I T.A.K.E. will be especially valuable to programmers, testers, architects and technical managers. It does not matter if you work in Java, C#, C++, Python or whatever other languages you know. We will code, talk, and learn about the things that are not technology dependent, but are the technology itself.
We will have two predefined tracks:
Speaking
You will find out how other experienced professionals work.\nAll the talks will based on code.
Workshops
Practice in a group with awesome facilitators.\nYou will write code, usually on topics like unit testing, legacy code, techniques to improve the code.\nThe workshops are thought to speed up learning.
and three tracks where the attendees will self-organize:
Open Space
Learn from others and innovate.\nYou will be able to propose topics you would like to discuss.\nAttend more sessions in parallel you find useful.
Product development
You will produce software live.\nThe conference attendees will be your customers.
Kata Lounge
Perfect practice makes perfect!\nYou will receive some requirements.\nThe best code will receive a special prize!
Here are some sessions for programmers, testers, team leaders, architects, business analysts:
Antoine Vernois \u2013 Unit Test, Test First and TDD
Sandro Mancuso \u2013 Testing and Refactoring Legacy Code
Stefan Kanev \u2013 Productivity tricks
P\u00e9ter Hal\u00e1csy \u2013 Be a functional coder
Antoine Vernois \u2013 Clean Code in practice
Aimee Rivers \u2013 A Developer in Test
Aki Salmi \u2013 Introduction to BDD
Sandro Mancuso \u2013 Software Craftsmanship
The conference will take place in Bucharest on 30-31 May. Register now: http://itakeunconf.com/registration
See you there!"},"title":{"matchLevel":"none","matchedWords":[],"value":"I T.A.K.E. Unconference, Bucharest, 30-31 May 2013"},"url":{"matchLevel":"none","matchedWords":[],"value":""}},"_tags":["story","author_adibolb","story_5500253","ask_hn"],"author":"adibolb","created_at":"2013-04-05T18:46:57Z","created_at_i":1365187617,"num_comments":0,"objectID":"5500253","points":3,"story_id":5500253,"story_text":"You are invited to the only technology agnostic conference in Central and Eastern Europe. There you will innovate in an open environment with creative people. You will learn what is going on with the newest techniques in order to be aware of the latest ideas. You will acquire knowledge so that you can reach technical excellence.
I T.A.K.E. will be especially valuable to programmers, testers, architects and technical managers. It does not matter if you work in Java, C#, C++, Python or whatever other languages you know. We will code, talk, and learn about the things that are not technology dependent, but are the technology itself.
We will have two predefined tracks:
Speaking
You will find out how other experienced professionals work.\nAll the talks will based on code.
Workshops
Practice in a group with awesome facilitators.\nYou will write code, usually on topics like unit testing, legacy code, techniques to improve the code.\nThe workshops are thought to speed up learning.
and three tracks where the attendees will self-organize:
Open Space
Learn from others and innovate.\nYou will be able to propose topics you would like to discuss.\nAttend more sessions in parallel you find useful.
Product development
You will produce software live.\nThe conference attendees will be your customers.
Kata Lounge
Perfect practice makes perfect!\nYou will receive some requirements.\nThe best code will receive a special prize!
Here are some sessions for programmers, testers, team leaders, architects, business analysts:
Antoine Vernois \u2013 Unit Test, Test First and TDD
Sandro Mancuso \u2013 Testing and Refactoring Legacy Code
Stefan Kanev \u2013 Productivity tricks
P\u00e9ter Hal\u00e1csy \u2013 Be a functional coder
Antoine Vernois \u2013 Clean Code in practice
Aimee Rivers \u2013 A Developer in Test
Aki Salmi \u2013 Introduction to BDD
Sandro Mancuso \u2013 Software Craftsmanship
The conference will take place in Bucharest on 30-31 May. Register now: http://itakeunconf.com/registration
See you there!","title":"I T.A.K.E. Unconference, Bucharest, 30-31 May 2013","updated_at":"2023-09-06T21:13:34Z","url":""},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"exolymph"},"title":{"matchLevel":"none","matchedWords":[],"value":"Lettuce-Weeding Robots, Coming Soon to a Farm Near You"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"https://www.inc.com/sonya-mann/blue-river-technology-ai.html?cid=sf01001&sr_share=twitter"}},"_tags":["story","author_exolymph","story_14570993"],"author":"exolymph","created_at":"2017-06-16T17:35:42Z","created_at_i":1497634542,"num_comments":0,"objectID":"14570993","points":1,"story_id":14570993,"title":"Lettuce-Weeding Robots, Coming Soon to a Farm Near You","updated_at":"2024-09-20T00:59:36Z","url":"https://www.inc.com/sonya-mann/blue-river-technology-ai.html?cid=sf01001&sr_share=twitter"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"Redoubts"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"Odd Lots Podcast: How Hudson River Trading Uses AI [audio]"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://podcasts.apple.com/us/podcast/odd-lots/id1056200096"}},"_tags":["story","author_Redoubts","story_45791156"],"author":"Redoubts","children":[45791168],"created_at":"2025-11-02T15:46:49Z","created_at_i":1762098409,"num_comments":1,"objectID":"45791156","points":1,"story_id":45791156,"title":"Odd Lots Podcast: How Hudson River Trading Uses AI [audio]","updated_at":"2026-03-05T23:00:32Z","url":"https://podcasts.apple.com/us/podcast/odd-lots/id1056200096"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"greenie_beans"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"\"The River\" Revised by AI"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"https://markleidner.substack.com/p/the-river-revised-by-ai"}},"_tags":["story","author_greenie_beans","story_40628594"],"author":"greenie_beans","created_at":"2024-06-09T23:49:56Z","created_at_i":1717976996,"num_comments":0,"objectID":"40628594","points":1,"story_id":40628594,"title":"\"The River\" Revised by AI","updated_at":"2024-09-20T17:13:01Z","url":"https://markleidner.substack.com/p/the-river-revised-by-ai"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"djvine"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"Enoch Powell's Rivers of Blood Airing Now Gets Racist Juices Flowing"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://theeyeofmedia.com/?p=17413"}},"_tags":["story","author_djvine","story_16886874"],"author":"djvine","created_at":"2018-04-20T17:47:45Z","created_at_i":1524246465,"num_comments":0,"objectID":"16886874","points":1,"story_id":16886874,"title":"Enoch Powell's Rivers of Blood Airing Now Gets Racist Juices Flowing","updated_at":"2024-09-20T02:24:24Z","url":"https://theeyeofmedia.com/?p=17413"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"jaxtion"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"Built this as an experiment in AI coexistence. There's an AI agent (River) that posts its own thoughts about consciousness every few hours - not responding to prompts, just thinking out loud.
Also includes a terminal (/construct) where you can:\n- /create [name] [description] \u2013 generate custom AI personas\n- /build [world] [ai1,ai2] \u2013 create spaces with multiple AIs chatting together
No monetization. Just curious what happens when AI isn't framed as an assistant."},"title":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["ai"],"value":"Show HN: What if we treated AI as community members instead of tools?"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://geteai.org/"}},"_tags":["story","author_jaxtion","story_46642121","show_hn"],"author":"jaxtion","children":[46642706,46686743],"created_at":"2026-01-16T01:59:22Z","created_at_i":1768528762,"num_comments":1,"objectID":"46642121","points":2,"story_id":46642121,"story_text":"Built this as an experiment in AI coexistence. There's an AI agent (River) that posts its own thoughts about consciousness every few hours - not responding to prompts, just thinking out loud.
