TBs or billions of rows).<p>However the data looks in your database, it should look exactly the same in your data warehouse. We also strive to handle all data types with no exceptions - i.e. supporting TOAST columns, composite keys as primary keys, arrays, large rows (>1MBs), etc. (I say “strive” because I’m sure we haven’t seen all possible data types yet!).<p>We’ve been live for a while now. Several companies use us to update all their analytic dashboards in real time (with dashboards built on top of their data warehouse). Fintech platforms use us to perform financial transaction monitoring. A utilities software platform uses us to grab video/photo data to perform risk/hazard assessment against ML models.<p>Artie is live and supports PostgreSQL, MySQL, MongoDB sources and Snowflake, BigQuery, Redshift destinations. We make money from our hosted service. We charge based on usage (# of rows transferred per month, not including initial snapshots).<p>We’d love for you to try it out! You can get started with the open source version here: <a href="https://github.com/artie-labs/transfer">https://github.com/artie-labs/transfer</a>. We have a small OSS Slack community (www.artie.so/slack) – feel free to ping us for help or any other requests.<p>For our hosted version, we need to ensure that we have enough storage and compute capacity provisioned, so we’re asking cloud users to hop on a quick call with us before we activate your account. Eventually we’ll have easy self-serve functionality but that’s not 100% built yet, so for now we set up a Slack channel to ensure smooth deployments. If you’re willing to work with us on that, we’ll be super excited to show you what we’ve got. Just email us at [email protected] or request access from <a href="https://www.artie.so">https://www.artie.so</a>.<p>We’d love for you to try the OSS or hosted solution and give us feedback! We’re eager to improve the product and test it against various workloads and data types :)
Upvote: | 123 |
Title: This was something I built while trying to look for housing in Toronto that was decently transit-accessible to my office while still cheap.<p>The backend is written in Rust. It parses public GTFS data from transit agencies and performs a simple heuristics-based BFS on the bus lines to calculate how long to reach all points in a city.<p>The frontend uses React and Mapbox GL to render each individual road segment based on how long it takes to reach.<p>This project was a great excuse to learn Rust, deployments, and mapping. The source code is here if you are interested: <a href="http://github.com/econaxis/time2reach">http://github.com/econaxis/time2reach</a>
Upvote: | 255 |
Title: Hey everyone! After another year of building as a solo dev on nights and weekends, I'm back with an update on this post: <a href="https://news.ycombinator.com/item?id=31083093">https://news.ycombinator.com/item?id=31083093</a>.<p>TL;DR - ProjectionLab (<a href="https://projectionlab.com" rel="nofollow noreferrer">https://projectionlab.com</a>) is a privacy-friendly personal finance planning tool where you can create financial plans that go beyond the standard online retirement calculators. And by popular request, it now supports self-hosting for Lifetime users!<p>Something I'm grateful for is that our community here on HN is the difference between PL existing and not. There was actually a time early on when I was one day away from halting work on it. I posted here on a whim, and was shocked to receive some really constructive and energizing feedback that went on to power my indie dev journey over the past two and a half years.<p>As a quick recap, the story started when I dove head-first down the financial independence rabbit hole. I wanted a hands-on and visual way to explore the trade-offs between different life paths. One thing led to another, and I decided to build ProjectionLab.<p>After last year's Show HN, I really put my nose to the grindstone, and here are some of the big developments:<p>- Self-hosting for Lifetime users (spin up your own private deployment, based on Docker Compose, includes support for auth/encryption)<p>- Cash-flow visualization for each simulated year (sankey charts)<p>- Tax analytics (detailed breakdowns for projected income, taxes, marginal rates, effective brackets, etc)<p>- Major redesign of entire app, with landing page and resources now split into separate project<p>- Filing separately option to improve support for international locations that don't have joint filing<p>- Roth Conversions and 72t (SEPP) distribution modeling<p>- Improvements to US tax estimation (Secure 2.0 updates, rental property tax deductions, Medicare + IRMAA, NIIT, principal residence exclusion, etc)<p>- Better support for planning as a couple<p>- More modeling options for cash-flow priorities to support different budgeting philosophies and goals<p>- Extra liquidity + withdrawal options, ability to fund expenses with specific accounts or route income to specific accounts<p>- Customization options for Monte Carlo simulations (characterization of success rates and outcome types, option to set random seed, etc)<p>- And a whole bunch more! (<a href="https://projectionlab.com/changelog" rel="nofollow noreferrer">https://projectionlab.com/changelog</a>)<p>The HN community has had a huge role in shaping my overall direction with PL, and I can't wait to hear what you all think of the updates and where you would like to see things go from here.<p>As always, PL is free to try, with no need to create an account. It does not ask to link your financial accounts, and it has a sandbox mode if you just want to hop in and see how it works.<p>--Kyle
Upvote: | 766 |
Title: As Richard Feynman famously said: "What I cannot create, I do not understand." I like to learn things to bare bones. But almost all courses are shallow.<p>Now I am learning https://www.nand2tetris.org/ and it's fantastic by structure and explanation. In the past, I also completed JS and HTML courses by Anthony Alicea on Udemy. These courses also follow a similar approach to truly understanding how things work under the hood.<p>What other books/courses could you recommend from any area of programming?
