- As much as they tell you it's about showing how you "think", that's only true if you get to the solution with very little help\n- It's mostly luck that you'll get questions that you've seen before while practicing\n- It seems like most companies are using Leetcode medium/hard now for anything that isn't a new grad\n- You're racing against the clock to find the trick to the question
First point, many times I would get stumped and interviewer offers hint. The moment the interviewer needs to offer you a hint, it's likely you're done. Unfortunately the bar seems to be really high in that regard. This might be a bit of luck, where you get a compassionate enough interviewer who understands that nerves always come into play, but who knows.
Second point, it really builds from the first point. The more problems you do, the easier it is to pattern match, and the more likely you are to not get stumped or need hints.
Third point, I think that things like CCTCI and LeetCode have essentially forced companies to raise the bar. Now everyone will grind leetcode to get through, so it creates a feedback cycle where interview committees think that their questions might be too easy, so they need to raise the bar.
Finally, you only really have about 15-20 mins to solve a problem in a given 45 min interview. You lose about 5-10 mins of chat with the interviewer. You're left with 30 mins, of which you probably need to think through the problem for about 10-15 mins to find the optimal solution. DON'T bother implementing inefficient solutions. Others will have grinded leetcode and will "figure" out the optimal solution in a few minutes. If you can't do same, you've likely lost the interview at that point unfortunately.
Beyond that, any tips here to help a fellow developer? Any inside perspectives that suggest what I'm saying is completely biased and wrong?"},"title":{"matchLevel":"none","matchedWords":[],"value":"Ask HN: How to Improve at Coding Interviews"}},"_tags":["story","author_rafiki6","story_23478957","ask_hn"],"author":"rafiki6","created_at":"2020-06-10T16:28:48Z","created_at_i":1591806528,"num_comments":0,"objectID":"23478957","points":4,"story_id":23478957,"story_text":"I've been actively struggling with FAANG interviews. Came to a few realizations I wanted to share:
- As much as they tell you it's about showing how you "think", that's only true if you get to the solution with very little help\n- It's mostly luck that you'll get questions that you've seen before while practicing\n- It seems like most companies are using Leetcode medium/hard now for anything that isn't a new grad\n- You're racing against the clock to find the trick to the question
First point, many times I would get stumped and interviewer offers hint. The moment the interviewer needs to offer you a hint, it's likely you're done. Unfortunately the bar seems to be really high in that regard. This might be a bit of luck, where you get a compassionate enough interviewer who understands that nerves always come into play, but who knows.
Second point, it really builds from the first point. The more problems you do, the easier it is to pattern match, and the more likely you are to not get stumped or need hints.
Third point, I think that things like CCTCI and LeetCode have essentially forced companies to raise the bar. Now everyone will grind leetcode to get through, so it creates a feedback cycle where interview committees think that their questions might be too easy, so they need to raise the bar.
Finally, you only really have about 15-20 mins to solve a problem in a given 45 min interview. You lose about 5-10 mins of chat with the interviewer. You're left with 30 mins, of which you probably need to think through the problem for about 10-15 mins to find the optimal solution. DON'T bother implementing inefficient solutions. Others will have grinded leetcode and will "figure" out the optimal solution in a few minutes. If you can't do same, you've likely lost the interview at that point unfortunately.
Beyond that, any tips here to help a fellow developer? Any inside perspectives that suggest what I'm saying is completely biased and wrong?","title":"Ask HN: How to Improve at Coding Interviews","updated_at":"2024-09-20T06:26:22Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"Nimish29"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"So I watch a lot of educational YouTube videos for work. I'd always find myself wanting to share the good stuff with my network on LinkedIn or tweet about it.
But here's the thing that drove me crazy: I'd spend ages trying to summarize a 20-minute video into a decent post. Copy some quotes, try to capture the main points, format it nicely... and it would take me like 30 minutes just to write one post.
Then I started thinking - if this is painful for me as just a consumer, imagine how much worse it is for actual content creators and YouTubers + social media managers. They're making all this great content on YouTube but then have to spend hours manually converting it for LinkedIn, Twitter, etc. just to reach different audiences.
I figured there had to be a better way, so I built this thing called ThinkYT.
You just paste a YouTube URL and it turns into posts for different platforms. No more sitting there rewinding videos trying to get the exact quote or struggling to remember what the third point was.
What used to take me 30+ minutes now takes like 2 minutes.
I would love to hear your feedback!"},"title":{"matchLevel":"none","matchedWords":[],"value":"Show HN: I Built this to repurpose YouTube Videos into posts to save time"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://thinkyt.com"}},"_tags":["story","author_Nimish29","story_44969201","show_hn"],"author":"Nimish29","children":[44969210],"created_at":"2025-08-21T05:04:13Z","created_at_i":1755752653,"num_comments":1,"objectID":"44969201","points":1,"story_id":44969201,"story_text":"So I watch a lot of educational YouTube videos for work. I'd always find myself wanting to share the good stuff with my network on LinkedIn or tweet about it.
But here's the thing that drove me crazy: I'd spend ages trying to summarize a 20-minute video into a decent post. Copy some quotes, try to capture the main points, format it nicely... and it would take me like 30 minutes just to write one post.
Then I started thinking - if this is painful for me as just a consumer, imagine how much worse it is for actual content creators and YouTubers + social media managers. They're making all this great content on YouTube but then have to spend hours manually converting it for LinkedIn, Twitter, etc. just to reach different audiences.
I figured there had to be a better way, so I built this thing called ThinkYT.
You just paste a YouTube URL and it turns into posts for different platforms. No more sitting there rewinding videos trying to get the exact quote or struggling to remember what the third point was.
What used to take me 30+ minutes now takes like 2 minutes.
I would love to hear your feedback!","title":"Show HN: I Built this to repurpose YouTube Videos into posts to save time","updated_at":"2026-03-05T22:32:17Z","url":"https://thinkyt.com"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"etairl"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"Hi HN,\nI've been running this on my own dependency tree for the past few months. Probus is a vulnerability scanner that uses three agents. One picks the files worth deep-scanning. One writes raw findings. The third reads the code on its own and rejects any finding that doesn't have a real attack vector.\nWhile building it I pointed it at projects I use day to day. Bugs that came out of this and got reported as PRs:
n8n: password-reset JWTs being logged at debug level (n8n-io/n8n#29405)\nVercel AI SDK: role: "system" injection in createAgentUIStream, a runtime schema bypass in ToolLoopAgent, and a prototype-property collision in getMediaTypeFromUrl (vercel/ai#14749, #14750 merged, #14751 merged)\nLangGraph.js: NoSQL injection in MongoDBSaver via unvalidated thread_id / checkpoint_ns / checkpoint_id types (langchain-ai/langgraphjs#2353)\nbrowser-use: path traversal in remote-fetched templates.json fields (browser-use/browser-use#4777)\nHaystack: SSRF and arbitrary file read via unrestricted OpenAPI $ref resolution, path traversal in the image converter, and unbounded HTTP body reads in LinkContentFetcher (deepset-ai/haystack#11226, #11228, #11229)
The false positive rate got low enough that I'd rather have other people running it than keep it private, so it's now public under Apache 2.0.\nHow it works:
Analyst (1 LLM call): reads the repo and picks 50 to 500 files to deep-scan based on entry points, third-party surface, and dangerous sinks.\nResearcher (per file): walks call chains and writes raw findings.\nQA (per file): re-reads the code against each claim with no access to the researcher's reasoning, and rejects anything that doesn't have a real attack vector. Keeping the QA agent isolated from the researcher is what got noise under control. If it sees the researcher's reasoning, it just agrees with it.
Each agent runs in its own query() session through the Claude Agent SDK with a filesystem sandbox scoped to the target repo.\nCost is tuned for open models. About $0.50 per file with Qwen 3.6 plus DeepSeek v4 Pro on OpenRouter. OpenAI is around 2.5x that. Anthropic is around 10x.\nnpm install -g probus\nprobus scan ./my-app\nThings I'd like feedback on:
The QA prompt took the most iteration. Happy to walk through it if anyone is working on similar verifier-agent patterns.\nI want to publish a public benchmark against a vulhub-style corpus. Suggestions on which repos to run it against would be helpful.\nThe analyst step is a single LLM call right now. On large monorepos it sometimes misses things. Thinking about a hierarchical version.
https://github.com/etairl/Probus"},"title":{"matchLevel":"none","matchedWords":[],"value":"Show HN: Probus, AI vuln scanner (PRs merged in Vercel AI SDK, n8n, LangGraph)"}},"_tags":["story","author_etairl","story_48022411","show_hn"],"author":"etairl","created_at":"2026-05-05T13:39:05Z","created_at_i":1777988345,"num_comments":0,"objectID":"48022411","points":1,"story_id":48022411,"story_text":"Hi HN,\nI've been running this on my own dependency tree for the past few months. Probus is a vulnerability scanner that uses three agents. One picks the files worth deep-scanning. One writes raw findings. The third reads the code on its own and rejects any finding that doesn't have a real attack vector.\nWhile building it I pointed it at projects I use day to day. Bugs that came out of this and got reported as PRs:
n8n: password-reset JWTs being logged at debug level (n8n-io/n8n#29405)\nVercel AI SDK: role: "system" injection in createAgentUIStream, a runtime schema bypass in ToolLoopAgent, and a prototype-property collision in getMediaTypeFromUrl (vercel/ai#14749, #14750 merged, #14751 merged)\nLangGraph.js: NoSQL injection in MongoDBSaver via unvalidated thread_id / checkpoint_ns / checkpoint_id types (langchain-ai/langgraphjs#2353)\nbrowser-use: path traversal in remote-fetched templates.json fields (browser-use/browser-use#4777)\nHaystack: SSRF and arbitrary file read via unrestricted OpenAPI $ref resolution, path traversal in the image converter, and unbounded HTTP body reads in LinkContentFetcher (deepset-ai/haystack#11226, #11228, #11229)
The false positive rate got low enough that I'd rather have other people running it than keep it private, so it's now public under Apache 2.0.\nHow it works:
Analyst (1 LLM call): reads the repo and picks 50 to 500 files to deep-scan based on entry points, third-party surface, and dangerous sinks.\nResearcher (per file): walks call chains and writes raw findings.\nQA (per file): re-reads the code against each claim with no access to the researcher's reasoning, and rejects anything that doesn't have a real attack vector. Keeping the QA agent isolated from the researcher is what got noise under control. If it sees the researcher's reasoning, it just agrees with it.
Each agent runs in its own query() session through the Claude Agent SDK with a filesystem sandbox scoped to the target repo.\nCost is tuned for open models. About $0.50 per file with Qwen 3.6 plus DeepSeek v4 Pro on OpenRouter. OpenAI is around 2.5x that. Anthropic is around 10x.\nnpm install -g probus\nprobus scan ./my-app\nThings I'd like feedback on:
The QA prompt took the most iteration. Happy to walk through it if anyone is working on similar verifier-agent patterns.\nI want to publish a public benchmark against a vulhub-style corpus. Suggestions on which repos to run it against would be helpful.\nThe analyst step is a single LLM call right now. On large monorepos it sometimes misses things. Thinking about a hierarchical version.
https://github.com/etairl/Probus","title":"Show HN: Probus, AI vuln scanner (PRs merged in Vercel AI SDK, n8n, LangGraph)","updated_at":"2026-05-05T13:41:53Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"austinchankl"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"I am a developer at startup company in Malaysia. Recently I've been tasked to study our website Google Analytics data.
First pain point: I tried the GA dashboard and it was diffucult to navigate and find the data I need!
Second pain point: I'm not familiar with analyzing data, so I turn to ChatGPT
Third pain point: ChatGPT can't access my Google Analytics data, I have to export and upload the data manually every times.
So to bridge the gap, I'm experimenting and building https://rumius.com (in progress), essentially an AI Data Analyst Agent that can access data in realtime. I believe it's something useful to product managers too.