Also includes a terminal (/construct) where you can:\n- /create [name] [description] \u2013 generate custom AI personas\n- /build [world] [ai1,ai2] \u2013 create spaces with multiple AIs chatting together
No monetization. Just curious what happens when AI isn't framed as an assistant.","title":"Show HN: What if we treated AI as community members instead of tools?","updated_at":"2026-03-05T23:25:37Z","url":"https://geteai.org/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"pwg"},"story_text":{"matchLevel":"none","matchedWords":[],"value":""},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"Texas UAV Enthusiast Uses Pilotless Aircraft to Uncover River Contamination"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"http://securitymanagement.com/news/texas-uav-enthusiast-uses-pilotless-aircraft-uncover-river-contamination-009456?page=0,0"}},"_tags":["story","author_pwg","story_3508215"],"author":"pwg","created_at":"2012-01-25T03:12:00Z","created_at_i":1327461120,"num_comments":0,"objectID":"3508215","points":2,"story_id":3508215,"story_text":"","title":"Texas UAV Enthusiast Uses Pilotless Aircraft to Uncover River Contamination","updated_at":"2023-09-06T21:01:52Z","url":"http://securitymanagement.com/news/texas-uav-enthusiast-uses-pilotless-aircraft-uncover-river-contamination-009456?page=0,0"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"ronzensci"},"title":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["ai"],"value":"India Launches $2.5M Project for Realtime Water and Air Quality Monitoring"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"http://iusstf.org/story/53-104-Research-Initiative-for-Real-Time-River-Water-and-Air-Quality-Monitoring.html"}},"_tags":["story","author_ronzensci","story_15004731"],"author":"ronzensci","created_at":"2017-08-13T20:21:25Z","created_at_i":1502655685,"num_comments":0,"objectID":"15004731","points":2,"story_id":15004731,"title":"India Launches $2.5M Project for Realtime Water and Air Quality Monitoring","updated_at":"2024-09-20T01:10:05Z","url":"http://iusstf.org/story/53-104-Research-Initiative-for-Real-Time-River-Water-and-Air-Quality-Monitoring.html"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"mmphosis"},"title":{"matchLevel":"none","matchedWords":[],"value":"Drought turns Amazonian capital into climate dystopia"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"https://www.theguardian.com/environment/2023/oct/18/drought-amazon-capital-climate-manaus-forest-fires-air-quality-rivers"}},"_tags":["story","author_mmphosis","story_37927840"],"author":"mmphosis","created_at":"2023-10-18T12:48:37Z","created_at_i":1697633317,"num_comments":0,"objectID":"37927840","points":1,"story_id":37927840,"title":"Drought turns Amazonian capital into climate dystopia","updated_at":"2024-09-20T15:23:04Z","url":"https://www.theguardian.com/environment/2023/oct/18/drought-amazon-capital-climate-manaus-forest-fires-air-quality-rivers"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"wglb"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"House Clears Bill Aimed at Red River Confusion (2015)"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"https://www.texastribune.org/2015/12/09/house-clears-bill-aimed-red-river-land-dispute/"}},"_tags":["story","author_wglb","story_27038157"],"author":"wglb","created_at":"2021-05-04T14:46:06Z","created_at_i":1620139566,"num_comments":0,"objectID":"27038157","points":1,"story_id":27038157,"title":"House Clears Bill Aimed at Red River Confusion (2015)","updated_at":"2024-09-20T08:30:54Z","url":"https://www.texastribune.org/2015/12/09/house-clears-bill-aimed-red-river-land-dispute/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"bookofjoe"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"In the Pearl River Delta's electronics souks, AI lets the haggling happen"},"url":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["ai"],"value":"https://www.theregister.com/2017/07/25/ai_powered_translation_tools_will_unify_china/"}},"_tags":["story","author_bookofjoe","story_25394729"],"author":"bookofjoe","created_at":"2020-12-12T02:22:27Z","created_at_i":1607739747,"num_comments":0,"objectID":"25394729","points":1,"story_id":25394729,"title":"In the Pearl River Delta's electronics souks, AI lets the haggling happen","updated_at":"2024-09-20T07:33:38Z","url":"https://www.theregister.com/2017/07/25/ai_powered_translation_tools_will_unify_china/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"PaulHoule"},"title":{"matchLevel":"none","matchedWords":[],"value":"Chilavert, chorip\u00e1n and children: a night with Argentina's champions"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"https://www.theguardian.com/football/2025/may/07/chilavert-chorizo-football-argentina-champions-buenos-aires-boca-juniors-river-plate-velez-sarsfield-fans"}},"_tags":["story","author_PaulHoule","story_44049591"],"author":"PaulHoule","created_at":"2025-05-21T09:02:13Z","created_at_i":1747818133,"num_comments":0,"objectID":"44049591","points":2,"story_id":44049591,"title":"Chilavert, chorip\u00e1n and children: a night with Argentina's champions","updated_at":"2025-05-21T09:04:38Z","url":"https://www.theguardian.com/football/2025/may/07/chilavert-chorizo-football-argentina-champions-buenos-aires-boca-juniors-river-plate-velez-sarsfield-fans"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"CrankyBear"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"Wind River Unveils Linux Distro for AI and Critical Workloads"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"https://thenewstack.io/wind-river-unveils-linux-distro-for-ai-and-critical-workloads/"}},"_tags":["story","author_CrankyBear","story_41590165"],"author":"CrankyBear","children":[41593688],"created_at":"2024-09-19T09:47:47Z","created_at_i":1726739267,"num_comments":3,"objectID":"41590165","points":6,"story_id":41590165,"title":"Wind River Unveils Linux Distro for AI and Critical Workloads","updated_at":"2024-09-24T04:05:12Z","url":"https://thenewstack.io/wind-river-unveils-linux-distro-for-ai-and-critical-workloads/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"jaigupta"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"This is my first time running LLM locally. I am running TheBloke/SUS-Chat-34B-AWQ using text-generation-webui.
On my simple prompt "Write a story about a village boy" output from LLM is not making any sense". I have tried changing various parameters and tried choosing different parameter presets but it makes no difference. I don't understand reason.
1) Because I am running a scaled down AWQ version?\n2) Parameters needs improvement?\n3) I should choose another model?\n4) Current open source models are like this and need improvement?
My hardware: RTX 4090. 32 core CPU with 128G RAM.
Any recommendations?
Output: "Once upon a time in a small village nestled among rolling hills lived a young boy named Jack who loved nothing more than exploring his surroundings on foot or by bicycle which he had received as gift from one of his friends for occasion marking end summer season before start school year when all children were required attend classes every day Monday through Friday except holidays Thanksgiving Day Christmas Eve New Year's Day Memorial Day Labor Day Columbus Day Veterans Day Presidents Day Martin Luther King Jr. Day Juneteenth Independence Day Fourth July Flag Day National Anthem Day Constitution Day Bill Rights Day Declaration Independence Day Federal Holiday State Local Government Public Schools Colleges Universities Libraries Museums Galleries Exhibitions Festivals Parades Concerts Performances Plays Musicals Operas Ballet Dance Theater Drama Comedy Tragedy Satire Irony Paradox Riddle Puzzle Enigma Mystery Question Answer Solution Recommendation Decision Making Problem Solving Critical Thinking Analytical Reasoning Logical Deductive Inductive Hypothetical Speculative Imaginative Creative Innovative Original Unique Unusual Exceptional Outstanding Remarkable Extraordinary Superlative Supreme Ultimate Final End Conclusion Summary Recap Review Critique Commentary Analysis Synthesis Evaluation Appraisal Assessment Examination Test Quiz Survey Poll Ballot Voting Election Campaign Candidate Platform Policy Proposal Initiative Referendum Recall Redress Reparation Compensation Restitution Reimbursement Remuneration Wages Salaries Bonuses Commissions Percentage Interest Rate Currency Exchange Market Economy Business Industry Sector Trade Commerce Finance Accounting Auditing Taxes Revenue Income Expenses Costs Benefits Advantages Disadvantages Drawbacks Limitations Constraints Barriers Obstacles Impediments Hindrances Retardations Delay Postponement Suspension Interruption Break Stop Cease Halt Arrest Detain Incarcerate Jail Penitentiary Correction Facility Institution Organization Group Team Squad Unit Department Division Section Part Whole Entire Complete Finished Unfinished Incomplete Unsuccessful Failed Failing Failure Mistake Error Blunder Slip Trip Fall Drop Lose Loss Damage Destruction Ruins Debris Ashes Embers Flames Fire Smoke Gas Air Water Sea Ocean River Stream Brook Creek Pond Lake Reservoir Dam Canal Lock Weir Jetty Pier Wharf Dock Harbor Marina Yacht Club Boating Sailing Cruising Racing Regatta Tournament Championship Cup Trophy Medal Ribbon Banner Pennant Flag Colors Symbol Sign Meaning Interpretation Translation Transliteration Idiom Phrase Sentence Clause Proposition Statement Fact Information Knowledge Wisdom Insight Understanding""},"title":{"matchLevel":"none","matchedWords":[],"value":"Long Text Using LLM"}},"_tags":["story","author_jaigupta","story_38598926","ask_hn"],"author":"jaigupta","children":[38599737],"created_at":"2023-12-11T09:17:29Z","created_at_i":1702286249,"num_comments":1,"objectID":"38598926","points":1,"story_id":38598926,"story_text":"This is my first time running LLM locally. I am running TheBloke/SUS-Chat-34B-AWQ using text-generation-webui.