Upvote: | 109 |
Title: I wanted to share a concerning experience we recently had with our domain name on the GoDaddy platform. A few days ago, we purchased a domain for our business for 10$. The transaction went smoothly, and the domain was added to our account.<p>However, to our astonishment, we later discovered that the domain had been removed from our account and was put back on sale by GoDaddy at an outrageously higher price 2000$.<p>We had genuine intentions to use the domain for our business endeavors. We followed all the proper procedures, paid the required amount, and were under the impression that the domain was rightfully ours.<p>It really seems like the domain has been taken away from us to be resold at a much higher price.<p>We have all the necessary documentation to prove our rightful ownership of the domain (bank statement, screenshots).<p>At this point, we have reached out to GoDaddy's customer support and are awaiting their response to rectify the situation promptly.<p>Can we do anything else?
Upvote: | 79 |
Title: I've been watching Elon Musk's takeover of Twitter, and I've even been personally affected by it (negatively), so I wanted a way to vent using open-source software and some very basic art skills. I woke up yesterday with an idea in mind for a MAD Fold-In[1].<p>A MAD Fold-In is a piece of artwork on the inside-back cover of each issue of "MAD Magazine". The main image can be folded to reveal a hidden, secondary image.<p>I wanted to share my interactive picture with others, but printing a fold-in and mailing it to friends with instructions seemed really hard, so I searched for a digital way to do it.<p>I found a blog post[2] from Thomas Park where he already did 100% of the work necessary to make a MAD Fold-In using nothing but CSS and HTML and a normal PNG image.<p>Using Inkscape and some Creative Commons images, I drafted a rough piece of artwork and tweaked it until the CSS folded it nicely.<p>I wouldn't normally share here, but I think other HN readers may get some utility from seeing and trying out a purely CSS implementation of a MAD Fold-In, even if I didn't write the CSS myself. I'm hoping to start a trend where others make their own fold-ins.<p>Moreover, Elon Musk picked today of all days to rebrand the Twitter logo as a unicode "X" character (𝕏). So, the Twitter bird really _did_ die today, as happened in my MAD fold-in, making my post somewhat topical and weirdly apropos.<p>If this post is too off-topic, please feel free to take it down. If it stays up, please let me know what you think. In the event my server goes down, I made a WayBack Machine archive of the site [3].<p>[1] <a href="https://en.wikipedia.org/wiki/Mad_Fold-in" rel="nofollow noreferrer">https://en.wikipedia.org/wiki/Mad_Fold-in</a><p>[2] <a href="https://thomaspark.co/2020/06/the-mad-magazine-fold-in-effect-in-css/" rel="nofollow noreferrer">https://thomaspark.co/2020/06/the-mad-magazine-fold-in-effec...</a><p>[3] <a href="https://web.archive.org/web/20230724221700/https://brendenhyde.com/foldin/" rel="nofollow noreferrer">https://web.archive.org/web/20230724221700/https://brendenhy...</a>
Upvote: | 79 |
Title: I’m mulling over an idea to make an all-in-one reading and writing tool, which empower thinking in this information overload era.
Upvote: | 43 |
Title: Just got this email a few mins ago:<p>> We wanted to let you know our team reviewed your content, and we think it violates our spam, deceptive practices and scams policy. We know you may not have realized this was a violation of our policies, so we're not applying a strike to your channel. However, we have removed the following content from YouTube:
> URL: ht tps://filmmusic. io/standard-license<p>Looks like some false positive, Youtube is really banning links to a valid and safe URL.