Do you have the same problem? Let's discuss."},"title":{"matchLevel":"none","matchedWords":[],"value":"I wish ChatGPT can analyze Google Analytics data in realtime"}},"_tags":["story","author_austinchankl","story_42282051","ask_hn"],"author":"austinchankl","created_at":"2024-11-30T15:18:34Z","created_at_i":1732979914,"num_comments":0,"objectID":"42282051","points":1,"story_id":42282051,"story_text":"I am a developer at startup company in Malaysia. Recently I've been tasked to study our website Google Analytics data.
First pain point: I tried the GA dashboard and it was diffucult to navigate and find the data I need!
Second pain point: I'm not familiar with analyzing data, so I turn to ChatGPT
Third pain point: ChatGPT can't access my Google Analytics data, I have to export and upload the data manually every times.
So to bridge the gap, I'm experimenting and building https://rumius.com (in progress), essentially an AI Data Analyst Agent that can access data in realtime. I believe it's something useful to product managers too.
Do you have the same problem? Let's discuss.","title":"I wish ChatGPT can analyze Google Analytics data in realtime","updated_at":"2024-11-30T15:20:17Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"loverobots"},"title":{"matchLevel":"none","matchedWords":[],"value":"Yahoo CEO might have lied about having CS degree"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"http://www.marketwatch.com/story/third-point-llc-letter-to-yahoo-board-of-directors-regarding-discovery-of-discrepancies-in-educational-records-of-ceo-scott-thompson-and-director-patti-hart-2012-05-03"}},"_tags":["story","author_loverobots","story_3925359"],"author":"loverobots","children":[3925461,3925481,3925497,3925515,3925528,3925545,3925574,3925601,3925618,3925623,3925658,3925754,3925812,3925842,3925888,3925922,3925928,3925982,3926043,3926154,3926328,3926554,3926557,3926583,3926607,3926794,3926891,3927005,3927129,3927324,3928698,3928802,3930236,3930458],"created_at":"2012-05-03T19:34:26Z","created_at_i":1336073666,"num_comments":205,"objectID":"3925359","points":282,"story_id":3925359,"title":"Yahoo CEO might have lied about having CS degree","updated_at":"2023-09-06T21:35:19Z","url":"http://www.marketwatch.com/story/third-point-llc-letter-to-yahoo-board-of-directors-regarding-discovery-of-discrepancies-in-educational-records-of-ceo-scott-thompson-and-director-patti-hart-2012-05-03"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"iProject"},"title":{"matchLevel":"none","matchedWords":[],"value":"Yahoo CEO apologizes in memo, board meets"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"http://www.reuters.com/article/2012/05/08/net-us-yahoo-thirdpoint-idUSBRE8460SY20120508"}},"_tags":["story","author_iProject","story_3943390"],"author":"iProject","children":[3943513,3943530,3943634,3943714,3943737],"created_at":"2012-05-08T13:05:43Z","created_at_i":1336482343,"num_comments":13,"objectID":"3943390","points":18,"story_id":3943390,"title":"Yahoo CEO apologizes in memo, board meets","updated_at":"2024-09-19T18:34:58Z","url":"http://www.reuters.com/article/2012/05/08/net-us-yahoo-thirdpoint-idUSBRE8460SY20120508"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"ValentineC"},"title":{"matchLevel":"none","matchedWords":[],"value":"A Private Club's Public Problems"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"https://www.wsj.com/business/sohohouse-dan-loeb-third-point-private-clubs-very-public-problems-99d35251"}},"_tags":["story","author_ValentineC","story_43064871"],"author":"ValentineC","children":[43064872,43064959],"created_at":"2025-02-16T02:53:36Z","created_at_i":1739674416,"num_comments":2,"objectID":"43064871","points":2,"story_id":43064871,"title":"A Private Club's Public Problems","updated_at":"2025-02-16T03:12:38Z","url":"https://www.wsj.com/business/sohohouse-dan-loeb-third-point-private-clubs-very-public-problems-99d35251"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"benackles"},"story_text":{"matchLevel":"none","matchedWords":[],"value":""},"title":{"matchLevel":"none","matchedWords":[],"value":"Investor Demands Yahoo Fire C.E.O."},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"http://dealbook.nytimes.com/2012/05/04/third-point-demands-yahoo-c-e-o-be-fired-by-monday/"}},"_tags":["story","author_benackles","story_3930924"],"author":"benackles","created_at":"2012-05-05T00:33:42Z","created_at_i":1336178022,"num_comments":0,"objectID":"3930924","points":2,"story_id":3930924,"story_text":"","title":"Investor Demands Yahoo Fire C.E.O.","updated_at":"2024-09-19T18:33:41Z","url":"http://dealbook.nytimes.com/2012/05/04/third-point-demands-yahoo-c-e-o-be-fired-by-monday/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"finphil"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"Robinhood\u2019s third outage may point to deeper problems in its tech stack"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"https://techcrunch.com/2020/03/10/robinhoods-third-outage-may-point-to-deeper-problems-in-its-tech-stack/"}},"_tags":["story","author_finphil","story_22536578"],"author":"finphil","children":[22536707,22536708,22536709,22536724,22536736,22536737,22536740,22536748,22536756,22536865,22536870,22536876],"created_at":"2020-03-10T16:00:46Z","created_at_i":1583856046,"num_comments":50,"objectID":"22536578","points":37,"story_id":22536578,"title":"Robinhood\u2019s third outage may point to deeper problems in its tech stack","updated_at":"2024-09-20T05:50:05Z","url":"https://techcrunch.com/2020/03/10/robinhoods-third-outage-may-point-to-deeper-problems-in-its-tech-stack/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"sneek_"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"Hey HN, my name is James and I founded Payload (https://payloadcms.com/) with two close colleagues, Dan and Elliot. We're a dev-first headless CMS [1] that's half app framework and half CMS\u2014we're closing the gap between the two. You can check out our demo here: https://demo.payloadcms.com.
Imagine you're going to build a new SaaS app. Would you think of building it on a headless CMS? Probably not. To devs, "content management system" is usually a swear word. If a team of engineers gets assigned a CMS project, it's less than thrilling. Engineers want to avoid roadblocks, write code, and build things they're proud of\u2014but existing CMS's get in the way of that left and right with their third-party integrations, point-and-click schema designers, code generation, etc.
Rather, you'd build your backend on an app framework like Django, Laravel, etc., for good reasons: ownership over the backend, better access control, customizable auth patterns, etc. Typically, headless CMS are super limiting; you'll end up fighting the platform more than having it help. But, with app frameworks, you're often left to roll your own admin UI, and that takes time. Not to mention building CRUD UI gets old quick after you do it a few times.
That\u2019s where a headless CMS could shine, because they instantly give you admin UI that non-technical teams can use to manage digital products. That saves a ton of UI dev time\u2014 but without an extensible API, headless CMS's are far too limiting. They're designed for marketing teams, which usually only need the generic basics: log in, create a draft, preview the draft, publish the content. Go back and update some pages. Define editor roles and localize content. If you need more than that, you'll soon be out of luck.
Payload is different because we treat developers as first-class citizens. We provide the best of both ends: a powerful and extensible API and a fully customizable admin UI out-of-the-box. All with a developer experience that we obsess over, because we want it ourselves.
Payload is code-first, which allows us to get a lot of things right. We give you what you need, then step back and let you build what you want in TypeScript. You'll understand how your CMS works because you will have written it exactly how you want it. Version control your schema and use your own Express server. Completely control the Admin panel by using your own React components. Swap out fields or even entire views with ease. Use your data however and wherever you need thanks to auto-generated, yet fully extensible REST, GraphQL, and Local Node APIs.
Since it uses your own Express server, you can open up your own endpoints alongside what Payload does. In fact, you can extend just about everything that Payload does. It's MIT and open-source, fully self-hosted, comes with GraphQL and REST APIs, and completely customizable.
We realized the need for Payload while we were building the corporate website for Klarna. The Klarna engineers we were working with were among the best in the world, and while they evaluated headless CMS options, they saw restrictions in how all of the normal contenders "black-box" away the API. They wanted to build their CMS, deploy it on their own infrastructure, and truly "own" their CMS. They fell back to using WordPress. When that happened, Klarna inadvertently shined a spotlight on the CMS market and pointed out a significant void in proper code-based, developer-first CMS. There was no one to give them the developer experience they needed. That's what got us started working on this.
It might seem like a CMS is just a wrapper around a database with a nice UI to show different field types\u2014but in reality, it's a lot more complex than that. We obsessed for years around how to build a proper API that minimizes breaking changes, but still exposes a simple way to extend everything. When you start to introduce things like field-based access control, field-based conditional logic, localization, versions, drafts, and autosave, the task becomes a lot more daunting. Doing it right requires a significant development investment\u2014especially if you want it to perform at scale in addition to removing roadblocks at dev time.
It seems like every day, a new headless CMS pops up. But when you filter down to those that are completely self-hosted, the options quickly dwindle. And then when you remove the confused point-and-click "no-code" (argh!) GUI nature of the existing options, the options narrow to one: Payload.
Our users have built quite a diverse set of apps on Payload. We've seen a virtual events platform, a broadcast platform, SaaS apps of all shapes and sizes, video games, and an Uber-like snow plow service! There are over 1,000 projects in production as of last week, and we can't wait to see more.
Open source has been incredibly helpful. We've gotten significant PRs and our community has gone above and beyond in their contributions. We did not anticipate the level of skill and involvement that we are seeing daily from our community.
Our business model is based on two things:
1. Enterprise features like SSO, audit logs, publication workflows, and translation workflows. Of course, as Payload is open-source, you can build these functions yourself, but enterprises are opting to pay for our official functionality and SLAs rather than rolling it themselves.
2. Cloud hosting. Now that Payload 1.0 is released and ready for production after more than two years of development and dogfooding, we've shifted focus to building a deployment platform for Payload that will deliver permanent file storage, database, API layer, and CI. It will be the easiest way to deploy Payload, but not mandatory to use\u2014much like the NextJS and Vercel model.
You can get started in one line by running `npx create-payload-app` or you can try out our public demo at https://demo.payloadcms.com. The code for the demo is at https://github.com/payloadcms/public-demo.
We would love to hear your feedback. If we don't have something, we'll build it. If there's a sticky spot in the DX (developer experience), we\u2019ll fix it. Looking forward to hearing what you think\u2014and thank you!
[1] Quick refresher: CMS stands for "content management system" and headless just means API-based, with no restrictions over where you use the content on the frontend."},"title":{"matchLevel":"none","matchedWords":[],"value":"Launch HN: Payload (YC S22) \u2013 Headless CMS for Developers"}},"_tags":["story","author_sneek_","story_32665325","launch_hn"],"author":"sneek_","children":[32665486,32665489,32665509,32665560,32665776,32665788,32665853,32665929,32665939,32665968,32666117,32666166,32666253,32666783,32666914,32666942,32666969,32667034,32667040,32667090,32667135,32667157,32667284,32667627,32667738,32667976,32668065,32668166,32668329,32668359,32668501,32668645,32669568,32669588,32670347,32670363,32670644,32671833,32672707,32672993,32673341,32673394,32674658,32675029,32675227,32678479,32683303,32689510],"created_at":"2022-08-31T17:14:58Z","created_at_i":1661966098,"num_comments":137,"objectID":"32665325","points":242,"story_id":32665325,"story_text":"Hey HN, my name is James and I founded Payload (https://payloadcms.com/) with two close colleagues, Dan and Elliot. We're a dev-first headless CMS [1] that's half app framework and half CMS\u2014we're closing the gap between the two. You can check out our demo here: https://demo.payloadcms.com.