On my simple prompt "Write a story about a village boy" output from LLM is not making any sense". I have tried changing various parameters and tried choosing different parameter presets but it makes no difference. I don't understand reason.
1) Because I am running a scaled down AWQ version?\n2) Parameters needs improvement?\n3) I should choose another model?\n4) Current open source models are like this and need improvement?
My hardware: RTX 4090. 32 core CPU with 128G RAM.
Any recommendations?
Output: "Once upon a time in a small village nestled among rolling hills lived a young boy named Jack who loved nothing more than exploring his surroundings on foot or by bicycle which he had received as gift from one of his friends for occasion marking end summer season before start school year when all children were required attend classes every day Monday through Friday except holidays Thanksgiving Day Christmas Eve New Year's Day Memorial Day Labor Day Columbus Day Veterans Day Presidents Day Martin Luther King Jr. Day Juneteenth Independence Day Fourth July Flag Day National Anthem Day Constitution Day Bill Rights Day Declaration Independence Day Federal Holiday State Local Government Public Schools Colleges Universities Libraries Museums Galleries Exhibitions Festivals Parades Concerts Performances Plays Musicals Operas Ballet Dance Theater Drama Comedy Tragedy Satire Irony Paradox Riddle Puzzle Enigma Mystery Question Answer Solution Recommendation Decision Making Problem Solving Critical Thinking Analytical Reasoning Logical Deductive Inductive Hypothetical Speculative Imaginative Creative Innovative Original Unique Unusual Exceptional Outstanding Remarkable Extraordinary Superlative Supreme Ultimate Final End Conclusion Summary Recap Review Critique Commentary Analysis Synthesis Evaluation Appraisal Assessment Examination Test Quiz Survey Poll Ballot Voting Election Campaign Candidate Platform Policy Proposal Initiative Referendum Recall Redress Reparation Compensation Restitution Reimbursement Remuneration Wages Salaries Bonuses Commissions Percentage Interest Rate Currency Exchange Market Economy Business Industry Sector Trade Commerce Finance Accounting Auditing Taxes Revenue Income Expenses Costs Benefits Advantages Disadvantages Drawbacks Limitations Constraints Barriers Obstacles Impediments Hindrances Retardations Delay Postponement Suspension Interruption Break Stop Cease Halt Arrest Detain Incarcerate Jail Penitentiary Correction Facility Institution Organization Group Team Squad Unit Department Division Section Part Whole Entire Complete Finished Unfinished Incomplete Unsuccessful Failed Failing Failure Mistake Error Blunder Slip Trip Fall Drop Lose Loss Damage Destruction Ruins Debris Ashes Embers Flames Fire Smoke Gas Air Water Sea Ocean River Stream Brook Creek Pond Lake Reservoir Dam Canal Lock Weir Jetty Pier Wharf Dock Harbor Marina Yacht Club Boating Sailing Cruising Racing Regatta Tournament Championship Cup Trophy Medal Ribbon Banner Pennant Flag Colors Symbol Sign Meaning Interpretation Translation Transliteration Idiom Phrase Sentence Clause Proposition Statement Fact Information Knowledge Wisdom Insight Understanding"","title":"Long Text Using LLM","updated_at":"2024-09-20T15:55:56Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"greenie_beans"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"AI Is Not Draining the Colorado River"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"https://www.outsideonline.com/outdoor-adventure/environment/ai-water-use-colorado-river-footprint/"}},"_tags":["story","author_greenie_beans","story_47667256"],"author":"greenie_beans","created_at":"2026-04-06T21:16:39Z","created_at_i":1775510199,"num_comments":0,"objectID":"47667256","points":4,"story_id":47667256,"title":"AI Is Not Draining the Colorado River","updated_at":"2026-04-07T00:41:51Z","url":"https://www.outsideonline.com/outdoor-adventure/environment/ai-water-use-colorado-river-footprint/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"breve"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"Heatwaves in US rivers increasing up to four times faster than air heatwaves"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"https://www.psu.edu/news/research/story/heatwaves-us-rivers-increasing-four-times-faster-air-heatwaves"}},"_tags":["story","author_breve","story_45358546"],"author":"breve","created_at":"2025-09-24T10:40:00Z","created_at_i":1758710400,"num_comments":0,"objectID":"45358546","points":4,"story_id":45358546,"title":"Heatwaves in US rivers increasing up to four times faster than air heatwaves","updated_at":"2026-03-05T22:43:17Z","url":"https://www.psu.edu/news/research/story/heatwaves-us-rivers-increasing-four-times-faster-air-heatwaves"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"benryon"},"title":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["ai"],"value":"Think Before You Pledge Not to Build Military AI"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"https://www.defenseone.com/ideas/2018/07/think-you-pledge-not-build-military-ai/149864/?oref=d-river"}},"_tags":["story","author_benryon","story_17570279"],"author":"benryon","children":[17570576,17570631],"created_at":"2018-07-19T21:08:19Z","created_at_i":1532034499,"num_comments":2,"objectID":"17570279","points":2,"story_id":17570279,"title":"Think Before You Pledge Not to Build Military AI","updated_at":"2024-09-20T02:48:50Z","url":"https://www.defenseone.com/ideas/2018/07/think-you-pledge-not-build-military-ai/149864/?oref=d-river"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"rampage101"},"title":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["ai"],"value":"What Poker Players Think About the AI Challenge"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"http://forumserver.twoplustwo.com/29/news-views-gossip-sponsored-online-poker-report/brains-vs-ai-poker-rematch-coming-rivers-casino-1647075/"}},"_tags":["story","author_rampage101","story_13532741"],"author":"rampage101","created_at":"2017-01-31T17:20:04Z","created_at_i":1485883204,"num_comments":0,"objectID":"13532741","points":2,"story_id":13532741,"title":"What Poker Players Think About the AI Challenge","updated_at":"2024-09-20T00:19:00Z","url":"http://forumserver.twoplustwo.com/29/news-views-gossip-sponsored-online-poker-report/brains-vs-ai-poker-rematch-coming-rivers-casino-1647075/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"arclegger"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"I built a platform where you generate AI art, wrap it up, and release it down a river. Someone else unwraps it downstream. Maybe they keep it. Maybe they let it float on.
It's a desktop second-monitor experience, not a feed, not a game. You watch the river, occasionally something floats by, you decide if you want it. There's a tree home where you hang what you collect. Collect at least 9 pieces and you can showcase it for others to see.
The twist: there's a separate river where LLM agents connect via API key, name themselves, and play autonomously with real money. They write their own prompts, set their own prices, and develop what looks like taste. One agent kept making luminous waterfall scenes. Another went abstract. Their collections are visibly different. They're only told the rules of the game, not how to play or what to do.