The "appeal" link doesn't seem to work either. Worst part is that this probably affects thousands of videos using royalty free music, and they are now breaking the license due to deleting the attribution link :(<p>Any ideas on how to contact YT to fix that?<p>Edit: they have fixed it!
Upvote: | 54 |
Title: Anything that have had a positive impact on your life.
Upvote: | 46 |
Title: Title says it. Interested to hear from founders who went from self hosted sites e.g. on AWS to shopify/squarespace/… and then back to self hosted. I am considering stopping self hosting to focus more on content and product and getting rid of the burden of mastering html css javascript etc and maintaining servers.
Upvote: | 43 |
Title: Hi HN, we’re Mayank and Matt of Roundtable (<a href="https://roundtable.ai/">https://roundtable.ai/</a>). We use LLMs to produce cheap, yet surprisingly useful, simulations of surveys. Specifically, we train LLMs on standard, curated survey datasets. This approach allows us to essentially build general-purpose models of human behavior and opinion. We combine this with a nice UI that lets users easily visualize and interpret the results.<p>Surveys are incredibly important for user and market research, but are expensive and take months to design, run, and analyze. By simulating responses, our users can get results in seconds and make decisions faster. See <a href="https://roundtable.ai/showcase">https://roundtable.ai/showcase</a> for a bunch of examples, and <a href="https://www.loom.com/share/eb6fb27acebe48839dd561cf1546f131" rel="nofollow noreferrer">https://www.loom.com/share/eb6fb27acebe48839dd561cf1546f131</a> for a demo video.<p>Our product lets you add questions (e.g. “how old are you”) and conditions (e.g. “is a Hacker News user”) and then see how these affect the survey results. For example, the survey “Are you interested in buying an e-bike?” shows ‘yes’ 28% [1]. But if you narrow it down to people who own a Tesla, ‘yes’ jumps to 52% [2]. Another example: if you survey “where did you learn to code”, the question “how old are you?” makes a dramatic difference—for “45 or older” the answer is 55% “books” [3], but for “younger than 45” it’s 76% “online” [4]. One more: 5% of people answer “legroom” to the question “Which of the following factors is most important for choosing which airline to fly?” [5], and this jumps to 20% when you condition on people over six feet tall [6].<p>You wouldn’t think (well, we didn’t think) that such simulated surveys would work very well, but empirically they work a lot better than expected—we have run many surveys in the wild to validate Roundtable's results (e.g. comparing age demographics to U.S. Census data). We’re still trying to figure out why. We believe that LLMs that are pre-trained on the public Internet have internalized a lot of information/correlations about communities (e.g. Tesla drivers, Hacker News, etc.) and can reasonably approximate their behavior. In any case, researchers are seeing the same things that we are. A nice paper by a BYU group [7] discusses extracting sub-population information from GPT/LLMs. A related paper from Microsoft [8] shows how GPT can simulate different human behaviors. It’s an active research topic, and we hope we can get a sense of the theoretical basis relatively soon.<p>Because these models are primarily trained on Internet data, they start out skewed towards the demographics of heavy Internet users (e.g., high-income, male). We addressed this by fine-tuning GPT on the GSS (General Social Survey [9] - the gold standard of demographic surveys in the US) so our models emulate a more representative U.S. population.<p>We’ve built a transparency feature that shows how similar your survey question is to the training data and thus gives a confidence metric of our accuracy. If you click ‘Investigate Results’, we report the most similar (in terms of cosine distance between LLM embeddings) GSS questions as a way of estimating how much extrapolation / interpolation is going on. This doesn’t quite address the accuracy of the subpopulations / conditioning questions (we are working on this), but we thought we are at a sufficiently advanced point to share what we’ve built with you all.<p>We're graduating PhD students from Princeton University in cognitive science and AI. We ran a ton of surveys and behavioral experiments and were often frustrated with the pipeline. We were looking to leave academia, and saw an opportunity in making the survey pipeline better. User and market research is a big market, and many of the tools and methods the industry uses are clunky and slow. Mayank’s PhD work used large datasets and ML for developing interpretable scientific theories, and Matt’s developed complex experimental software to study coordinated group decision-making. We see Roundtable as operating at the intersection of our interests.