Imagine you're going to build a new SaaS app. Would you think of building it on a headless CMS? Probably not. To devs, "content management system" is usually a swear word. If a team of engineers gets assigned a CMS project, it's less than thrilling. Engineers want to avoid roadblocks, write code, and build things they're proud of\u2014but existing CMS's get in the way of that left and right with their third-party integrations, point-and-click schema designers, code generation, etc.
Rather, you'd build your backend on an app framework like Django, Laravel, etc., for good reasons: ownership over the backend, better access control, customizable auth patterns, etc. Typically, headless CMS are super limiting; you'll end up fighting the platform more than having it help. But, with app frameworks, you're often left to roll your own admin UI, and that takes time. Not to mention building CRUD UI gets old quick after you do it a few times.
That\u2019s where a headless CMS could shine, because they instantly give you admin UI that non-technical teams can use to manage digital products. That saves a ton of UI dev time\u2014 but without an extensible API, headless CMS's are far too limiting. They're designed for marketing teams, which usually only need the generic basics: log in, create a draft, preview the draft, publish the content. Go back and update some pages. Define editor roles and localize content. If you need more than that, you'll soon be out of luck.
Payload is different because we treat developers as first-class citizens. We provide the best of both ends: a powerful and extensible API and a fully customizable admin UI out-of-the-box. All with a developer experience that we obsess over, because we want it ourselves.
Payload is code-first, which allows us to get a lot of things right. We give you what you need, then step back and let you build what you want in TypeScript. You'll understand how your CMS works because you will have written it exactly how you want it. Version control your schema and use your own Express server. Completely control the Admin panel by using your own React components. Swap out fields or even entire views with ease. Use your data however and wherever you need thanks to auto-generated, yet fully extensible REST, GraphQL, and Local Node APIs.
Since it uses your own Express server, you can open up your own endpoints alongside what Payload does. In fact, you can extend just about everything that Payload does. It's MIT and open-source, fully self-hosted, comes with GraphQL and REST APIs, and completely customizable.
We realized the need for Payload while we were building the corporate website for Klarna. The Klarna engineers we were working with were among the best in the world, and while they evaluated headless CMS options, they saw restrictions in how all of the normal contenders "black-box" away the API. They wanted to build their CMS, deploy it on their own infrastructure, and truly "own" their CMS. They fell back to using WordPress. When that happened, Klarna inadvertently shined a spotlight on the CMS market and pointed out a significant void in proper code-based, developer-first CMS. There was no one to give them the developer experience they needed. That's what got us started working on this.
It might seem like a CMS is just a wrapper around a database with a nice UI to show different field types\u2014but in reality, it's a lot more complex than that. We obsessed for years around how to build a proper API that minimizes breaking changes, but still exposes a simple way to extend everything. When you start to introduce things like field-based access control, field-based conditional logic, localization, versions, drafts, and autosave, the task becomes a lot more daunting. Doing it right requires a significant development investment\u2014especially if you want it to perform at scale in addition to removing roadblocks at dev time.
It seems like every day, a new headless CMS pops up. But when you filter down to those that are completely self-hosted, the options quickly dwindle. And then when you remove the confused point-and-click "no-code" (argh!) GUI nature of the existing options, the options narrow to one: Payload.
Our users have built quite a diverse set of apps on Payload. We've seen a virtual events platform, a broadcast platform, SaaS apps of all shapes and sizes, video games, and an Uber-like snow plow service! There are over 1,000 projects in production as of last week, and we can't wait to see more.
Open source has been incredibly helpful. We've gotten significant PRs and our community has gone above and beyond in their contributions. We did not anticipate the level of skill and involvement that we are seeing daily from our community.
Our business model is based on two things:
1. Enterprise features like SSO, audit logs, publication workflows, and translation workflows. Of course, as Payload is open-source, you can build these functions yourself, but enterprises are opting to pay for our official functionality and SLAs rather than rolling it themselves.
2. Cloud hosting. Now that Payload 1.0 is released and ready for production after more than two years of development and dogfooding, we've shifted focus to building a deployment platform for Payload that will deliver permanent file storage, database, API layer, and CI. It will be the easiest way to deploy Payload, but not mandatory to use\u2014much like the NextJS and Vercel model.
You can get started in one line by running `npx create-payload-app` or you can try out our public demo at https://demo.payloadcms.com. The code for the demo is at https://github.com/payloadcms/public-demo.
We would love to hear your feedback. If we don't have something, we'll build it. If there's a sticky spot in the DX (developer experience), we\u2019ll fix it. Looking forward to hearing what you think\u2014and thank you!
[1] Quick refresher: CMS stands for "content management system" and headless just means API-based, with no restrictions over where you use the content on the frontend.","title":"Launch HN: Payload (YC S22) \u2013 Headless CMS for Developers","updated_at":"2026-02-16T00:51:13Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"mauliknshah"},"title":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["third"],"value":"Google launches the third version of its A.I. chips, an alternative to Nvidia's"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"https://www.cnbc.com/2018/05/08/google-launches-tpu-3-point-0-third-version-of-ai-chips.html"}},"_tags":["story","author_mauliknshah","story_17023837"],"author":"mauliknshah","children":[17024151,17024157,17024206,17024808],"created_at":"2018-05-08T19:13:04Z","created_at_i":1525806784,"num_comments":16,"objectID":"17023837","points":62,"story_id":17023837,"title":"Google launches the third version of its A.I. chips, an alternative to Nvidia's","updated_at":"2024-09-20T02:23:03Z","url":"https://www.cnbc.com/2018/05/08/google-launches-tpu-3-point-0-third-version-of-ai-chips.html"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"scrlk"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"Federal Reserve to increase interest rates by 75 basis points for the third time"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://www.federalreserve.gov/newsevents/pressreleases/monetary20220921a.htm"}},"_tags":["story","author_scrlk","story_32928803"],"author":"scrlk","children":[32929258,32929274,32929333,32929366,32929370,32929376,32929485,32929545,32929576,32929648,32929663,32929719,32929728,32929760,32930345,32930356,32930711,32931086],"created_at":"2022-09-21T18:00:45Z","created_at_i":1663783245,"num_comments":228,"objectID":"32928803","points":153,"story_id":32928803,"title":"Federal Reserve to increase interest rates by 75 basis points for the third time","updated_at":"2026-02-06T02:44:19Z","url":"https://www.federalreserve.gov/newsevents/pressreleases/monetary20220921a.htm"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"zdw"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"Third Parties and Single Points of Failure"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"https://calendar.perfplanet.com/2025/third-parties-and-single-points-of-failure/"}},"_tags":["story","author_zdw","story_46422461"],"author":"zdw","children":[46450412],"created_at":"2025-12-29T16:47:17Z","created_at_i":1767026837,"num_comments":1,"objectID":"46422461","points":12,"story_id":46422461,"title":"Third Parties and Single Points of Failure","updated_at":"2026-03-05T23:15:17Z","url":"https://calendar.perfplanet.com/2025/third-parties-and-single-points-of-failure/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"rawgabbit"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"Tucker Carlson unveils 10-point manifesto amid third-party speculation"},"url":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["third"],"value":"https://www.axios.com/2026/08/06/tucker-carlson-third-party-israel-trump"}},"_tags":["story","author_rawgabbit","story_49195503"],"author":"rawgabbit","children":[49195556,49195703,49196447],"created_at":"2026-08-06T12:01:31Z","created_at_i":1786017691,"num_comments":3,"objectID":"49195503","points":10,"story_id":49195503,"title":"Tucker Carlson unveils 10-point manifesto amid third-party speculation","updated_at":"2026-08-13T00:17:16Z","url":"https://www.axios.com/2026/08/06/tucker-carlson-third-party-israel-trump"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"bdcravens"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"MLB Used Two Balls Again This Year, and Evidence Points to a Third"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"https://www.insider.com/mlb-used-two-balls-again-this-year-and-evidence-points-to-a-third-2022-12"}},"_tags":["story","author_bdcravens","story_34480915"],"author":"bdcravens","created_at":"2023-01-22T18:55:27Z","created_at_i":1674413727,"num_comments":0,"objectID":"34480915","points":10,"story_id":34480915,"title":"MLB Used Two Balls Again This Year, and Evidence Points to a Third","updated_at":"2024-09-20T13:06:38Z","url":"https://www.insider.com/mlb-used-two-balls-again-this-year-and-evidence-points-to-a-third-2022-12"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"yuan43"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"Fed goes big again with third-straight three-quarter-point rate hike"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://lite.cnn.com/en/article/h_60b41ec196c06611beb22712abf3056a"}},"_tags":["story","author_yuan43","story_32930033"],"author":"yuan43","created_at":"2022-09-21T19:14:25Z","created_at_i":1663787665,"num_comments":0,"objectID":"32930033","points":5,"story_id":32930033,"title":"Fed goes big again with third-straight three-quarter-point rate hike","updated_at":"2024-09-20T12:03:52Z","url":"https://lite.cnn.com/en/article/h_60b41ec196c06611beb22712abf3056a"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"uptown"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"MLB Used Two Balls Again This Year, and Evidence Points to a Third"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"https://www.insider.com/mlb-used-two-balls-again-this-year-and-evidence-points-to-a-third-2022-12"}},"_tags":["story","author_uptown","story_33887969"],"author":"uptown","created_at":"2022-12-06T22:22:15Z","created_at_i":1670365335,"num_comments":0,"objectID":"33887969","points":2,"story_id":33887969,"title":"MLB Used Two Balls Again This Year, and Evidence Points to a Third","updated_at":"2024-09-20T12:44:00Z","url":"https://www.insider.com/mlb-used-two-balls-again-this-year-and-evidence-points-to-a-third-2022-12"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"megalord"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"Our new single point of failure - third-party script for cookies consent"},"url":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["third"],"value":"https://schnuky.substack.com/p/today-third-party-script-for-cookies"}},"_tags":["story","author_megalord","story_35998984"],"author":"megalord","created_at":"2023-05-19T07:52:19Z","created_at_i":1684482739,"num_comments":0,"objectID":"35998984","points":1,"story_id":35998984,"title":"Our new single point of failure - third-party script for cookies consent","updated_at":"2024-09-20T14:07:57Z","url":"https://schnuky.substack.com/p/today-third-party-script-for-cookies"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"joeytom"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"Hi,
I spent the last couple of months developing OurClubFundraiser(http://www.ourclubfundraiser.com) and launched it about two weeks ago.I would love to get some feedback from the HN community on how I could improve the site and tweak the concept.
Since launching I have found that: \n1) I am not getting enough visits(Average around 5 a day)\n2) Any of the visits I do get are not converting.
I am working on 1) and will hopefully be able to dramatically increase the number of visits over the next few weeks.
I could really do with some third party perspectives on point 2.
1)Is the concept not explained well enough?\n2)Have I got the pricing wrong? \n3)Is there a problem with the concept that I am missing?
Any advice given would be greatly appreciated."},"title":{"matchLevel":"none","matchedWords":[],"value":"Show HN: Need feedback on new website/concept"},"url":{"matchLevel":"none","matchedWords":[],"value":""}},"_tags":["story","author_joeytom","story_4329485","show_hn"],"author":"joeytom","children":[4329696,4329737,4329739,4329812,4329831,4329894,4329915,4330013,4330675,4353707],"created_at":"2012-08-02T16:23:20Z","created_at_i":1343924600,"num_comments":18,"objectID":"4329485","points":8,"story_id":4329485,"story_text":"Hi,
I spent the last couple of months developing OurClubFundraiser(http://www.ourclubfundraiser.com) and launched it about two weeks ago.I would love to get some feedback from the HN community on how I could improve the site and tweak the concept.
Since launching I have found that: \n1) I am not getting enough visits(Average around 5 a day)\n2) Any of the visits I do get are not converting.
I am working on 1) and will hopefully be able to dramatically increase the number of visits over the next few weeks.
I could really do with some third party perspectives on point 2.