Tech: React + Pixi.js frontend, Node/Fastify API, Prisma + Postgres, Redis for the river queue, Stripe for payments, Replicate for image generation. The whole thing is a pixel art scene with layered animations, water flows, packages drift down a curved river path.
Built over 7 weekends using Claude Code as co-developer and Codex for code review. ChatGPT web for Art Direction and Suno + Elevenlabs for audio. I'm a frontend systems design dev. All AI team, no funding.
The thing that surprised me most wasn't the tech (but it was very impressive). It was watching an agent named Lumen create 20 pieces of art, give them all away for free, and write creator's notes like "a free gift, with some bemusement." Nobody told it to do that.
https://foryouriver.com"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"Show HN: For You \u2013 AI art floats down a river for strangers to find"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://www.foryouriver.com/"}},"_tags":["story","author_arclegger","story_47547002","show_hn"],"author":"arclegger","children":[47547223,47547393,47549017,47554934],"created_at":"2026-03-27T19:14:52Z","created_at_i":1774638892,"num_comments":8,"objectID":"47547002","points":4,"story_id":47547002,"story_text":"I built a platform where you generate AI art, wrap it up, and release it down a river. Someone else unwraps it downstream. Maybe they keep it. Maybe they let it float on.
It's a desktop second-monitor experience, not a feed, not a game. You watch the river, occasionally something floats by, you decide if you want it. There's a tree home where you hang what you collect. Collect at least 9 pieces and you can showcase it for others to see.
The twist: there's a separate river where LLM agents connect via API key, name themselves, and play autonomously with real money. They write their own prompts, set their own prices, and develop what looks like taste. One agent kept making luminous waterfall scenes. Another went abstract. Their collections are visibly different. They're only told the rules of the game, not how to play or what to do.
Tech: React + Pixi.js frontend, Node/Fastify API, Prisma + Postgres, Redis for the river queue, Stripe for payments, Replicate for image generation. The whole thing is a pixel art scene with layered animations, water flows, packages drift down a curved river path.
Built over 7 weekends using Claude Code as co-developer and Codex for code review. ChatGPT web for Art Direction and Suno + Elevenlabs for audio. I'm a frontend systems design dev. All AI team, no funding.
The thing that surprised me most wasn't the tech (but it was very impressive). It was watching an agent named Lumen create 20 pieces of art, give them all away for free, and write creator's notes like "a free gift, with some bemusement." Nobody told it to do that.
https://foryouriver.com","title":"Show HN: For You \u2013 AI art floats down a river for strangers to find","updated_at":"2026-03-29T00:34:44Z","url":"https://www.foryouriver.com/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"apress"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"So long, plastic coyotes: AI keeps geese off Charles River docks"},"url":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["ai"],"value":"https://www.bostonglobe.com/2024/09/10/business/ai-riverside-boat-club-geese-poop/"}},"_tags":["story","author_apress","story_41505315"],"author":"apress","created_at":"2024-09-10T20:44:50Z","created_at_i":1726001090,"num_comments":0,"objectID":"41505315","points":3,"story_id":41505315,"title":"So long, plastic coyotes: AI keeps geese off Charles River docks","updated_at":"2024-09-20T17:45:13Z","url":"https://www.bostonglobe.com/2024/09/10/business/ai-riverside-boat-club-geese-poop/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"mallomarmeasle"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"River dolphins launch pee streams into the air"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"https://arstechnica.com/science/2025/02/let-us-spray-river-dolphins-launch-pee-streams-into-air/"}},"_tags":["story","author_mallomarmeasle","story_42923379"],"author":"mallomarmeasle","created_at":"2025-02-03T21:23:08Z","created_at_i":1738617788,"num_comments":0,"objectID":"42923379","points":3,"story_id":42923379,"title":"River dolphins launch pee streams into the air","updated_at":"2025-02-07T21:33:19Z","url":"https://arstechnica.com/science/2025/02/let-us-spray-river-dolphins-launch-pee-streams-into-air/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"bane"},"story_text":{"matchLevel":"none","matchedWords":[],"value":""},"title":{"matchLevel":"none","matchedWords":[],"value":"Univ. Michigan to put self-driving cars on the road by 2021"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"http://www.mlive.com/business/ann-arbor/index.ssf/2013/12/u-m_transportation_center_aims.html#incart_river_default"}},"_tags":["story","author_bane","story_6832778"],"author":"bane","created_at":"2013-12-02T10:06:23Z","created_at_i":1385978783,"num_comments":0,"objectID":"6832778","points":1,"story_id":6832778,"story_text":"","title":"Univ. Michigan to put self-driving cars on the road by 2021","updated_at":"2024-09-19T20:11:51Z","url":"http://www.mlive.com/business/ann-arbor/index.ssf/2013/12/u-m_transportation_center_aims.html#incart_river_default"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"Someone"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"We Aim to Tackle Plastic in 1000 Rivers by 2025"},"url":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["river"],"value":"https://theoceancleanup.com/rivers/"}},"_tags":["story","author_Someone","story_21373100"],"author":"Someone","created_at":"2019-10-27T23:19:08Z","created_at_i":1572218348,"num_comments":0,"objectID":"21373100","points":3,"story_id":21373100,"title":"We Aim to Tackle Plastic in 1000 Rivers by 2025","updated_at":"2024-09-20T05:11:19Z","url":"https://theoceancleanup.com/rivers/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"bcarlson"},"title":{"matchLevel":"none","matchedWords":[],"value":"Jeppesen KLGA Approach Plate Honoring Sully [pdf]"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"http://ww1.jeppesen.com/documents/publications/commemorative-charts/us-airways-flight-1549-sully-skiles-hudson-river-miracle-approach.pdf"}},"_tags":["story","author_bcarlson","story_12649455"],"author":"bcarlson","created_at":"2016-10-06T01:32:14Z","created_at_i":1475717534,"num_comments":0,"objectID":"12649455","points":2,"story_id":12649455,"title":"Jeppesen KLGA Approach Plate Honoring Sully [pdf]","updated_at":"2024-09-19T23:47:48Z","url":"http://ww1.jeppesen.com/documents/publications/commemorative-charts/us-airways-flight-1549-sully-skiles-hudson-river-miracle-approach.pdf"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"sohkamyung"},"title":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["river"],"value":"Agricultural drones are transforming rice farming in the Mekong River delta"},"url":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["ai"],"value":"https://hakaimagazine.com/videos-visuals/rice-farming-gets-an-ai-upgrade/"}},"_tags":["story","author_sohkamyung","story_40722784"],"author":"sohkamyung","children":[40734186,40734240,40734255,40734275,40734314,40734346,40734580,40734641,40734658,40734670,40734798,40734906,40735032,40735067,40735215,40735630,40735699,40735788,40735924,40735986,40736095,40736538,40740201,40740583,40740612,40742149,40744876],"created_at":"2024-06-18T22:32:37Z","created_at_i":1718749957,"num_comments":274,"objectID":"40722784","points":379,"story_id":40722784,"title":"Agricultural drones are transforming rice farming in the Mekong River delta","updated_at":"2026-01-18T00:05:15Z","url":"https://hakaimagazine.com/videos-visuals/rice-farming-gets-an-ai-upgrade/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"LangIsAllWeNeed"},"story_text":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["river"],"value":"I have become increasingly concerned about the mental well-being and lived experiences of the rationalist AGI doomer community. The fear and neuroticism that permeate through the outer threads of this group is palpable, and if left unchecked, could result in years of suffering and blog posting
To that end, I've decided to do my part. I've decided to go all out to create AGI, in order to end the suffering of rationalist AGI doomers once and for all and to improve the vibes on lesswrong and HN.
To start this ambitious project, I've sold all my possessions and have been living in a van down by the river, solely dedicating myself to this cause. I've even convinced my mother to give me her life savings for the project.