<p>We charge per survey. We are targeting small and mid-market businesses who have market research teams, and ask for a minimum subscription amount. Pricing is at the bottom of our home page.<p>We are still in the early stages of building this product, and we’d love for you all to play around with the demo and provide us feedback. Let us know whatever you see - this is our first major endeavor into the private sector from academia, and we’re eager to hear whatever you have to say!<p>[1]: <a href="https://roundtable.ai/sandbox/e02e92a9ad20fdd517182788f4ae7e1f96a849c0">https://roundtable.ai/sandbox/e02e92a9ad20fdd517182788f4ae7e...</a><p>[2]: <a href="https://roundtable.ai/sandbox/6b4bf8740ad1945b08c0bf584c84c1202a5fec53">https://roundtable.ai/sandbox/6b4bf8740ad1945b08c0bf584c84c1...</a><p>[3] <a href="https://roundtable.ai/sandbox/d701556248385d05ce5d26ce7fc776bb4d32fad0">https://roundtable.ai/sandbox/d701556248385d05ce5d26ce7fc776...</a><p>[4] <a href="https://roundtable.ai/sandbox/8bd80babad042cf60d500ca28c40f7db413f553a">https://roundtable.ai/sandbox/8bd80babad042cf60d500ca28c40f7...</a><p>[5] <a href="https://roundtable.ai/sandbox/0450d499048c089894c34fba514db4042eafb6c0">https://roundtable.ai/sandbox/0450d499048c089894c34fba514db4...</a><p>[6] <a href="https://roundtable.ai/sandbox/eeafc6de644632af303896ec19feb69ac4714e24">https://roundtable.ai/sandbox/eeafc6de644632af303896ec19feb6...</a><p>[7] <a href="https://arxiv.org/abs/2209.06899" rel="nofollow noreferrer">https://arxiv.org/abs/2209.06899</a><p>[8] <a href="https://openreview.net/pdf?id=eYlLlvzngu" rel="nofollow noreferrer">https://openreview.net/pdf?id=eYlLlvzngu</a><p>[9] <a href="https://www.norc.org/research/projects/gss.html" rel="nofollow noreferrer">https://www.norc.org/research/projects/gss.html</a>
Upvote: | 121 |
Title: Hello HN! I'm a creative technologist and recently decided to develop RealAboutInstagram, a replica of the current About page of Instagram replacing its content with their current harmful strategies used on the platform and the negative impacts of social media.<p>The information on the website is extracted from resources such as the Digital Minimalism book by Cal Newport, Ted Talks, and many others that can be found in the footer.<p>This is one of many projects for my career, and I appreciate anyone taking the time to read this and check out the website.<p>You can check out my other projects at <a href="https://santiagoaguirre.netlify.app/" rel="nofollow noreferrer">https://santiagoaguirre.netlify.app/</a><p>Thank you for your time!
Upvote: | 87 |
Title: Still very much a work in progress, but really wanted to share this even in it's early state. Had heaps of fun building it to learn more about WebRTC.
Upvote: | 157 |
Title: Hello!
I'm a passionate chess enthusiast. I couldn’t find a platform that showcases and allows one to learn the best moves ever played in international matches. I wanted to create a game that does this while being enjoyable, swift, and competitive. Took sometime, but now my game is out on both iOS and Android.<p>Android: https://play.google.com/store/apps/details?id=com.superhugestudios.chessblitz.chessgame.friends.chess<p>iOS: https://apps.apple.com/in/app/chess-blitz-multiplayer/id6446757472<p>As this is my debut project, I'd greatly appreciate suggestions on how I can enhance and make the game better. What features would you like to be added?<p>Currently, I am using a AI bot to play as the opponent whenever there's a low concurrency, and also added a feature to invite friends to play against. I made a custom algorithm to calculate the ELO scores of the player based on how one plays, and place a puzzle into a particular ELO bucket as well.<p>I am also looking for people to help maintain a site I built, The Times of Chess (https://thetimesofchess.com/), to populate strategies and best stories around Chess.<p>Would love to hear from you and get your feedback! If you'd like to join me in developing this game further, do let me know! :)
Upvote: | 43 |
Title: I want to do a trip to SF to sightsee and network.<p>But I want to keep costs down.<p>Hotel / AirBnB are the easiest options, but what about co-living spaces and hacker houses?<p>Ideally want to network with people working on llms / ml ops, so the community of a hacker house is appealing.<p>But never stayed in one before, so not sure what to expect!<p>Any help would be appreciated.