1)Is the concept not explained well enough?\n2)Have I got the pricing wrong? \n3)Is there a problem with the concept that I am missing?
Any advice given would be greatly appreciated.","title":"Show HN: Need feedback on new website/concept","updated_at":"2024-09-19T18:44:08Z","url":""},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"brundolf"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"The community seems to be coalescing around TypeScript, to a point where even third-party projects like Deno support it exclusively. I haven't been keeping up with Flow for a few years, but I never hear about it anymore and last time I used it it had lots of issues.
Is it in maintenance mode, or is it still hoping to compete one day? Does it have a dedicated userbase or just legacy code?"},"title":{"matchLevel":"none","matchedWords":[],"value":"Ask HN: Is Flow (JS type system) dead?"}},"_tags":["story","author_brundolf","story_27542116","ask_hn"],"author":"brundolf","children":[27542205,27542703,27548533],"created_at":"2021-06-17T17:29:03Z","created_at_i":1623950943,"num_comments":3,"objectID":"27542116","points":5,"story_id":27542116,"story_text":"The community seems to be coalescing around TypeScript, to a point where even third-party projects like Deno support it exclusively. I haven't been keeping up with Flow for a few years, but I never hear about it anymore and last time I used it it had lots of issues.
Is it in maintenance mode, or is it still hoping to compete one day? Does it have a dedicated userbase or just legacy code?","title":"Ask HN: Is Flow (JS type system) dead?","updated_at":"2024-09-20T08:47:18Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"deathmonger5000"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"Hi HN,
Circus Chief is a tool for managing coding agent sessions from a browser. It's specifically optimized for small screens. It supports Claude Code, OpenAI Codex, and Google Gemini CLI agents.
Features
Agents can operate Circus Chief itself. Agents can spawn sessions, schedule sessions, interact with the Kanban board \u2014 anything you can do in the UI, an agent can also do.
Schedule work ahead of time.
Automatically reschedule when you hit usage limits.
Configurable, chainable prompt templates.
User-defined commands - agents can run them and see the results, and so can you in the UI. Handy for local CI or routine workflow steps.
Worktree-per-session isolation, or elect to work on a specific branch.
Shared canvas - this is a place for shared artifacts that don't belong in the code. Useful for iterating on planning documents.
Bring your own provider. Use subscription auth for Anthropic, OpenAI, or Google, or point sessions at third-party providers with Anthropic- or OpenAI-compatible endpoints.
If you want to try it:
npx circuschief\nDevelopment process
I built this almost entirely from my phone and tablet \u2014 starting with Claude Code using both Anthropic and z.ai, then adding Codex to the mix around a month ago.
It was vibe coded in the sense that I was not reviewing many individual lines of code. I did review and iterate on plans with the agents, but often stayed out of implementation details. If a plan made sense then I would usually have the agent review it a few times before just letting the agent implement without any intervention. I've enjoyed building this way, but I don't deeply understand a lot of the details of the implementation. That's bitten me exactly once so far, see this issue: https://github.com/ferrislucas/Circus-Chief/issues/944
I used Circus Chief to build Circus Chief from a very early stage. I bootstrapped with Vibe Kanban, then switched to dogfooding once Circus Chief was usable. A lot of the app has been built through the app itself.
By the numbers
2,184 agent sessions since Christmas 2025 \u2014 1,423 were child/follow-up sessions.
Agent mix: 1,818 Claude Code sessions, 363 Codex sessions, and 3 Gemini sessions.
50,929 conversation messages across those sessions.
41 sessions (1.9%) contained strong profanity in user-authored messages.
990 sessions linked to GitHub PRs, across 685 distinct PRs, 759 branches, and 731 worktrees.
Average logged agent call: 131.3 seconds.
2,410 commits on main; 883 pull requests total \u2014 847 merged (~96%).
Repo scale: 887 tracked files and 310K current lines of code (excludes lockfiles, fixtures, and recorded test cassettes).
Peak day: 76 commits on March 21, 2026, first at 12:49am, last at 11:57pm \u2014 one every 18 minutes for 23 hours.
Busiest merge day: 28 pull requests shipped \u2014 New Year's Day, 2026.
Largest PR by diff: 140K insertions across 92 files \u2014 replacing mock Claude fixtures with VCR-style record/replay.
Code: https://github.com/ferrislucas/Circus-Chief
Analytics disclosure: the npm build includes PostHog for anonymous page-view analytics (session recording off, DNT respected). Disable with --no-analytics, in Settings \u2192 General, or build from source without a PostHog key."},"title":{"matchLevel":"none","matchedWords":[],"value":"Show HN: Circus Chief \u2013 Claude Code, Codex, and Gemini from Your Phone"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://github.com/ferrislucas/Circus-Chief"}},"_tags":["story","author_deathmonger5000","story_48370360","show_hn"],"author":"deathmonger5000","created_at":"2026-06-02T13:59:56Z","created_at_i":1780408796,"num_comments":0,"objectID":"48370360","points":4,"story_id":48370360,"story_text":"Hi HN,
Circus Chief is a tool for managing coding agent sessions from a browser. It's specifically optimized for small screens. It supports Claude Code, OpenAI Codex, and Google Gemini CLI agents.
Features
Agents can operate Circus Chief itself. Agents can spawn sessions, schedule sessions, interact with the Kanban board \u2014 anything you can do in the UI, an agent can also do.
Schedule work ahead of time.
Automatically reschedule when you hit usage limits.
Configurable, chainable prompt templates.
User-defined commands - agents can run them and see the results, and so can you in the UI. Handy for local CI or routine workflow steps.
Worktree-per-session isolation, or elect to work on a specific branch.
Shared canvas - this is a place for shared artifacts that don't belong in the code. Useful for iterating on planning documents.
Bring your own provider. Use subscription auth for Anthropic, OpenAI, or Google, or point sessions at third-party providers with Anthropic- or OpenAI-compatible endpoints.
If you want to try it:
npx circuschief\nDevelopment process
I built this almost entirely from my phone and tablet \u2014 starting with Claude Code using both Anthropic and z.ai, then adding Codex to the mix around a month ago.
It was vibe coded in the sense that I was not reviewing many individual lines of code. I did review and iterate on plans with the agents, but often stayed out of implementation details. If a plan made sense then I would usually have the agent review it a few times before just letting the agent implement without any intervention. I've enjoyed building this way, but I don't deeply understand a lot of the details of the implementation. That's bitten me exactly once so far, see this issue: https://github.com/ferrislucas/Circus-Chief/issues/944
I used Circus Chief to build Circus Chief from a very early stage. I bootstrapped with Vibe Kanban, then switched to dogfooding once Circus Chief was usable. A lot of the app has been built through the app itself.
By the numbers
2,184 agent sessions since Christmas 2025 \u2014 1,423 were child/follow-up sessions.
Agent mix: 1,818 Claude Code sessions, 363 Codex sessions, and 3 Gemini sessions.
50,929 conversation messages across those sessions.
41 sessions (1.9%) contained strong profanity in user-authored messages.
990 sessions linked to GitHub PRs, across 685 distinct PRs, 759 branches, and 731 worktrees.
Average logged agent call: 131.3 seconds.
2,410 commits on main; 883 pull requests total \u2014 847 merged (~96%).
Repo scale: 887 tracked files and 310K current lines of code (excludes lockfiles, fixtures, and recorded test cassettes).
Peak day: 76 commits on March 21, 2026, first at 12:49am, last at 11:57pm \u2014 one every 18 minutes for 23 hours.
Busiest merge day: 28 pull requests shipped \u2014 New Year's Day, 2026.
Largest PR by diff: 140K insertions across 92 files \u2014 replacing mock Claude fixtures with VCR-style record/replay.
Code: https://github.com/ferrislucas/Circus-Chief
Analytics disclosure: the npm build includes PostHog for anonymous page-view analytics (session recording off, DNT respected). Disable with --no-analytics, in Settings \u2192 General, or build from source without a PostHog key.","title":"Show HN: Circus Chief \u2013 Claude Code, Codex, and Gemini from Your Phone","updated_at":"2026-06-05T11:43:19Z","url":"https://github.com/ferrislucas/Circus-Chief"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"kafkabenchmark"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"I have inherited an application which uses Kafkas request/reply configuration to provide a (mostly) synchronous configuration, which seems to work well enough. However I have now been approached by management to figure out what sort of scale this can handle. I am a bit green with Kafka so did a bit of looking around, however my normal load testing tools seem to not support Kafka request reply.
Gatling doesn't want to cover this, and points to a third party plug in, which appears to be undocumented and archived.
JMeter can generate load via Pepperbox, but doesn't consume.
There appears to be a lot of information on generating load, but nothing about measuring throughput in this set up. Is this a step too far in validating an asynchronous service, or are there setups that can be used to understand the limits of our infrastructure?"},"title":{"matchLevel":"none","matchedWords":[],"value":"Ask HN: Benchmarking Kafka Request Reply"}},"_tags":["story","author_kafkabenchmark","story_37135021","ask_hn"],"author":"kafkabenchmark","created_at":"2023-08-15T15:16:58Z","created_at_i":1692112618,"num_comments":0,"objectID":"37135021","points":2,"story_id":37135021,"story_text":"I have inherited an application which uses Kafkas request/reply configuration to provide a (mostly) synchronous configuration, which seems to work well enough. However I have now been approached by management to figure out what sort of scale this can handle. I am a bit green with Kafka so did a bit of looking around, however my normal load testing tools seem to not support Kafka request reply.
Gatling doesn't want to cover this, and points to a third party plug in, which appears to be undocumented and archived.
JMeter can generate load via Pepperbox, but doesn't consume.
There appears to be a lot of information on generating load, but nothing about measuring throughput in this set up. Is this a step too far in validating an asynchronous service, or are there setups that can be used to understand the limits of our infrastructure?","title":"Ask HN: Benchmarking Kafka Request Reply","updated_at":"2024-09-20T14:48:56Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"cjohnson318"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"I can do something like `isEspanol ? Cosas : Things` everywhere, but I'd like to be able to support a third language at some point if I needed to. Does anyone have strong opinions about building an app for more than one language in a sane and sustainable way?"},"title":{"matchLevel":"none","matchedWords":[],"value":"Ask HN: Advice on Writing Bilingual App?"}},"_tags":["story","author_cjohnson318","story_34795019","ask_hn"],"author":"cjohnson318","children":[34795171,34798606],"created_at":"2023-02-14T19:55:41Z","created_at_i":1676404541,"num_comments":4,"objectID":"34795019","points":1,"story_id":34795019,"story_text":"I can do something like `isEspanol ? Cosas : Things` everywhere, but I'd like to be able to support a third language at some point if I needed to. Does anyone have strong opinions about building an app for more than one language in a sane and sustainable way?","title":"Ask HN: Advice on Writing Bilingual App?","updated_at":"2024-09-20T13:20:00Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"iluvdata"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"What are good data augmentation techniques which can work with 3d data specifically depth being third dimension. Also, any pointers on tools around them would be of great help."},"title":{"matchLevel":"none","matchedWords":[],"value":"Ask HN: Data augmentation tools for 3d object detection"}},"_tags":["story","author_iluvdata","story_18626819","ask_hn"],"author":"iluvdata","children":[18626913,18627709,18629151],"created_at":"2018-12-07T11:36:00Z","created_at_i":1544182560,"num_comments":6,"objectID":"18626819","points":28,"story_id":18626819,"story_text":"What are good data augmentation techniques which can work with 3d data specifically depth being third dimension. Also, any pointers on tools around them would be of great help.","title":"Ask HN: Data augmentation tools for 3d object detection","updated_at":"2024-09-20T03:28:26Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"Spike"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"Hi everyone,
I need your advice on starting a company with two other partners, and my question really is about equity levels. Here's the story so far...All three of us have been working for a university and for various reasons we all have decided to move away from the academic world and go in industry. One of the partners (let's call him X) has good links with industry in China and he managed to bring money over to start a project, which is the design and development of a point-of-care device. The contract terms between the university and the funding company made clear that any IP arising from this work will belong to the company, with us three being the named inventors in the patents generated.