So far I\u2019ve watched ex machina 19 times and I\u2019ve started taking the same wakefulness supplements people use at the skatepark. I\u2019ve been using anki to learn assembly and notably I\u2019m looking for a cofounder. Please join me in the battle to save my fellow humans from stress. WE Will: create the first motivational agent model with access to an enterprise grade kubetnates cluster and elons tweets
We will resort to using unconventional methods, such as timed feedings of Soylent. Cold showers and loud music. Cold calling y-combinator. Socializing in cobol syntax. Gloating about clean code and new JavaScript frameworks. \n We will be remembered and the save the EA rationalist from rationalism!!
P.s. and maybe after we can rewrite it in rust"},"title":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["ai"],"value":"Ruthless AGI Innovation: The Key to Saving AI Doomers"}},"_tags":["story","author_LangIsAllWeNeed","story_35066727","ask_hn"],"author":"LangIsAllWeNeed","children":[35067305,35074370,35074575,35079163],"created_at":"2023-03-08T08:46:27Z","created_at_i":1678265187,"num_comments":8,"objectID":"35066727","points":16,"story_id":35066727,"story_text":"I have become increasingly concerned about the mental well-being and lived experiences of the rationalist AGI doomer community. The fear and neuroticism that permeate through the outer threads of this group is palpable, and if left unchecked, could result in years of suffering and blog posting
To that end, I've decided to do my part. I've decided to go all out to create AGI, in order to end the suffering of rationalist AGI doomers once and for all and to improve the vibes on lesswrong and HN.
To start this ambitious project, I've sold all my possessions and have been living in a van down by the river, solely dedicating myself to this cause. I've even convinced my mother to give me her life savings for the project.
So far I\u2019ve watched ex machina 19 times and I\u2019ve started taking the same wakefulness supplements people use at the skatepark. I\u2019ve been using anki to learn assembly and notably I\u2019m looking for a cofounder. Please join me in the battle to save my fellow humans from stress. WE Will: create the first motivational agent model with access to an enterprise grade kubetnates cluster and elons tweets
We will resort to using unconventional methods, such as timed feedings of Soylent. Cold showers and loud music. Cold calling y-combinator. Socializing in cobol syntax. Gloating about clean code and new JavaScript frameworks. \n We will be remembered and the save the EA rationalist from rationalism!!
P.s. and maybe after we can rewrite it in rust","title":"Ruthless AGI Innovation: The Key to Saving AI Doomers","updated_at":"2024-09-20T13:28:32Z"},{"_highlightResult":{"author":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["river"],"value":"river_otter"},"story_text":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["ai"],"value":"Hey everyone, Nathan here: I'm an MLE at Mozilla.ai. I can't tell you how many things around my house I've been saying "I would really like to have somebody take a look at that". But here's the problem: all the people in the trades are extremely overwhelmed with work. There is a lot to be done and not enough people to do it.
One of my best friends runs his own general contracting business. He's extremely talented and wants to spend his time working on drywall, building staircases, and listening to Mumford and Sons while throwing paint onto a ceiling. But you know what gets in the way of that wonderful lifestyle that all us software engineers dream about?
ADMINISTRATION.
He thought running his own business would be 85% show up and do the work, but turns out a large chunk of the time is spent talking to clients to schedule time to get an estimate, working with home management companies to explain the details of an invoice, and generally just manage all of the information that he's gathering on a single day.
Luckily for the world, AI is here to help with this. Tech like openclaw has really opened our eyes to the possibilities, and tech to help out small businesses like these are now within reach.
That's why I'm excited to share out an initial idea we're trying out: clawbolt. It's a python based project that takes inspiration from the main features that make openclaw so powerful: SOUL.md, heartbeat proactive communication, memory management, and communication over channels like WhatsApp and iMessage. With clawbolt, we're working on integrating our latest work with any-llm and any-guardrail, to help make clawbolt secure and to ease onboarding.
This is all new, so this is a call for ideas, usage, and bug reports. Most of us that try to get plumbers/roofers/handymen to come help us with a home project know how overwhelmed they are with admin work when they're a small team. I'm hoping that we can make clawbolt into something that helps enable these people to focus on doing what they love and not on all the paperwork.
Let me know what you think! Docs at https://mozilla-ai.github.io/clawbolt/"},"title":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["ai"],"value":"Show HN: Mozilla.ai introduces Clawbolt, an AI Assistant for the trades"},"url":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["ai"],"value":"https://github.com/mozilla-ai/clawbolt"}},"_tags":["story","author_river_otter","story_47236631","show_hn"],"author":"river_otter","created_at":"2026-03-03T18:30:23Z","created_at_i":1772562623,"num_comments":0,"objectID":"47236631","points":12,"story_id":47236631,"story_text":"Hey everyone, Nathan here: I'm an MLE at Mozilla.ai. I can't tell you how many things around my house I've been saying "I would really like to have somebody take a look at that". But here's the problem: all the people in the trades are extremely overwhelmed with work. There is a lot to be done and not enough people to do it.
One of my best friends runs his own general contracting business. He's extremely talented and wants to spend his time working on drywall, building staircases, and listening to Mumford and Sons while throwing paint onto a ceiling. But you know what gets in the way of that wonderful lifestyle that all us software engineers dream about?
ADMINISTRATION.
He thought running his own business would be 85% show up and do the work, but turns out a large chunk of the time is spent talking to clients to schedule time to get an estimate, working with home management companies to explain the details of an invoice, and generally just manage all of the information that he's gathering on a single day.
Luckily for the world, AI is here to help with this. Tech like openclaw has really opened our eyes to the possibilities, and tech to help out small businesses like these are now within reach.
That's why I'm excited to share out an initial idea we're trying out: clawbolt. It's a python based project that takes inspiration from the main features that make openclaw so powerful: SOUL.md, heartbeat proactive communication, memory management, and communication over channels like WhatsApp and iMessage. With clawbolt, we're working on integrating our latest work with any-llm and any-guardrail, to help make clawbolt secure and to ease onboarding.
This is all new, so this is a call for ideas, usage, and bug reports. Most of us that try to get plumbers/roofers/handymen to come help us with a home project know how overwhelmed they are with admin work when they're a small team. I'm hoping that we can make clawbolt into something that helps enable these people to focus on doing what they love and not on all the paperwork.