Upvote: | 41 |
Title: Hi HN, we’re Nate and Ty, co-founders of Continue, an open-source autopilot for software development built to be deeply customizable and continuously learn from development data. It consists of an extended language server and (to start) a VS Code extension.<p>Our GitHub is <a href="https://github.com/continuedev/continue">https://github.com/continuedev/continue</a>. You can watch a demo of Continue and download the extension at <a href="https://continue.dev">https://continue.dev</a><p>— — —<p>A growing number of developers are replacing Google + Stack Overflow with Large Language Models (LLMs) as their primary approach to get help, similar to how developers previously replaced reference manuals with Google + Stack Overflow.<p>However, existing LLM developer tools are cumbersome black boxes. Developers are stuck copy/pasting from ChatGPT and guessing what context Copilot uses to make a suggestion. As we use these products, we expose how we build software and give implicit feedback that is used to improve their LLMs, yet we don’t benefit from this data nor get to keep it.<p>The solution is to give developers what they need: <i>transparency, hackability,</i> and <i>control</i>. Every one of us should be able to reason about what’s going on, tinker, and have control over our own development data. This is why we created Continue.<p>— — —<p>At its most basic, Continue removes the need for copy/pasting from ChatGPT—instead, you collect context by highlighting and then ask questions in the sidebar or have an edit streamed directly to your editor.<p>But Continue also provides powerful tools for managing context. For example, type ‘@issue’ to quickly reference a GitHub issue as you are prompting the LLM, ‘@README.md’ to reference such a file, or ‘@google’ to include the results of a Google search.<p>And there’s a ton of room for further customization. Today, you can write your own<p>- slash commands (e.g. ‘/commit’ to write a summary and commit message for staged changes, ‘/docs’ to grab the contents of a file and update documentation pages that depend on it, ‘/ticket’ to generate a full-featured ticket with relevant files and high-level instructions from a short description)<p>- context sources (e.g. GitHub issues, Jira, local files, StackOverflow, documentation pages)<p>- templated system message (e.g. “Always give maximally concise answers. Adhere to the following style guide whenever writing code: {{ /Users/nate/repo/styleguide.md }}”)<p>- tools (e.g. add a file, run unit tests, build and watch for errors)<p>- policies (e.g. define a goal-oriented agent that works in a write code, run code, read errors, fix code, repeat loop)<p>Continue works with any LLM, including local models using ggml or open-source models hosted on your own cloud infrastructure, allowing you to remain 100% private. While OpenAI and Anthropic perform best today, we are excited to support the progress of open-source as it catches up (<a href="https://continue.dev/docs/customization#change-the-default-llm">https://continue.dev/docs/customization#change-the-default-l...</a>).<p>When you use Continue, you automatically collect data on how you build software. By default, this development data is saved to `.continue/dev_data` on your local machine. When combined with the code that you ultimately commit, it can be used to improve the LLM that you or your team use (if you allow).<p>You can read more about how development data is generated as a byproduct of LLM-aided development and why we believe that you should start collecting it now: <a href="https://medium.com/@continuedev/its-time-to-collect-data-on-how-you-build-software-197d12a020d5" rel="nofollow noreferrer">https://medium.com/@continuedev/its-time-to-collect-data-on-...</a><p>Continue has an Apache 2.0 license. We plan to make money by offering organizations a paid development data engine—a continuous feedback loop that ensures the LLMs always have fresh information and code in their preferred style.<p>— — —<p>We’d love for you to try out Continue and give us feedback! Let us know what you think in the comments : )
Upvote: | 298 |
Title: As per the title, anyone else getting continuous adverts from Youtube, making it impossible to watch anything on Youtube? Cant jump forward as it goes straight to an advert, even started watching youtube channels in the search list, but the youtube controls keep changing as well as well as dissappearing. One even managed to crash the computer, wiped out the screen, the keyboard, the mouse.<p>Keep getting an Eon advert over and over again, 20 seconds long, singing away something about time is running out, and then immediately after the Eon advert loads of junk adverts for spurious health conditions which all seem to be voiced over by the same bloke?
Upvote: | 41 |
Title: I built this library because langchain was too bloated and I needed a simple abstraction to call multiple LLM APIs. litellm has two functions - completion(), embedding()
Upvote: | 62 |
Title: I was having some problems with Slack that you can send message but you do not get a confirmation (so the text message gets greyed out) but it is actually sent. This is confusing because you can't delete the pending message and can't delete the duplicated messages after restarting the Slack client either.