They were happy with the outcome and through X's prior links with company (X is a board member in the company) it was agreed to form a new company. They agreed to invest around \u00a3300k in return for about 40% stake in the new company. The remaining 60% would be divided as follows: 20% for X (he claims he had bought in the past seven patents from his previous company which went into liquidation, investing \u00a3120k - later he said it was actually \u00a3150k), 20-25% for options for other paeople that will eventually join the company at a later stage) and the remaining 15-20% for myself and the third partner.
At this point I need to point out that myself and the third partner are making things happening (in terms of engineering) and without us it would impossible (or at least it would take a lot of time for him to build a new team) to start the company now. Also, it is important to note that X has no other option but to leave university, as his contract is due for re-evaluation in September, with negative predictions for the outcome.
Some more info about the forecast of the new company: we will be using the manufacturing facilities of the funding company and we will have direct access to their sales network, They say the market is already there, they have clients and last year they had a turnover of more than $10m (it might be more, i can't remember accurately). My salary will be slightly less than the market value (probably around \u00a335k/yr) and I suspect X will be asking for a lot more, as he takes for granted that he will be the technical director and acting CEO of the company.
There are many details that we're not told of, because X is taking the excuse that his contact in he funding company does not speak English fluently, that's why all their conversations are in Chinese. We get the occasional email in English circulated among all of us, but nothing mentioning stake levels. I feel I'm heading straight towards a not-so-good deal, while I know for a fact that X will not go below 20% as he immediately points his finger towards his \u00a3150k investment in previous patents, while making claims of the type \"I have achieved with great efforts your involvement in the new company as co-founders with a stake, rather than just employees\". I have started to grow tired of his bull and I'm trying to figure out how I can go about sealing a good deal for me. I have previously discussed with him the matter of equity, as one day he came to us and said \"when the time comes to decide equity, you must be modest\" and at that point I told him that modesty has no place in business, I will only be fair and I won't settle for anything less than fair, I will literally walk away if I'm not happy with what's offered to me.
I have the feeling that this is the ONLY time I can bargain with him, or them (the funding company), as this is the only time I feel that if i break the team, he won't be able to go ahead, at least not according to his time plans. I'm fighting with myself, as one side of me wants to grasp this opportunity of potentially making enough for the rest of my life, while my other side is telling me that X is trying to manipulate the whole situation based on the trust he has with the funding company, and that without him they wouldn't invest, which is probably true. Still, I can't trust him, at all.
Apologies for the long story, I just needed to draw, with as much detail as possible,the whole picture. I would appreciate any thoughts you might have, I desperately need some advice from people who understand business. I'm very good at what I'm doing but when it comes to business, it's all chinese to me....
Thank you in advance.
S."},"title":{"matchLevel":"none","matchedWords":[],"value":"Startup company stake - Please advise..."},"url":{"matchLevel":"none","matchedWords":[],"value":""}},"_tags":["story","author_Spike","story_686561","ask_hn"],"author":"Spike","children":[686580,686591,686682,686720,687116],"created_at":"2009-07-03T20:24:34Z","created_at_i":1246652674,"num_comments":7,"objectID":"686561","points":5,"story_id":686561,"story_text":"Hi everyone,
I need your advice on starting a company with two other partners, and my question really is about equity levels. Here's the story so far...All three of us have been working for a university and for various reasons we all have decided to move away from the academic world and go in industry. One of the partners (let's call him X) has good links with industry in China and he managed to bring money over to start a project, which is the design and development of a point-of-care device. The contract terms between the university and the funding company made clear that any IP arising from this work will belong to the company, with us three being the named inventors in the patents generated.
They were happy with the outcome and through X's prior links with company (X is a board member in the company) it was agreed to form a new company. They agreed to invest around \u00a3300k in return for about 40% stake in the new company. The remaining 60% would be divided as follows: 20% for X (he claims he had bought in the past seven patents from his previous company which went into liquidation, investing \u00a3120k - later he said it was actually \u00a3150k), 20-25% for options for other paeople that will eventually join the company at a later stage) and the remaining 15-20% for myself and the third partner.
At this point I need to point out that myself and the third partner are making things happening (in terms of engineering) and without us it would impossible (or at least it would take a lot of time for him to build a new team) to start the company now. Also, it is important to note that X has no other option but to leave university, as his contract is due for re-evaluation in September, with negative predictions for the outcome.
Some more info about the forecast of the new company: we will be using the manufacturing facilities of the funding company and we will have direct access to their sales network, They say the market is already there, they have clients and last year they had a turnover of more than $10m (it might be more, i can't remember accurately). My salary will be slightly less than the market value (probably around \u00a335k/yr) and I suspect X will be asking for a lot more, as he takes for granted that he will be the technical director and acting CEO of the company.
There are many details that we're not told of, because X is taking the excuse that his contact in he funding company does not speak English fluently, that's why all their conversations are in Chinese. We get the occasional email in English circulated among all of us, but nothing mentioning stake levels. I feel I'm heading straight towards a not-so-good deal, while I know for a fact that X will not go below 20% as he immediately points his finger towards his \u00a3150k investment in previous patents, while making claims of the type \"I have achieved with great efforts your involvement in the new company as co-founders with a stake, rather than just employees\". I have started to grow tired of his bull and I'm trying to figure out how I can go about sealing a good deal for me. I have previously discussed with him the matter of equity, as one day he came to us and said \"when the time comes to decide equity, you must be modest\" and at that point I told him that modesty has no place in business, I will only be fair and I won't settle for anything less than fair, I will literally walk away if I'm not happy with what's offered to me.
I have the feeling that this is the ONLY time I can bargain with him, or them (the funding company), as this is the only time I feel that if i break the team, he won't be able to go ahead, at least not according to his time plans. I'm fighting with myself, as one side of me wants to grasp this opportunity of potentially making enough for the rest of my life, while my other side is telling me that X is trying to manipulate the whole situation based on the trust he has with the funding company, and that without him they wouldn't invest, which is probably true. Still, I can't trust him, at all.
Apologies for the long story, I just needed to draw, with as much detail as possible,the whole picture. I would appreciate any thoughts you might have, I desperately need some advice from people who understand business. I'm very good at what I'm doing but when it comes to business, it's all chinese to me....
Thank you in advance.
S.","title":"Startup company stake - Please advise...","updated_at":"2024-09-19T16:43:28Z","url":""},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"yewenjie"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"What is the best MCP server in the market for adding long-term memory to AI conversations? Bonus points if it's open source and bonus points if supported by third-party benchmarks."},"title":{"matchLevel":"none","matchedWords":[],"value":"Ask HN: MCP server for long-term memory?"}},"_tags":["story","author_yewenjie","story_44442810","ask_hn"],"author":"yewenjie","children":[44442873],"created_at":"2025-07-02T12:17:00Z","created_at_i":1751458620,"num_comments":1,"objectID":"44442810","points":2,"story_id":44442810,"story_text":"What is the best MCP server in the market for adding long-term memory to AI conversations? Bonus points if it's open source and bonus points if supported by third-party benchmarks.","title":"Ask HN: MCP server for long-term memory?","updated_at":"2025-07-03T06:31:09Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"davidajackson"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"I thought I would start a thread on here to discuss some of the challenges in Generative AI particularly with enforcing on-topic conversation and tone. It seems this poses substantial cost restrictions on deployment, as using on topic checks and classifiers for inference prompts does not scale well. It seems an O(N) problem that has high cost and low performance with inference. In addition, though embedding methods could be used to ensure on topic conversations, I want to specifically discuss the problem of combined context, where a predefined cluster is isolated, but combined alternative context in the prompt 'dilutes' the vector such that it no longer points to a defined cluster, or has a cosine similarity such that it does no trigger a match or non match decisively. As an example, consider the sentence. "I would like to discuss politics, and I also enjoy the stock market." Though sentence splitters and or inference topic modeling could be used, that seems it could get expensive. Consider also the problem of impersonation, where the model is instructed to act like a volatile character. Finally, I throw out there a stranger version of prompts meant to throw model results off, where perhaps the letters are switched in such as a way that when the model computes the answer by decoding the answer, it does not consider the final context (rotate the following letters by 1 forward: "GH", which would produce "HI"). I am least concerned about the third but I am simply pointing it out. I know there are a lot of bright minds on here and thought it might warrant a conversation of how people have overcome these challenges and what ideas people might have."},"title":{"matchLevel":"none","matchedWords":[],"value":"Ask HN: How Would You Build Restrictions for Deployed Model Prompts/Answers?"}},"_tags":["story","author_davidajackson","story_40002948","ask_hn"],"author":"davidajackson","children":[40003132],"created_at":"2024-04-11T15:02:55Z","created_at_i":1712847775,"num_comments":1,"objectID":"40002948","points":2,"story_id":40002948,"story_text":"I thought I would start a thread on here to discuss some of the challenges in Generative AI particularly with enforcing on-topic conversation and tone. It seems this poses substantial cost restrictions on deployment, as using on topic checks and classifiers for inference prompts does not scale well. It seems an O(N) problem that has high cost and low performance with inference. In addition, though embedding methods could be used to ensure on topic conversations, I want to specifically discuss the problem of combined context, where a predefined cluster is isolated, but combined alternative context in the prompt 'dilutes' the vector such that it no longer points to a defined cluster, or has a cosine similarity such that it does no trigger a match or non match decisively. As an example, consider the sentence. "I would like to discuss politics, and I also enjoy the stock market." Though sentence splitters and or inference topic modeling could be used, that seems it could get expensive. Consider also the problem of impersonation, where the model is instructed to act like a volatile character. Finally, I throw out there a stranger version of prompts meant to throw model results off, where perhaps the letters are switched in such as a way that when the model computes the answer by decoding the answer, it does not consider the final context (rotate the following letters by 1 forward: "GH", which would produce "HI"). I am least concerned about the third but I am simply pointing it out. I know there are a lot of bright minds on here and thought it might warrant a conversation of how people have overcome these challenges and what ideas people might have.","title":"Ask HN: How Would You Build Restrictions for Deployed Model Prompts/Answers?","updated_at":"2024-09-20T16:44:17Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"simulationship"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"To succeed on Hacker News (HN), you have to completely drop the "marketing" and "YouTube hook" tone. The HN community heavily downvotes clickbait, sensationalism, and marketing fluff.\nThey love "Show HN" posts, open-source projects, CLI tools, local LLMs, and clever technical solutions to messy data problems (like parsing poorly scanned government PDFs).\nHere are the best titles and the exact description (to use either as a text post or your first comment) tailored specifically for the Hacker News audience.\n The Hacker News Titles\nChoose one of these. On HN, titles should be strictly factual, descriptive, and avoid emojis.\nOption 1 (The Classic HN Format - Recommended):\nShow HN: epstein-search \u2013 A CLI to query the unsealed court files with local LLMs\nOption 2 (Focus on the tech pipeline):\nShow HN: I built a local RAG CLI to make the Epstein PDFs searchable\nOption 3 (Straight to the point):\nShow HN: epstein-search \u2013 Query the Epstein document dumps offline via CLI\n The Hacker News Description (First Comment or Text Body)\nIf you submit the GitHub URL directly, immediately post this as the first comment. If you submit a text post, put this in the body. Keep the tone humble, technical, and open to feedback.