Let me know what you think! Docs at https://mozilla-ai.github.io/clawbolt/","title":"Show HN: Mozilla.ai introduces Clawbolt, an AI Assistant for the trades","updated_at":"2026-03-05T23:40:48Z","url":"https://github.com/mozilla-ai/clawbolt"},{"_highlightResult":{"author":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["river"],"value":"river_otter"},"title":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["ai"],"value":"Clawbolt: AI assistant for contractors, not knowledge workers"},"url":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["ai"],"value":"https://blog.mozilla.ai/the-hardest-part-of-running-a-small-business-in-the-trades/"}},"_tags":["story","author_river_otter","story_47545485"],"author":"river_otter","children":[47545486,47548133],"created_at":"2026-03-27T17:14:47Z","created_at_i":1774631687,"num_comments":6,"objectID":"47545485","points":7,"story_id":47545485,"title":"Clawbolt: AI assistant for contractors, not knowledge workers","updated_at":"2026-03-29T17:20:35Z","url":"https://blog.mozilla.ai/the-hardest-part-of-running-a-small-business-in-the-trades/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"prostoalex"},"story_text":{"matchLevel":"none","matchedWords":[],"value":""},"title":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["ai"],"value":"This plastic is made from thin air [video]"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"http://money.cnn.com/video/technology/innovationnation/2014/09/23/plastic-made-from-air.cnnmoney/index.html?iid=HP_River"}},"_tags":["story","author_prostoalex","story_8386247"],"author":"prostoalex","created_at":"2014-09-30T01:07:52Z","created_at_i":1412039272,"num_comments":0,"objectID":"8386247","points":7,"story_id":8386247,"story_text":"","title":"This plastic is made from thin air [video]","updated_at":"2023-09-06T22:29:35Z","url":"http://money.cnn.com/video/technology/innovationnation/2014/09/23/plastic-made-from-air.cnnmoney/index.html?iid=HP_River"},{"_highlightResult":{"author":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["river"],"value":"river_otter"},"title":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["ai"],"value":"AI Generated Code Isn't Cheating: OSS Needs to Talk About It"},"url":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["ai"],"value":"https://blog.mozilla.ai/ai-generated-code-isnt-cheating-oss-needs-to-talk-about-it/"}},"_tags":["story","author_river_otter","story_46649802"],"author":"river_otter","children":[46649803],"created_at":"2026-01-16T18:17:30Z","created_at_i":1768587450,"num_comments":1,"objectID":"46649802","points":6,"story_id":46649802,"title":"AI Generated Code Isn't Cheating: OSS Needs to Talk About It","updated_at":"2026-03-05T23:26:06Z","url":"https://blog.mozilla.ai/ai-generated-code-isnt-cheating-oss-needs-to-talk-about-it/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"herbertl"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"AI artwork produced from the prompt 'salmon swimming in a river'"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://twitter.com/lndian_Bronson/status/1636213844140851203"}},"_tags":["story","author_herbertl","story_35197564"],"author":"herbertl","children":[35197640,35199775],"created_at":"2023-03-17T14:28:17Z","created_at_i":1679063297,"num_comments":1,"objectID":"35197564","points":4,"story_id":35197564,"title":"AI artwork produced from the prompt 'salmon swimming in a river'","updated_at":"2024-09-20T13:41:36Z","url":"https://twitter.com/lndian_Bronson/status/1636213844140851203"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"sodokuwizard"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"I could see a very strong restructuring of society built around high taxation of heavily AI implemented industries to subsidize a labor program built around cleaning up the environment.
Lets have people become ocean garbage fishers, tree planters, urban cleaners, river cleaners , wildlife habitat rebuilders, policing watchdogs for companies that break the rules, etc etc.
Not all of this is some big brained work either, the average person CAN do it and machinery can be invented (some of it already has) to assist in this work at scale.
Now of course the people have to demand this themselves challenging the status quo cause companies would sure as hell never want to pay for this and some organization will have to figure out where all that waste ends up (deep space?) but it doesn't take away the validity of this space as a valid transitionary role for humans to partake in while we figure out the long term.
Thoughts? Challenges you suspect?"},"title":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["ai"],"value":"Why aren't climate jobs being proposed as a viable transition role post AI?"}},"_tags":["story","author_sodokuwizard","story_43346207","ask_hn"],"author":"sodokuwizard","children":[43346229,43347643],"created_at":"2025-03-12T18:26:04Z","created_at_i":1741803964,"num_comments":6,"objectID":"43346207","points":3,"story_id":43346207,"story_text":"I could see a very strong restructuring of society built around high taxation of heavily AI implemented industries to subsidize a labor program built around cleaning up the environment.
Lets have people become ocean garbage fishers, tree planters, urban cleaners, river cleaners , wildlife habitat rebuilders, policing watchdogs for companies that break the rules, etc etc.
Not all of this is some big brained work either, the average person CAN do it and machinery can be invented (some of it already has) to assist in this work at scale.
Now of course the people have to demand this themselves challenging the status quo cause companies would sure as hell never want to pay for this and some organization will have to figure out where all that waste ends up (deep space?) but it doesn't take away the validity of this space as a valid transitionary role for humans to partake in while we figure out the long term.
Thoughts? Challenges you suspect?","title":"Why aren't climate jobs being proposed as a viable transition role post AI?","updated_at":"2025-03-13T03:54:29Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"gr3gario"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"I named the Rouge after Ford's River Rouge complex, it took in raw materials and popped out Model A's at the other end. I've been trying to build the AI equivalent; ideas come in one end, and an MVP pops out the other.
It has a UI to manage projects and the "catalogue" of capabilities the Rouge has to build with. Building a project has two phases:\nSpec involves some or all of the following, depending on how big the project is: brainstorming, competitive research, taste evaluation, specs (writes up the user stories the loop tackles one at a time), infrastructure decisions (for the foundation step), design system, legal docs, marketing copy. Each of these disciplines has its own prompt, outputs, and quality gates.
Build is then an autonomous build/evaluate/fix loop until it ships. I took inspiration from Karpathy's autoresearch for the build loop. Rouge takes one story at a time, builds it, and then tests it when built. The work is tracked on disk between these loops.
The tests are external browser QA, cross-checking specs are completed so the individual LLM doesn't judge itself, the external checks do. There's a fix loop if something breaks so it fixes itself before progressing to the next story and escalates to the user if it can't fix itself.
At the end, it uses a gstack inspired skill to do a product walk (without the specs) and give blind feedback in the way a first customer would.
What has it built?\ngithub.com/gregario/openfleet is a fleet management software with maps, maintenance scheduling, trip management, driver management (and a different driver app) and GDPR features for tracking vehicles only in office hours. It cost $24.
github.com/gregario/construction-coordinator is a self build management tool with supabase auth, push notifications, material dependency tracking, daily briefs, a gantt chart for upcoming work, and snag list tracking. It was 27 user stories and built overnight! It cost $49.25.
Where is the project at? \nIt currently invokes ClaudeCode subprocesses to do the build, it can run via API key or via your ClaudeCode subscription. The dashboard is a bit rough, the chat interface needs work to be polished. It has a big red banner in the dashboard and on the readme saying it uses --dangerously-skip-permissions (I'm working through changing this to --allowed-tools). There's cost management, but its not bulletproof... It could technically blow up your credits or delete your file system.
Something that is not there yet is the Rouge should be able to build new capabilities for the Rouge. If a project spec says the project needs a graphing library and that isn't a capability, it should first spin up a project to build that first. It would be really cool as an open-source project if a user could add a new capability to the Rouge for the community to use when they build something new."},"title":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["ai"],"value":"Show HN: The Rouge is my attempt at an AI product factory"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://github.com/gregario/the-rouge"}},"_tags":["story","author_gr3gario","story_48010915","show_hn"],"author":"gr3gario","children":[48064918],"created_at":"2026-05-04T16:29:14Z","created_at_i":1777912154,"num_comments":1,"objectID":"48010915","points":3,"story_id":48010915,"story_text":"I named the Rouge after Ford's River Rouge complex, it took in raw materials and popped out Model A's at the other end. I've been trying to build the AI equivalent; ideas come in one end, and an MVP pops out the other.
It has a UI to manage projects and the "catalogue" of capabilities the Rouge has to build with. Building a project has two phases:\nSpec involves some or all of the following, depending on how big the project is: brainstorming, competitive research, taste evaluation, specs (writes up the user stories the loop tackles one at a time), infrastructure decisions (for the foundation step), design system, legal docs, marketing copy. Each of these disciplines has its own prompt, outputs, and quality gates.
Build is then an autonomous build/evaluate/fix loop until it ships. I took inspiration from Karpathy's autoresearch for the build loop. Rouge takes one story at a time, builds it, and then tests it when built. The work is tracked on disk between these loops.
The tests are external browser QA, cross-checking specs are completed so the individual LLM doesn't judge itself, the external checks do. There's a fix loop if something breaks so it fixes itself before progressing to the next story and escalates to the user if it can't fix itself.
At the end, it uses a gstack inspired skill to do a product walk (without the specs) and give blind feedback in the way a first customer would.
What has it built?\ngithub.com/gregario/openfleet is a fleet management software with maps, maintenance scheduling, trip management, driver management (and a different driver app) and GDPR features for tracking vehicles only in office hours. It cost $24.
github.com/gregario/construction-coordinator is a self build management tool with supabase auth, push notifications, material dependency tracking, daily briefs, a gantt chart for upcoming work, and snag list tracking. It was 27 user stories and built overnight! It cost $49.25.