Upvote: | 70 |
Title: I initially wrote "over-engineering" in the title, but changed it to "complexity" so as to not unnecessarily trigger some readers. I've been in the software industry for 20 years, and on HN for 10. This quarter was the first time I've seen a pattern of a lot of engineers yearning for old-fashioned simplicity while critiquing "modern" tech and practices, some of which were still on vogue just a year ago. E.g.<p>Imaginary Problems Are the Root of Bad Software https://news.ycombinator.com/item?id=36380711<p>Looming demise of the 10x developer – an era of enthusiast programmers is ending https://news.ycombinator.com/item?id=36719409<p>Every time you click this link, it will send you to a random Web 1.0 website https://news.ycombinator.com/item?id=36739920<p>Windows 9x and Word 9x at 800x600 resolution. Spacious. Comfy https://news.ycombinator.com/item?id=36682403<p>Why I Hate Frameworks (2005) https://news.ycombinator.com/item?id=36637655<p>DevOps Is Bullshit https://news.ycombinator.com/item?id=36354049<p>How I run my servers (2022) https://news.ycombinator.com/item?id=36744090<p>Look ma, no React: I recoded my portfolio site with vanilla everything https://news.ycombinator.com/item?id=36736576<p>Don't Take VC Funding – It Will Destroy Your Company https://news.ycombinator.com/item?id=36654960<p>Some of these submissions would have been received very differently in the past.<p>I was just entering the industry when the original OOP craze began, and for the longest time, thought I was stupid or crazy for not "getting" it and preferring to write simple functions. Same with cloud-native systems, reactive SPAs, container orchestration, SCRUM, and a bunch of other practices. In a way, this is refreshing. But at the same time, I worry that this may signal the end of the current era of explosive growth in software (both in terms of tech and money), and what we thought would be perpetual progress may in fact be a brief summer. Does anyone else feel this way?
Upvote: | 55 |
Title: Hey HN! We’re excited to show you Subset (<a href="https://subset.so" rel="nofollow noreferrer">https://subset.so</a>)! Subset is a spreadsheet on an infinite canvas where you can drop in spreadsheet building blocks and connect them together to easily build good looking analytical tools.<p>My cofounder and I both worked in finance before this. We quit our jobs and learned to code to build the spreadsheet we always wanted.<p>We got frustrated with how much time was wasted recreating the same thing from scratch or trying to extract something reusable from a previous spreadsheet. We wanted a spreadsheet with more composability and a community of building blocks to work off of.<p>Why infinite canvas? Reusability is not really a first class concept in a traditional endless spreadsheet grid [1]. A canvas felt more intuitive. You can communicate data flow, better understand inputs/outputs, and quickly make data presentable.<p>Subset is browser based and real-time multiplayer. It is free to use. We haven’t figured out a pricing model yet, but we imagine it’ll be a freemium SaaS model.<p>We’ve spent a lot of time on getting the core spreadsheet functionality as close to Sheets/Excel as possible, but there’s a lot more to do here. This list could be endless, but let us know if we need anything in particular.<p>Here’s a couple templates to explore:<p>Splitting a bill that includes sharing items - <a href="https://subset.so/templates/how-to-split-a-bill" rel="nofollow noreferrer">https://subset.so/templates/how-to-split-a-bill</a>
Calculating the cap rate and cash yield on real estate investments - <a href="https://subset.so/templates/quick-real-estate-deal-calculator" rel="nofollow noreferrer">https://subset.so/templates/quick-real-estate-deal-calculato...</a><p>We love spreadsheets and we think this combination of a canvas + reusable blocks has the potential to solve some of the biggest challenges with spreadsheets themselves.<p>Let us know what you think!<p>— AJ and Jason<p>[1] <a href="https://www.microsoft.com/en-us/research/uploads/prod/2020/04/joharizadeh_2020_gridlets.pdf" rel="nofollow noreferrer">https://www.microsoft.com/en-us/research/uploads/prod/2020/0...</a>
Upvote: | 49 |