\nHi HN,\nWhen the Epstein court documents and flight logs were unsealed, they were released the way most legal drops are: thousands of pages of messy, poorly scanned, unsearchable PDFs. Standard Ctrl+F doesn't work well due to OCR errors, and the sheer volume makes manual parsing a nightmare.\nTo solve this, I built epstein-search, an open-source Python CLI tool that lets you search and synthesize the documents using a Retrieval-Augmented Generation (RAG) pipeline directly in your terminal.\nHow it works:\nIt parses and chunks the original unsealed PDF files.\nYou can run queries against the dataset using API-based models (OpenAI/Anthropic) if you want speed.\nPrivacy-first: If you don't want your queries logged by a third-party API, you can point it directly to a local model (via Ollama or Llama.cpp) to run the entire search and retrieval process 100% offline.\nThe goal was to make this data accessible to researchers and OSINT investigators without requiring them to manually read thousands of pages of court dockets or hand over their search queries to OpenAI.\nRepo is here: https://github.com/simulationship/epstein-search"},"title":{"matchLevel":"none","matchedWords":[],"value":"Show HN: A CLI to query the unsealed court files with local LLMs"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://github.com/simulationship/epstein-search"}},"_tags":["story","author_simulationship","story_47157154","show_hn"],"author":"simulationship","created_at":"2026-02-25T20:11:07Z","created_at_i":1772050267,"num_comments":0,"objectID":"47157154","points":2,"story_id":47157154,"story_text":"To succeed on Hacker News (HN), you have to completely drop the "marketing" and "YouTube hook" tone. The HN community heavily downvotes clickbait, sensationalism, and marketing fluff.\nThey love "Show HN" posts, open-source projects, CLI tools, local LLMs, and clever technical solutions to messy data problems (like parsing poorly scanned government PDFs).\nHere are the best titles and the exact description (to use either as a text post or your first comment) tailored specifically for the Hacker News audience.\n The Hacker News Titles\nChoose one of these. On HN, titles should be strictly factual, descriptive, and avoid emojis.\nOption 1 (The Classic HN Format - Recommended):\nShow HN: epstein-search \u2013 A CLI to query the unsealed court files with local LLMs\nOption 2 (Focus on the tech pipeline):\nShow HN: I built a local RAG CLI to make the Epstein PDFs searchable\nOption 3 (Straight to the point):\nShow HN: epstein-search \u2013 Query the Epstein document dumps offline via CLI\n The Hacker News Description (First Comment or Text Body)\nIf you submit the GitHub URL directly, immediately post this as the first comment. If you submit a text post, put this in the body. Keep the tone humble, technical, and open to feedback.\nHi HN,\nWhen the Epstein court documents and flight logs were unsealed, they were released the way most legal drops are: thousands of pages of messy, poorly scanned, unsearchable PDFs. Standard Ctrl+F doesn't work well due to OCR errors, and the sheer volume makes manual parsing a nightmare.\nTo solve this, I built epstein-search, an open-source Python CLI tool that lets you search and synthesize the documents using a Retrieval-Augmented Generation (RAG) pipeline directly in your terminal.\nHow it works:\nIt parses and chunks the original unsealed PDF files.\nYou can run queries against the dataset using API-based models (OpenAI/Anthropic) if you want speed.\nPrivacy-first: If you don't want your queries logged by a third-party API, you can point it directly to a local model (via Ollama or Llama.cpp) to run the entire search and retrieval process 100% offline.\nThe goal was to make this data accessible to researchers and OSINT investigators without requiring them to manually read thousands of pages of court dockets or hand over their search queries to OpenAI.\nRepo is here: https://github.com/simulationship/epstein-search","title":"Show HN: A CLI to query the unsealed court files with local LLMs","updated_at":"2026-03-05T23:37:22Z","url":"https://github.com/simulationship/epstein-search"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"seanpk"},"title":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["third"],"value":"Third-Party Services in BlueOak Server (starting with Bos-Couchdb)"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"http://www.pointsource.com/blog/third-party-services-in-blueoak-server?utm_campaign=blueoak&utm_medium=social&utm_source=hackernews"}},"_tags":["story","author_seanpk","story_11223587"],"author":"seanpk","created_at":"2016-03-04T13:37:03Z","created_at_i":1457098623,"num_comments":0,"objectID":"11223587","points":1,"story_id":11223587,"title":"Third-Party Services in BlueOak Server (starting with Bos-Couchdb)","updated_at":"2024-09-19T22:54:00Z","url":"http://www.pointsource.com/blog/third-party-services-in-blueoak-server?utm_campaign=blueoak&utm_medium=social&utm_source=hackernews"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"rsaha7"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"Hi HN community, I have been working on benchmarking publicly available LLMs these past couple of weeks. More precisely, I am interested on the finetuning piece since a lot of businesses are starting to entertain the idea of self-hosting LLMs trained on their proprietary data rather than relying on third party APIs.
To this point, I am tracking the following 4 pillars of evaluation that businesses are typically look into:\n- Performance\n- Time to train an LLM\n- Cost to train an LLM\n- Inference (throughput / latency / cost per token)
For each LLM, my aim is to benchmark them for popular tasks, i.e., classification and summarization. Moreover, I would like to compare them against each other.
So far, I have benchmarked Flan-T5-Large, Falcon-7B and RedPajama and have found them to be very efficient in low-data situations, i.e., when there are very few annotated samples. Llama2-7B/13B and Writer\u2019s Palmyra are in the pipeline.
But there\u2019s so many LLMs out there! In case this work interests you, would be great to join forces.
GitHub repo attached \u2014 feedback is always welcome :)
Happy hacking!"},"title":{"matchLevel":"none","matchedWords":[],"value":"Show HN: finetune LLMs via the Finetuning Hub"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://github.com/georgian-io/LLM-Finetuning-Hub"}},"_tags":["story","author_rsaha7","story_37381296","show_hn"],"author":"rsaha7","children":[37382956,37383225,37384606,37385295,37413547],"created_at":"2023-09-04T15:16:42Z","created_at_i":1693840602,"num_comments":17,"objectID":"37381296","points":80,"story_id":37381296,"story_text":"Hi HN community, I have been working on benchmarking publicly available LLMs these past couple of weeks. More precisely, I am interested on the finetuning piece since a lot of businesses are starting to entertain the idea of self-hosting LLMs trained on their proprietary data rather than relying on third party APIs.
To this point, I am tracking the following 4 pillars of evaluation that businesses are typically look into:\n- Performance\n- Time to train an LLM\n- Cost to train an LLM\n- Inference (throughput / latency / cost per token)
For each LLM, my aim is to benchmark them for popular tasks, i.e., classification and summarization. Moreover, I would like to compare them against each other.
So far, I have benchmarked Flan-T5-Large, Falcon-7B and RedPajama and have found them to be very efficient in low-data situations, i.e., when there are very few annotated samples. Llama2-7B/13B and Writer\u2019s Palmyra are in the pipeline.
But there\u2019s so many LLMs out there! In case this work interests you, would be great to join forces.
GitHub repo attached \u2014 feedback is always welcome :)
Happy hacking!","title":"Show HN: finetune LLMs via the Finetuning Hub","updated_at":"2026-05-13T06:25:33Z","url":"https://github.com/georgian-io/LLM-Finetuning-Hub"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"gtkid190"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"Friends keep telling me I should stfu and stop complaining but I'm genuinely depressed and looking for constructive feedback. I failed out of my universities CS program and ended up doing philosophy with a double minor in math and stats. It's my fault. I got cocky. My university is extremely cut throat when it comes to the CS program so even if I wouldn't have gotten in with really pushing myself, I would not be complaining right now. I failed intro to CS twice and passed on the 3rd try with a 52. Switched to philosophy after taking being put on suspension. Recovered my GPA (a bit) and finished university last year. Even though I shat the bed with university, I did a lot of hackathons, audited courses because of pure interest (through MIT open courseware), and getting internships. I ended up joining Shopify as a fullstack. But now, I really want to go back to school and learn things the traditional way because it feels too overwhelming. I can't work in the states because of tn visa regulations. My gpa is pathetic because of the failed courses in early years. I really don't have the patience to start a second degree at a third tier college/university because it's pointless to do so. I want to get into one of those online masters programs but let's be honest, with a pathetic 2.44 cgpa there is no way any masters program is gonna let me in. Seeking advice on what to do at this point"},"title":{"matchLevel":"none","matchedWords":[],"value":"Ask HN: Feeling Hopeless and Lost at 23"}},"_tags":["story","author_gtkid190","story_34665256","ask_hn"],"author":"gtkid190","children":[34665480,34665553,34665575,34665576,34665581,34665598,34665632,34665638,34665649,34665667,34665670,34665679,34665688,34665696,34665707,34665720,34665721,34665726,34665734,34665737,34665739,34665758,34665770,34665773,34665782,34665786,34665805,34665823,34665825,34665864,34665870,34665885,34665902,34665916,34665952,34665984,34665997,34666011,34666047,34666089,34666169,34666368,34666483,34666541,34666712,34666770,34666835,34666894,34667232,34667985,34668026,34668789,34668823,34668952,34669878,34671858,34672409],"created_at":"2023-02-05T14:42:21Z","created_at_i":1675608141,"num_comments":64,"objectID":"34665256","points":49,"story_id":34665256,"story_text":"Friends keep telling me I should stfu and stop complaining but I'm genuinely depressed and looking for constructive feedback. I failed out of my universities CS program and ended up doing philosophy with a double minor in math and stats. It's my fault. I got cocky. My university is extremely cut throat when it comes to the CS program so even if I wouldn't have gotten in with really pushing myself, I would not be complaining right now. I failed intro to CS twice and passed on the 3rd try with a 52. Switched to philosophy after taking being put on suspension. Recovered my GPA (a bit) and finished university last year. Even though I shat the bed with university, I did a lot of hackathons, audited courses because of pure interest (through MIT open courseware), and getting internships. I ended up joining Shopify as a fullstack. But now, I really want to go back to school and learn things the traditional way because it feels too overwhelming. I can't work in the states because of tn visa regulations. My gpa is pathetic because of the failed courses in early years. I really don't have the patience to start a second degree at a third tier college/university because it's pointless to do so. I want to get into one of those online masters programs but let's be honest, with a pathetic 2.44 cgpa there is no way any masters program is gonna let me in. Seeking advice on what to do at this point","title":"Ask HN: Feeling Hopeless and Lost at 23","updated_at":"2025-10-11T11:53:31Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"pcc"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"I have been looking at a variety of open source billing implementations, but can't seem to find anything that supports a \"billing on behalf of\" paradigm properly (or where it can be bolted on easily). It seems as if all the open source billing systems have evolved from telecoms or ISP billing problems. They are too complex and with the wrong focus for my problem domain.
So I thought it seemed sensible to ask for insight from the larger community.
Its easy to implement some variation on the \"double entry universal ledger\" idea, and we've found this adequate for a whole range of accounting / billing problems. However, there seems to be a whole class of billing use cases that would be greatly facilitated by a more sophisticated approach:
- when a service provider bills a customer, a whole series of entries/transactions might actually result. For example, it may be a third party who is actually providing the service, in other words the service provider is actually billing on behalf of this third party. Also, at the time of this billing, some of the entries/transactions might relate to ascribing a portion of the billing to various OTHER third parties (eg commissions, service fees to be collected for disbursement to other providers which exist in the service chain, etc.). Then also, in some cases there are service tax ramifications, so there may be a tax portion to the billing which may likewise have to be split among various third parties. This all needs to be recorded at the time of the original billing, since it reflects earnings accrued to the various parties under general accounting principles.
- the primary requirement being that each third party must know at any point what money has been ascribed to it (even if no payment has been received yet) -- but no funds for the billing will actually be released to anyone, until the billed customer has paid over and cleared some funds (might be in the form of a chq or eft).
- at the same time, all information that goes into calculating a billing, must remain associated with that billing. (E.g. all formulas and all their parameters, such as tax rates etc -- these can obviously differ vastly for different billings). The reason being that a partial reversal could be made against a billing at any point in future, and money must then be \"pulled back\" proportionally from any third party to which it has been ascribed, according to the original rules which were in effect at the time of that billing. (The rules might have changed in the interim, eg different third parties might have become associated with an account, commission rate might have changed etc).