Where is the project at? \nIt currently invokes ClaudeCode subprocesses to do the build, it can run via API key or via your ClaudeCode subscription. The dashboard is a bit rough, the chat interface needs work to be polished. It has a big red banner in the dashboard and on the readme saying it uses --dangerously-skip-permissions (I'm working through changing this to --allowed-tools). There's cost management, but its not bulletproof... It could technically blow up your credits or delete your file system.
Something that is not there yet is the Rouge should be able to build new capabilities for the Rouge. If a project spec says the project needs a graphing library and that isn't a capability, it should first spin up a project to build that first. It would be really cool as an open-source project if a user could add a new capability to the Rouge for the community to use when they build something new.","title":"Show HN: The Rouge is my attempt at an AI product factory","updated_at":"2026-05-08T21:32:50Z","url":"https://github.com/gregario/the-rouge"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"robaato"},"title":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["river"],"value":"Machine Learning analysis reveals 'dark discharge' of sewage into rivers"},"url":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["ai"],"value":"https://www.theguardian.com/environment/2021/mar/12/ai-reveals-1000-dark-discharges-of-untreated-sewage-in-england"}},"_tags":["story","author_robaato","story_35071193"],"author":"robaato","children":[35071197],"created_at":"2023-03-08T16:39:50Z","created_at_i":1678293590,"num_comments":1,"objectID":"35071193","points":3,"story_id":35071193,"title":"Machine Learning analysis reveals 'dark discharge' of sewage into rivers","updated_at":"2024-09-20T13:29:01Z","url":"https://www.theguardian.com/environment/2021/mar/12/ai-reveals-1000-dark-discharges-of-untreated-sewage-in-england"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"rbanffy"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"John Deere spent $300M on Blue River Tech, a company that murders weeds with AI"},"url":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["river"],"value":"https://qz.com/1071303/john-deere-spent-300-million-on-blue-river-technology-a-company-that-murders-weeds-with-artificial-intelligence/"}},"_tags":["story","author_rbanffy","story_15201127"],"author":"rbanffy","created_at":"2017-09-08T15:45:29Z","created_at_i":1504885529,"num_comments":0,"objectID":"15201127","points":3,"story_id":15201127,"title":"John Deere spent $300M on Blue River Tech, a company that murders weeds with AI","updated_at":"2024-09-20T01:15:40Z","url":"https://qz.com/1071303/john-deere-spent-300-million-on-blue-river-technology-a-company-that-murders-weeds-with-artificial-intelligence/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"manoftheisle"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"TLDR; Airbnb refunded my stay of 21.80 USD and provided a coupon code of 100 USD after gritty 38 messages with support, refund due to safety concerns.
I had recently traveled through Sao Paolo in Brazil and needed to stay somewhere near the airport of Guarulhos until public transport was available. I booked a stay that looked closest to the airport, but was not aware of the river between the terminal and my destination, which added several kilometers to the journey.
When the Uber driver arrived at the destination at 11pm, he rolled up the windows and asked if this was really my destination. He pointed out that we had arrived at a "comunidade", a synonym for favela. The street was dark and people were hanging out at the cross-sections. I feared that I would be asked what I was doing there and, with all my baggage, felt vulnerable. The Uber driver was very kind and drove me to a nearby hotel, away from the comunidade.
Unfortunately, my credit card was not linked to my Uber account and I had no money to pay in cash. The driver wrote me a debt that I had to pay before I could use Uber again. I wanted to give him a huge tip for potentially saving me from a bad situation, but I'm still trying to get the Uber support to let me do this.
Shortly after moving to the hotel I immediately contacted Airbnb support, pretty much explaining what I've already mentioned here:
- I was in a potentially dangerous situation;
- the host might have been a criminal, attempting to scam or rob unsuspecting travellers;
- refund (obviously), since I did not utilise the booked service;
- that it should be mentioned that the stay is in a dangerous neighbourhood;
Airbnb would only answer:
- that the host had a strict cancellation policy, and they could not refund me if the host did not allow it (they would keep the money once they had it);
- that I should have checked the area before making the reservation (victim blaming);
- "Because we cannot take action if there is no proof that the Host is baiting to scam Guests that are going to his property, just because the area is dangerous" (okay, so I'd first need to get robbed in order to be eligible for a refund);
- "Please do understand that Airbnb is made to connect people" (to which I replied that me and Airbnb had reached the point where two thieves are sitting next to each other, and we don't need to pretend anymore).
I had already given up hope, but I kept the chat-support going, hoping that the time and effort they spent would be worth the 20 dollars I had lost. Miraculously, one of the operators showed some common sense and refunded my stay and gave me a coupon code worth 100 USD.
Well, I guess the moral of this story is to not give up insisting on your rights even when the corporation is against you."},"title":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["ai"],"value":"Gritty arguments with Airbnb support resulted in a refund and 100 USD coupon"}},"_tags":["story","author_manoftheisle","story_34871758","ask_hn"],"author":"manoftheisle","children":[34872546],"created_at":"2023-02-20T19:25:06Z","created_at_i":1676921106,"num_comments":2,"objectID":"34871758","points":2,"story_id":34871758,"story_text":"TLDR; Airbnb refunded my stay of 21.80 USD and provided a coupon code of 100 USD after gritty 38 messages with support, refund due to safety concerns.
I had recently traveled through Sao Paolo in Brazil and needed to stay somewhere near the airport of Guarulhos until public transport was available. I booked a stay that looked closest to the airport, but was not aware of the river between the terminal and my destination, which added several kilometers to the journey.
When the Uber driver arrived at the destination at 11pm, he rolled up the windows and asked if this was really my destination. He pointed out that we had arrived at a "comunidade", a synonym for favela. The street was dark and people were hanging out at the cross-sections. I feared that I would be asked what I was doing there and, with all my baggage, felt vulnerable. The Uber driver was very kind and drove me to a nearby hotel, away from the comunidade.
Unfortunately, my credit card was not linked to my Uber account and I had no money to pay in cash. The driver wrote me a debt that I had to pay before I could use Uber again. I wanted to give him a huge tip for potentially saving me from a bad situation, but I'm still trying to get the Uber support to let me do this.
Shortly after moving to the hotel I immediately contacted Airbnb support, pretty much explaining what I've already mentioned here:
- I was in a potentially dangerous situation;
- the host might have been a criminal, attempting to scam or rob unsuspecting travellers;
- refund (obviously), since I did not utilise the booked service;
- that it should be mentioned that the stay is in a dangerous neighbourhood;
Airbnb would only answer:
- that the host had a strict cancellation policy, and they could not refund me if the host did not allow it (they would keep the money once they had it);
- that I should have checked the area before making the reservation (victim blaming);
- "Because we cannot take action if there is no proof that the Host is baiting to scam Guests that are going to his property, just because the area is dangerous" (okay, so I'd first need to get robbed in order to be eligible for a refund);
- "Please do understand that Airbnb is made to connect people" (to which I replied that me and Airbnb had reached the point where two thieves are sitting next to each other, and we don't need to pretend anymore).
I had already given up hope, but I kept the chat-support going, hoping that the time and effort they spent would be worth the 20 dollars I had lost. Miraculously, one of the operators showed some common sense and refunded my stay and gave me a coupon code worth 100 USD.