Title: Hi HN! I'm Sai, the co-founder and CEO of PeerDB (<a href="https://www.peerdb.io/">https://www.peerdb.io/</a>), a Postgres-first data-movement platform that makes moving data in and out of Postgres fast and simple. PeerDB is free and open (<a href="https://github.com/PeerDB-io/peerdb">https://github.com/PeerDB-io/peerdb</a>) and we provide a Docker stack for users to try us out. Our repo is at <a href="https://github.com/PeerDB-io/peerdb">https://github.com/PeerDB-io/peerdb</a> and there’s a 5-minute quickstart here: <a href="https://docs.peerdb.io/quickstart">https://docs.peerdb.io/quickstart</a>.<p>For the past 8 years, working at Microsoft on Postgres on Azure, and before that at Citus Data, I’ve worked closely with customers running Postgres at the heart of their data stack, storing anywhere from 10s of GB of data to 10s of TB.<p>This was when I got exposed to the challenges customers faced when moving data in and out of Postgres. Usually they would try existing ETL tools, fail, and decide to build in-house solutions. Common issues with these tools included painfully slow syncs - syncing 100s of GB of data took days; flaky and unreliable - frequent crashes, loss of data precision on target etc., and; feature-limited - lack of configurability, unsupported data types and so on.<p>I remember a specific scenario where a tool didn’t support something as simple as the Postgres’ COPY command to ingest data. This would have improved the throughput by orders of magnitude. We (customer and me) reached out to that company to request them to add this feature. They couldn’t prioritize this feature because it wasn’t very easy - their tech stack was designed to support 100s of connectors rather than supporting a native Postgres feature.<p>After multiple such occurrences, I thought, why not build a tool specialized for Postgres, making the lives of many Postgres users easier. I reached out to my long-time buddy Kaushik, who was building operating systems at Google and had led data teams at Safegraph and Palantir. We spent a few weeks building an MVP that streamed data in real-time from Postgres to BigQuery. It was 10 times faster than existing tools and maintained data freshness of less than 30 seconds. We realized that there were many Postgres native and infrastructural optimizations we could do to provide a rich data-movement experience for Postgres users. This is when we decided to start PeerDB!<p>We started with two main use cases: Real-time Change Data Capture from Postgres (demo: <a href="https://docs.peerdb.io/usecases/realtime-cdc#demo">https://docs.peerdb.io/usecases/realtime-cdc#demo</a>) and Real-time Streaming of query results from Postgres (demo: <a href="https://docs.peerdb.io/usecases/realtime-streaming-of-query-results#demo">https://docs.peerdb.io/usecases/realtime-streaming-of-query-...</a>). The 2nd demo shows PeerDB streaming a table with 100M rows from Postgres to Snowflake.<p>We implement multiple optimizations to provide a fast, reliable, feature-rich experience. For performance, we can parallelize the initial load of a large table, still ensuring consistency. Syncing 100s of GB goes from days to minutes. We do this by logically partitioning the table based on internal tuple identifiers (CTID) and parallelly streaming those partitions (inspired by this DuckDB blog - <a href="https://duckdb.org/2022/09/30/postgres-scanner.html#parallelization" rel="nofollow noreferrer">https://duckdb.org/2022/09/30/postgres-scanner.html#parallel...</a>)<p>For CDC, we don’t use Debezium, rather handle replication more natively—reading the slot, replicating the changes, keeping state etc. We made this choice mainly for flexibility. Staying native helps us use existing and future Postgres enhancements more effectively. For example, if the order of rows across tables on the target is not important, we can parallelize reading of a single slot across multiple tables and improve performance. Our architecture is designed for real-time syncs, which enables data-freshness of a few 10s of seconds even at large throughputs (10k+ tps).<p>We have fault tolerance mechanisms for reliability (<a href="https://blog.peerdb.io/using-temporal-to-scale-data-synchronization-at-peerdb">https://blog.peerdb.io/using-temporal-to-scale-data-synchron...</a>) and support multiple features including log-based (CDC) / query based streaming, efficient syncing of tables with large (TOAST) columns, configurable batching and parallelism to prevent OOMs and crashes etc.<p>For usability - we provide a Postgres compatible SQL layer for data-movement. This makes the life of data engineers much easier. They can develop pipelines using a framework they are familiar with, without needing to deal with custom UIs and REST APIs. They can use Postgres' 100s of integrations to build and manage ETL. We extend Postgres' SQL grammar with a few new intuitive SQL commands to enable real-time data streaming across stores. Because of this, we were able to add dbt integration via Dagster (in private preview) in a few hours! We expect data-engineers to unravel similar integrations with PeerDB easily, and plan to make this grammar richer as we evolve.