Now it seems to me that this class of requirement surely cannot be that uncommon -- and yet I seem unable to find any reference to this class of problem in any open source billing system I can locate; neither can I seem to find any books or other materials which might cover implementation patterns and best practices for these classes of system.
For the project that needs this, we can of course design and code up something ourselves (e.g. a ledger that tracks only cash accounts with a separate billing structure tracking the rest, settling to and from the cash accounts); and if it turns out there is a broader need for this type of use case, the project might support open sourcing this code.
Thus on the one hand, I would be interested to learn of anyone else who has faced a similar type of use case, and how this was approached, and whether there would be value in creating a new open source project that speaks to this class of requirement.
On the other hand, it always seems prudent to write less code and see where one can piggy-back on existing projects -- so if anyone knows of any existing open source that can do this, please let us know!"},"title":{"matchLevel":"none","matchedWords":[],"value":"Ask YC: Billing Systems"},"url":{"matchLevel":"none","matchedWords":[],"value":""}},"_tags":["story","author_pcc","story_273264","ask_hn"],"author":"pcc","children":[273274,273276,273289,273453,273834],"created_at":"2008-08-11T19:47:09Z","created_at_i":1218484029,"num_comments":6,"objectID":"273264","points":7,"story_id":273264,"story_text":"I have been looking at a variety of open source billing implementations, but can't seem to find anything that supports a \"billing on behalf of\" paradigm properly (or where it can be bolted on easily). It seems as if all the open source billing systems have evolved from telecoms or ISP billing problems. They are too complex and with the wrong focus for my problem domain.
So I thought it seemed sensible to ask for insight from the larger community.
Its easy to implement some variation on the \"double entry universal ledger\" idea, and we've found this adequate for a whole range of accounting / billing problems. However, there seems to be a whole class of billing use cases that would be greatly facilitated by a more sophisticated approach:
- when a service provider bills a customer, a whole series of entries/transactions might actually result. For example, it may be a third party who is actually providing the service, in other words the service provider is actually billing on behalf of this third party. Also, at the time of this billing, some of the entries/transactions might relate to ascribing a portion of the billing to various OTHER third parties (eg commissions, service fees to be collected for disbursement to other providers which exist in the service chain, etc.). Then also, in some cases there are service tax ramifications, so there may be a tax portion to the billing which may likewise have to be split among various third parties. This all needs to be recorded at the time of the original billing, since it reflects earnings accrued to the various parties under general accounting principles.
- the primary requirement being that each third party must know at any point what money has been ascribed to it (even if no payment has been received yet) -- but no funds for the billing will actually be released to anyone, until the billed customer has paid over and cleared some funds (might be in the form of a chq or eft).
- at the same time, all information that goes into calculating a billing, must remain associated with that billing. (E.g. all formulas and all their parameters, such as tax rates etc -- these can obviously differ vastly for different billings). The reason being that a partial reversal could be made against a billing at any point in future, and money must then be \"pulled back\" proportionally from any third party to which it has been ascribed, according to the original rules which were in effect at the time of that billing. (The rules might have changed in the interim, eg different third parties might have become associated with an account, commission rate might have changed etc).
Now it seems to me that this class of requirement surely cannot be that uncommon -- and yet I seem unable to find any reference to this class of problem in any open source billing system I can locate; neither can I seem to find any books or other materials which might cover implementation patterns and best practices for these classes of system.
For the project that needs this, we can of course design and code up something ourselves (e.g. a ledger that tracks only cash accounts with a separate billing structure tracking the rest, settling to and from the cash accounts); and if it turns out there is a broader need for this type of use case, the project might support open sourcing this code.
Thus on the one hand, I would be interested to learn of anyone else who has faced a similar type of use case, and how this was approached, and whether there would be value in creating a new open source project that speaks to this class of requirement.
On the other hand, it always seems prudent to write less code and see where one can piggy-back on existing projects -- so if anyone knows of any existing open source that can do this, please let us know!","title":"Ask YC: Billing Systems","updated_at":"2024-09-19T16:24:30Z","url":""},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"felixheikka"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"I want to share my short story as I think it can be helpful to a lot of people on here.
This is for those of you who have tried every marketing strategy under the sun with no success (like I previously had).
Just two months ago I had been working hard for 7 months straight without getting any real interest or users to my product.
I had built two products and it was the same story for both: nothing.
When I started my third project I took the time to analyze what I\u2019d done wrong with my first two products.
A commonality I found was that I didn\u2019t talk to anyone before building them. I just made assumptions that I thought made sense.
I tried validating them afterwards and the response was what I suspected. No interest at all from my target audience.
My products were to generic and there were better alternatives.
For my third project I used this pain point of building products no one wanted and I realized this is the problem I want to solve. It\u2019s a problem I experienced myself and I desperately wanted a solution to it.
So this time I reached out to my target audience before jumping ahead. I created a post on Reddit and shared a survey that helped me understand my target audience better.
Specifically:\n- How they\u2019re currently solving the problem\n- How big of a pain it is to them\n- How much they would pay for a solution\n- Their opinion on my solution idea
The response was good so I decided to move ahead and build the MVP. Building it took about 30 days.
When it was finished, I shared it with the same people who responded to my first Reddit post.
That\u2019s how I got my first few users.
To keep growing I posted and engaged in communities relevant to my target audience on X and Reddit.
Just after two weeks I had gained 100 users which is more than I could ever have imagined, and about 100 more users than I had ever had for any of my other projects.
Finally getting users for my project was the best feeling, and in my opinion I owe it to validating the idea this time.
My project continued growing and less than two months later I find myself at $2200 revenue, which is crazy since it feels like it was yesterday I was struggling just to get anyone to sign up for my projects.
Anyways, I thought I\u2019d share this and maybe it can help you in case idea validation is something you skipped before you started building."},"title":{"matchLevel":"none","matchedWords":[],"value":"Validating my idea got me to $2200 revenue in less than 2 months"}},"_tags":["story","author_felixheikka","story_42249195","ask_hn"],"author":"felixheikka","children":[42249564,42249613,42249827],"created_at":"2024-11-26T19:42:29Z","created_at_i":1732650149,"num_comments":6,"objectID":"42249195","points":4,"story_id":42249195,"story_text":"I want to share my short story as I think it can be helpful to a lot of people on here.
This is for those of you who have tried every marketing strategy under the sun with no success (like I previously had).
Just two months ago I had been working hard for 7 months straight without getting any real interest or users to my product.
I had built two products and it was the same story for both: nothing.
When I started my third project I took the time to analyze what I\u2019d done wrong with my first two products.
A commonality I found was that I didn\u2019t talk to anyone before building them. I just made assumptions that I thought made sense.
I tried validating them afterwards and the response was what I suspected. No interest at all from my target audience.
My products were to generic and there were better alternatives.
For my third project I used this pain point of building products no one wanted and I realized this is the problem I want to solve. It\u2019s a problem I experienced myself and I desperately wanted a solution to it.
So this time I reached out to my target audience before jumping ahead. I created a post on Reddit and shared a survey that helped me understand my target audience better.
Specifically:\n- How they\u2019re currently solving the problem\n- How big of a pain it is to them\n- How much they would pay for a solution\n- Their opinion on my solution idea
The response was good so I decided to move ahead and build the MVP. Building it took about 30 days.
When it was finished, I shared it with the same people who responded to my first Reddit post.
That\u2019s how I got my first few users.
To keep growing I posted and engaged in communities relevant to my target audience on X and Reddit.
Just after two weeks I had gained 100 users which is more than I could ever have imagined, and about 100 more users than I had ever had for any of my other projects.
Finally getting users for my project was the best feeling, and in my opinion I owe it to validating the idea this time.
My project continued growing and less than two months later I find myself at $2200 revenue, which is crazy since it feels like it was yesterday I was struggling just to get anyone to sign up for my projects.
Anyways, I thought I\u2019d share this and maybe it can help you in case idea validation is something you skipped before you started building.","title":"Validating my idea got me to $2200 revenue in less than 2 months","updated_at":"2025-01-25T17:02:34Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"escapecharacter"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"I have personal notes, correspondence, code and documentation from nearly 20 years of work. These are spread across multiple (cloud) services, and searching across these fiefdoms has been impractical.
The problem goes like: "Ah, I remember having a conversation with someone about [algorithm], then recording an important insight. Let's find that."
This isn't a problem solved by an LLM. The blocker is that there isn't a way to run search code on all this plain text.
Services:
* Email (gmail, synced to my macOS disk with Apple Mail)
* Dropbox
* Notion
* Google Drive
* Obsidian
* Github
* Apple Notes
* Discord chats
* Trello
* My own blog
If I had everything synced to my mac's disk, maybe I could do a plaintext search there. However Spotlight's indexing is always incomplete and misses obvious files. My Dropbox is so large I don't sync it all locally.
Some services I no longer use, like Evernote. When I archived this service, I exported everything and moved it into my Dropbox. So, if I search Dropbox, it also searches old notes from Evernote. There's no way I could be doing this for all services I actively use.
The way I search now is I guess the service the result is most likely in, and search there. When finding no results, I search the next most likely service, ad nauseum\nFor my own blog, I used to use Google's site search, but I recently discovered this was incomplete: https://bsky.app/profile/dustinfreeman.bsky.social/post/3m5l5tto6pk27
I could imagine a solution where there's some 3rd party service that has access keys to all my services. But, let's be real, that's a huge amount of trust. Also, my access to all these services is 2FA'd with expiry, and so I'd be continually re-upping auth to this third party service. At that point, it makes sense to just do search how I do it now."},"title":{"matchLevel":"none","matchedWords":[],"value":"Ask HN: How can you search your personal data?"}},"_tags":["story","author_escapecharacter","story_45990113","ask_hn"],"author":"escapecharacter","children":[45994038],"created_at":"2025-11-20T07:58:23Z","created_at_i":1763625503,"num_comments":3,"objectID":"45990113","points":3,"story_id":45990113,"story_text":"I have personal notes, correspondence, code and documentation from nearly 20 years of work. These are spread across multiple (cloud) services, and searching across these fiefdoms has been impractical.
The problem goes like: "Ah, I remember having a conversation with someone about [algorithm], then recording an important insight. Let's find that."
This isn't a problem solved by an LLM. The blocker is that there isn't a way to run search code on all this plain text.
Services:
* Email (gmail, synced to my macOS disk with Apple Mail)
* Dropbox
* Notion
* Google Drive
* Obsidian
* Github
* Apple Notes
* Discord chats
* Trello
* My own blog
If I had everything synced to my mac's disk, maybe I could do a plaintext search there. However Spotlight's indexing is always incomplete and misses obvious files. My Dropbox is so large I don't sync it all locally.
Some services I no longer use, like Evernote. When I archived this service, I exported everything and moved it into my Dropbox. So, if I search Dropbox, it also searches old notes from Evernote. There's no way I could be doing this for all services I actively use.
The way I search now is I guess the service the result is most likely in, and search there. When finding no results, I search the next most likely service, ad nauseum\nFor my own blog, I used to use Google's site search, but I recently discovered this was incomplete: https://bsky.app/profile/dustinfreeman.bsky.social/post/3m5l5tto6pk27
I could imagine a solution where there's some 3rd party service that has access keys to all my services. But, let's be real, that's a huge amount of trust. Also, my access to all these services is 2FA'd with expiry, and so I'd be continually re-upping auth to this third party service. At that point, it makes sense to just do search how I do it now.","title":"Ask HN: How can you search your personal data?","updated_at":"2026-03-05T23:03:33Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"endymi0n"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"Hi HN,
I'm currently thinking about moving towards using an Applicant Tracking System for our company, but I feel a little unsure. As an applicant, I used to absolutely hate the rigid ATS systems of the first generation where instead of sending my CV as PDF, I had to fill in dozens of forms in order to just get back a cold standard mail after weeks.
That's why when we started our company, we went with a simple mailing list first. Thing is, we're slowly getting to the scale where this stops working and there are simply too many open positions and people working on them cluttering the list. I'd really like to have features such as human or tech-assisted pre-categorization to a job ad, multiple people in the pipeline, handover and just making sure we're answering everybody.
On the other hand, I'd really like to work with a system that doesn't have much more of a hurdle for applicants than sending a simple e-mail and that doesn't get in the way an interaction between humans. Bonus points for good integration in third-party software.
Has anyone used any ATS systems in the recent past and if you (dis)liked them, why?"},"title":{"matchLevel":"none","matchedWords":[],"value":"Ask HN: What is your experience with Applicant Tracking Systems?"}},"_tags":["story","author_endymi0n","story_18383055","ask_hn"],"author":"endymi0n","created_at":"2018-11-05T15:34:17Z","created_at_i":1541432057,"num_comments":0,"objectID":"18383055","points":2,"story_id":18383055,"story_text":"Hi HN,
I'm currently thinking about moving towards using an Applicant Tracking System for our company, but I feel a little unsure. As an applicant, I used to absolutely hate the rigid ATS systems of the first generation where instead of sending my CV as PDF, I had to fill in dozens of forms in order to just get back a cold standard mail after weeks.
That's why when we started our company, we went with a simple mailing list first. Thing is, we're slowly getting to the scale where this stops working and there are simply too many open positions and people working on them cluttering the list. I'd really like to have features such as human or tech-assisted pre-categorization to a job ad, multiple people in the pipeline, handover and just making sure we're answering everybody.
On the other hand, I'd really like to work with a system that doesn't have much more of a hurdle for applicants than sending a simple e-mail and that doesn't get in the way an interaction between humans. Bonus points for good integration in third-party software.
Has anyone used any ATS systems in the recent past and if you (dis)liked them, why?","title":"Ask HN: What is your experience with Applicant Tracking Systems?","updated_at":"2024-09-20T03:22:43Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"bquaresma"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"I have been seeing many blogs taking Reddit screenshots for adding them into their blogs so I decided to create a tool for that that allows some customization and at some point, could be integrated with third party APIs."},"title":{"matchLevel":"none","matchedWords":[],"value":"Show HN: Take Reddit Screenshots"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://www.redditpik.com"}},"_tags":["story","author_bquaresma","story_36324073","show_hn"],"author":"bquaresma","children":[36324148],"created_at":"2023-06-14T11:06:07Z","created_at_i":1686740767,"num_comments":2,"objectID":"36324073","points":1,"story_id":36324073,"story_text":"I have been seeing many blogs taking Reddit screenshots for adding them into their blogs so I decided to create a tool for that that allows some customization and at some point, could be integrated with third party APIs.","title":"Show HN: Take Reddit Screenshots","updated_at":"2024-09-20T14:21:33Z","url":"https://www.redditpik.com"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"georgel"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"I have been working on my startup since August, and released our app (Gruburg) in November. The day after it got released, Google HotPot was announced. Since what I made was basically the same as Google's new product, my spirits were destroyed.
Fast forward to today. At this point I am on the third version of the app, with social features and other things. I am even giving people money to use the app! Nothing, stuck with just organic(read: slow) growth. Foursquare releases their update today and its 1:1 with my app features. Why would anyone download another geolocation app, if they already have one that friends use?
Spirits destroyed, I am thinking of killing Gruburg. Being first (or having grand ideas) means nothing to me now, regardless of execution."},"title":{"matchLevel":"none","matchedWords":[],"value":"Why being first doesn't matter."},"url":{"matchLevel":"none","matchedWords":[],"value":""}},"_tags":["story","author_georgel","story_2307285","ask_hn"],"author":"georgel","children":[2307435,2307448],"created_at":"2011-03-10T00:14:13Z","created_at_i":1299716053,"num_comments":2,"objectID":"2307285","points":1,"story_id":2307285,"story_text":"I have been working on my startup since August, and released our app (Gruburg) in November. The day after it got released, Google HotPot was announced. Since what I made was basically the same as Google's new product, my spirits were destroyed.
Fast forward to today. At this point I am on the third version of the app, with social features and other things. I am even giving people money to use the app! Nothing, stuck with just organic(read: slow) growth. Foursquare releases their update today and its 1:1 with my app features. Why would anyone download another geolocation app, if they already have one that friends use?
Spirits destroyed, I am thinking of killing Gruburg. Being first (or having grand ideas) means nothing to me now, regardless of execution.","title":"Why being first doesn't matter.","updated_at":"2023-09-06T19:28:19Z","url":""},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"idontlikemyjob"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"I've been at three companies in just over a year and I'm already at the point where I don't like the third one. I don't know why this keeps happening, but I hope it's something I can fix.
The projects that I work on at these jobs aren't intellectually challenging to me. In addition to that, I seem to be stuck as a front end engineer. When I joined my current company, I made it clear that I want to do backend work, but I haven't done anything but frontend since I started.
There are certain problems that are definitely my fault; the least of which is my inability to deal with other people. Discussions about which design is better are pointless to me. Why don't we just try both and let the data decide? Meetings that are announced at the last second are irritating at best, because they distract me from my work, but I'm required to join anyways.
I want to work with people that are a lot better than me, but (at the risk of sounding arrogant) each time I move to a new company, my team seems to get worse. I'm looking for a place where there is a high level of rigour involved in the software development process.
How can I figure out why I can't enjoy any job that I take? How can I tell if I am the problem? How can I stop doing front end work? How can I identify a good team before accepting the job? Is it bad that I'm moving through jobs so quickly?
Any advice or thoughts are appreciated."},"title":{"matchLevel":"none","matchedWords":[],"value":"Ask HN: How can I enjoy my job?"}},"_tags":["story","author_idontlikemyjob","story_8225523","ask_hn"],"author":"idontlikemyjob","children":[8225614,8225639,8225640,8225649,8225751,8225828,8226183,8226200,8226289,8226533,8228421],"created_at":"2014-08-26T03:11:50Z","created_at_i":1409022710,"num_comments":30,"objectID":"8225523","points":21,"story_id":8225523,"story_text":"I've been at three companies in just over a year and I'm already at the point where I don't like the third one. I don't know why this keeps happening, but I hope it's something I can fix.
The projects that I work on at these jobs aren't intellectually challenging to me. In addition to that, I seem to be stuck as a front end engineer. When I joined my current company, I made it clear that I want to do backend work, but I haven't done anything but frontend since I started.
There are certain problems that are definitely my fault; the least of which is my inability to deal with other people. Discussions about which design is better are pointless to me. Why don't we just try both and let the data decide? Meetings that are announced at the last second are irritating at best, because they distract me from my work, but I'm required to join anyways.
I want to work with people that are a lot better than me, but (at the risk of sounding arrogant) each time I move to a new company, my team seems to get worse. I'm looking for a place where there is a high level of rigour involved in the software development process.
How can I figure out why I can't enjoy any job that I take? How can I tell if I am the problem? How can I stop doing front end work? How can I identify a good team before accepting the job? Is it bad that I'm moving through jobs so quickly?
Any advice or thoughts are appreciated.","title":"Ask HN: How can I enjoy my job?","updated_at":"2024-09-19T21:04:53Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"phs318u"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"I am a long-in-the-tooth IT professional working at a bank. I have just had yet another issue with our company's SOE (seems to be DNS config related, but of course I have no admin rights to try anything). I understand the need for locked down corporate environments but why-oh-why are most of them built by people who don't actually understand (or test) what they're doing? We regularly get "security" changes pushed through which break various bits of "legit" corporate internet as well as some applications we are permitted to use. I should point out our company uses a third-party to manage our fleet of PCs. My question to all is - has someone come up with a "SOE as a Service"? There's got to be a market for something like this where companies can basically sign up to, where the service is provided by those who actually know what they are doing. I tried searching for this and didn't really find anything turnkey."},"title":{"matchLevel":"none","matchedWords":[],"value":"Ask HN: SOE as a Service?"}},"_tags":["story","author_phs318u","story_29100905","ask_hn"],"author":"phs318u","children":[29101684],"created_at":"2021-11-03T22:00:03Z","created_at_i":1635976803,"num_comments":1,"objectID":"29100905","points":1,"story_id":29100905,"story_text":"I am a long-in-the-tooth IT professional working at a bank. I have just had yet another issue with our company's SOE (seems to be DNS config related, but of course I have no admin rights to try anything). I understand the need for locked down corporate environments but why-oh-why are most of them built by people who don't actually understand (or test) what they're doing? We regularly get "security" changes pushed through which break various bits of "legit" corporate internet as well as some applications we are permitted to use. I should point out our company uses a third-party to manage our fleet of PCs. My question to all is - has someone come up with a "SOE as a Service"? There's got to be a market for something like this where companies can basically sign up to, where the service is provided by those who actually know what they are doing. I tried searching for this and didn't really find anything turnkey.","title":"Ask HN: SOE as a Service?","updated_at":"2024-09-20T09:42:44Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"pwthornton"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["third","point"],"value":"\"Apple genuinely thinks that FCPX is a better piece of software than FCP7. It has the best technological foundation of any video editing software out there and is built for the 64-bit, multi-core, OpenCL world we are now entering. It is certainly more Pro than FCP7 in terms of technology and speed, and in many ways that\u2019s how Apple is defining Pro.
Looking at the new FCP, it seems fairly clear to me that Apple wants FCPX to be a solid foundation that works for most people\u2019s video editing needs. If you need more power or more features, Apple is pointing you in the direction of third-party plugins. It\u2019s very clear from Apple\u2019s literature about FCPX that third-party plugins are meant to bridge many of the gaps missing from FCPX.\""},"title":{"matchLevel":"none","matchedWords":[],"value":"Apple attempting to redefine Pro with FCPX"},"url":{"matchLevel":"none","matchedWords":[],"value":"http://interchangeproject.org/2011/07/05/apple-attempting-to-redefine-pro-and-usability-with-fcpx/"}},"_tags":["story","author_pwthornton","story_2751719"],"author":"pwthornton","created_at":"2011-07-11T17:52:48Z","created_at_i":1310406768,"num_comments":0,"objectID":"2751719","points":1,"story_id":2751719,"story_text":"\"Apple genuinely thinks that FCPX is a better piece of software than FCP7. It has the best technological foundation of any video editing software out there and is built for the 64-bit, multi-core, OpenCL world we are now entering. It is certainly more Pro than FCP7 in terms of technology and speed, and in many ways that\u2019s how Apple is defining Pro.
Looking at the new FCP, it seems fairly clear to me that Apple wants FCPX to be a solid foundation that works for most people\u2019s video editing needs. If you need more power or more features, Apple is pointing you in the direction of third-party plugins. It\u2019s very clear from Apple\u2019s literature about FCPX that third-party plugins are meant to bridge many of the gaps missing from FCPX.\"","title":"Apple attempting to redefine Pro with FCPX","updated_at":"2024-09-19T17:51:06Z","url":"http://interchangeproject.org/2011/07/05/apple-attempting-to-redefine-pro-and-usability-with-fcpx/"}],"hitsPerPage":50,"nbHits":522,"nbPages":11,"page":0,"params":"query=Third+Point&tags=story&hitsPerPage=50&advancedSyntax=true&analyticsTags=backend","processingTimeMS":18,"processingTimingsMS":{"_request":{"roundTrip":19},"afterFetch":{"format":{"highlighting":3,"total":3},"merge":{"mergeLoop":{"prepareNextHit":1,"total":1},"total":2},"total":2},"fetch":{"query":5,"scanning":8,"total":14},"total":18},"query":"Third Point","serverTimeMS":22}