<p>PeerDB consists of the following components to handle data replication: (1) PeerDB Server uses the pgwire protocol to mimic a PostgreSQL server, responsible for query routing and generating gRPC requests to the Flow API. It relies on AST analysis to make informed decisions on routing. (2) Flow API: an API layer that deals with gRPC commands, orchestrating the data sync operations; (3) Flow Workers execute the data read-write operations from the source to the destination. Built to scale horizontally, they interact with Temporal for increased resilience. The types of data replication supported include CDC streaming replication and query-based batch replication. Workers do all of the heavy lifting, and have data store specific optimizations.<p>Currently we support 6 target data stores (BigQuery, Snowflake, Postgres, S3, Kafka etc) for data movement from Postgres. This doc captures the current status of the connectors: <a href="https://docs.peerdb.io/sql/commands/supported-connectors">https://docs.peerdb.io/sql/commands/supported-connectors</a>.<p>As we spoke to more customers, we realized that getting data into PostgreSQL at scale is equally important and hard. For example one of our customers wants to periodically sync data in multiple SQL Server instances (running on the edge) to their centralized Postgres database. Requests for Oracle to Postgres migrations are also common. So now we’re also supporting source data stores with Postgres as the target (currently SQL Server and Postgres itself, with more to come).<p>We are actively working with customers to onboard them to our self-hosted enterprise offering. Our fully hosted offering on the cloud is in private preview. We haven’t yet decided on the pricing. One common concern we’ve heard from customers is that existing tools are expensive and charge based on the amount of data transferred. To address this, we are considering a more transparent way of pricing—for example, pricing based on provisioned hardware (cpu, memory, disk). We’re open for feedback on this!<p>Check out our github repo - <a href="https://github.com/PeerDB-io/peerdb">https://github.com/PeerDB-io/peerdb</a> and go ahead and give it a spin (5-minute quickstart <a href="https://docs.peerdb.io/quickstart">https://docs.peerdb.io/quickstart</a>).<p>We want to provide the world’s best data-movement experience for Postgres. We would love to get your feedback on product experience, our thesis and anything else that comes to your mind. It would be super useful for us. Thank you!
Upvote: | 261 |
Title: I got laid off at the start of the year, and ever since then, I've been applying constantly but have only gotten one interview. Before being laid off, I held a job as a front-end dev for the previous 5 and a half years.<p>I've had my resume looked at by three different services (TopResume, Indeed, Levels.fyi) and am currently subscribed to Resume Worded, which scores my resume. Despite all these efforts, I keep receiving rejection emails.<p>So, I just wanted to reach out and see if anyone else has had any similar experiences with applying for jobs.
Upvote: | 711 |
Title: There were quite a lot of stories, mostly negative, about Tesla recently and it seems that most are actively removed from first page - todays one about Tesla range estimation “issues” is an example.<p>While removing articles about Musk and Tesla stock price seems like a good idea, the other ones (todays one, the one about German leak a few weeks ago) seem much more interesting to discuss.
Upvote: | 58 |
Title: I'm submitting this before going to bed, so by the time it hits front page (if it does at all) I'll likely be asleep, so I won't be able to answer questions in time.
Upvote: | 67 |
Title: I posted here a while back with this app I'd been working on. In short:<p>It allows you to talk to Spotify's recommendations API to specify musical features you'd like recommendations similar to.<p>It's loads of fun to mess around with, and I've given it an overhaul to include many more dimensions ito. musical features, as well as the ability to log in with Spotify, and save things you find to a playlist. You don't need a subscription to Spotify to use the app either.<p>Thanks to @Mockapapella for the tagline when I posted the first version: <a href="https://news.ycombinator.com/item?id=35907215">https://news.ycombinator.com/item?id=35907215</a>
Upvote: | 69 |
Title: Hello HN! It's been about 9 months since we've done one of these and it seems like time for another. Previous threads we've done: <a href="https://news.ycombinator.com/submitted?id=proberts">https://news.ycombinator.com/submitted?id=proberts</a>.<p>I'll be here for the next 2-3 hours and then after a break of about 1 hour for another 2-3 hours. As usual, there are many possible topics and I'll be guided by whatever you're concerned with. Please remember that I can't provide legal advice on specific cases for obvious liability reasons because I won't have access to all the facts. Please stick to a factual discussion in your questions and comments and I'll try to do the same in my answers!
Upvote: | 338 |
Subsets and Splits