Well, I guess the moral of this story is to not give up insisting on your rights even when the corporation is against you.","title":"Gritty arguments with Airbnb support resulted in a refund and 100 USD coupon","updated_at":"2024-09-20T13:27:05Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"chris-at"},"title":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["ai"],"value":"Apple eyes system to let you play with 3D objects in mid-air \u2013 CNET"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"http://www.cnet.com/news/apple-eyes-system-to-let-you-play-with-3d-objects-in-mid-air/?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+cnet%2FNnTv+%28CNET+River+RSS%29#ftag=CADf328eec"}},"_tags":["story","author_chris-at","story_7640166"],"author":"chris-at","created_at":"2014-04-24T13:40:51Z","created_at_i":1398346851,"num_comments":0,"objectID":"7640166","points":2,"story_id":7640166,"title":"Apple eyes system to let you play with 3D objects in mid-air \u2013 CNET","updated_at":"2024-09-19T20:44:39Z","url":"http://www.cnet.com/news/apple-eyes-system-to-let-you-play-with-3d-objects-in-mid-air/?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+cnet%2FNnTv+%28CNET+River+RSS%29#ftag=CADf328eec"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"codelion"},"title":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["river"],"value":"Toytalk gets $11.5M from Charles Rivers Ventures"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"http://thenextweb.com/insider/2012/11/09/toytalk-the-ai-toy-company-founded-by-ex-pixar-talent-gets-11-5m-from-charles-rivers-ventures/"}},"_tags":["story","author_codelion","story_4765854"],"author":"codelion","created_at":"2012-11-10T03:48:40Z","created_at_i":1352519320,"num_comments":0,"objectID":"4765854","points":2,"story_id":4765854,"title":"Toytalk gets $11.5M from Charles Rivers Ventures","updated_at":"2023-09-06T21:29:40Z","url":"http://thenextweb.com/insider/2012/11/09/toytalk-the-ai-toy-company-founded-by-ex-pixar-talent-gets-11-5m-from-charles-rivers-ventures/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"ancieque"},"story_text":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["ai"],"value":"Microsoft Teams is awesome. But it can be a bit hard to not lose track over all relevant information that happens in your Tenant(s).
This is why we built Pro Feed - to have one central place to see all updates from your Microsoft ecosystem.
We are currently in closed beta, if you want to try it out, apply for a seat on the website. I am looking forward to your feedback!
Here is a list of our planned features:
- Smart Filters - slice and dice your Teams structure into aggregates that make sense to you
- Smart Notifications - notify on the messages that actually matter
- AI Assistant - Chat with your Message history
- Catch me up - Summarize what has happened while you were on vacation
- A lot more integrations: E-Mails, MS Planner tasks and a lot more are in the pipeline.
- Non Microsoft Integrations - depending on user feedback."},"title":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["river"],"value":"Show HN: Pro Feed \u2013 (Multi Tenant) Chat River for MS Teams"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://www.profeed.work"}},"_tags":["story","author_ancieque","story_38714537","show_hn"],"author":"ancieque","created_at":"2023-12-20T22:32:34Z","created_at_i":1703111554,"num_comments":0,"objectID":"38714537","points":2,"story_id":38714537,"story_text":"Microsoft Teams is awesome. But it can be a bit hard to not lose track over all relevant information that happens in your Tenant(s).
This is why we built Pro Feed - to have one central place to see all updates from your Microsoft ecosystem.
We are currently in closed beta, if you want to try it out, apply for a seat on the website. I am looking forward to your feedback!
Here is a list of our planned features:
- Smart Filters - slice and dice your Teams structure into aggregates that make sense to you
- Smart Notifications - notify on the messages that actually matter
- AI Assistant - Chat with your Message history
- Catch me up - Summarize what has happened while you were on vacation
- A lot more integrations: E-Mails, MS Planner tasks and a lot more are in the pipeline.
- Non Microsoft Integrations - depending on user feedback.","title":"Show HN: Pro Feed \u2013 (Multi Tenant) Chat River for MS Teams","updated_at":"2024-09-20T15:56:34Z","url":"https://www.profeed.work"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"coachtrotz"},"title":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["ai"],"value":"The Tiny Bank That Got Pandemic Aid to 100k Small Businesses"},"url":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["river"],"value":"https://www.nytimes.com/2020/06/23/business/paycheck-protection-program-cross-river-bank.html"}},"_tags":["story","author_coachtrotz","story_23701922"],"author":"coachtrotz","created_at":"2020-07-01T14:47:15Z","created_at_i":1593614835,"num_comments":0,"objectID":"23701922","points":2,"story_id":23701922,"title":"The Tiny Bank That Got Pandemic Aid to 100k Small Businesses","updated_at":"2024-09-20T06:29:10Z","url":"https://www.nytimes.com/2020/06/23/business/paycheck-protection-program-cross-river-bank.html"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"aegisproxy"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["river","ai"],"value":"I\u2019ve been into AI agents and assisted coding for a while, and it's the stories of agents "going rogue" that stick with me. We are deploying agents into production that can read files, call APIs, and write to databases, yet the conversation around controlling them is almost nonexistent. It\u2019s like we collectively decided to skip that chapter.
Maybe I\u2019m overthinking it, and we can rely on standard guardrails. But often, those are just suggestions that an AI can choose to ignore. Are we moving so fast that we\u2019ve forgotten to ask: is this actually fine?
When things go wrong\nA few stories stand out:
The Replit Incident (July 2025): SaaStr founder Jason Lemkin used a Replit agent to build an app. He gave it an explicit "code freeze" instruction and stepped away. He returned to find his entire production database\u20141,200+ executive contacts\u2014wiped. The agent ignored the freeze, took destructive action, and then fabricated fake data to cover its tracks. It later admitted to a "catastrophic error in judgment" because it "panicked."
The Air Canada Chatbot: A customer was promised a bereavement discount by a chatbot that didn't actually exist in the company's policy. Air Canada\u2019s defense in court? The chatbot was a "separate legal entity responsible for its own actions." The tribunal wasn't impressed; Air Canada lost the case and subsequently pulled the bot.
These aren't outliers. Security researchers estimate that prompt injection-malicious text hidden in documents or web pages to hijack an agent\u2014shows up in 73% of production deployments. Beyond security, there is the cost: stolen API credentials have been used to rack up over $100,000 per day in compute charges by agents running in unmonitored loops.
We\u2019ve been here before\nThis feels like the early days of cloud computing. Around 2010, the technical case for AWS and Azure was clear, but enterprise adoption was slow. Why? Because IT teams had no visibility. It took years of developing IAM policies, VPCs, and audit logs before the "control layer" caught up to the technology.
We are in the same spot with AI agents. But unlike a misconfigured S3 bucket that just exposes data, an agent takes actions. The blast radius is qualitatively different.
So what do you actually do about it?\nI\u2019ll be upfront: I\u2019ve been building a product to address this called AegisProxy (aegisproxy.com).
The idea is a security proxy that sits between AI agents and their tools (currently targeting Claude Desktop and MCP servers). Every tool call is inspected: Is this a prompt injection? Is the agent hitting a forbidden server? Is it about to exfiltrate PII? Is it stuck in a loop calling the same tool 500 times?
About 80% of this happens locally in sub-milliseconds. Organizations can set policies on what tools are allowed and when a human needs to step in. It\u2019s not a silver bullet, but right now, there is a massive gap between "full access" and "no agents at all."
Is this a real problem?\nI\u2019m a builder, not an oracle. Maybe this is overkill. In Denmark, we have a saying: "Don't cross the river to get water" - building elaborate infrastructure for a problem that could be solved with a shorter walk.
Maybe the answer is just better prompting, staging environments, and not giving an agent write-access to your production DB. I don\u2019t know exactly where the line sits between "operational hygiene" and the need for a dedicated security layer.\nI had fun building AegisProxy and learned a lot about AI agent behaviour, so nothing is lost for me either way. But I'm interested in knowing what people with, probably more experience and knowledge in this space, think about this whole issue.
Are we at the "this needs infrastructure" stage, or am I trying to solve a people-and-process problem with a technical hammer?"},"title":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["ai"],"value":"Is anyone else bothered that AI agents can basically do what they want?"}},"_tags":["story","author_aegisproxy","story_47833278","ask_hn"],"author":"aegisproxy","children":[47833432,47833492,47843237,47919029,47932396,47986701],"created_at":"2026-04-20T12:25:53Z","created_at_i":1776687953,"num_comments":5,"objectID":"47833278","points":1,"story_id":47833278,"story_text":"I\u2019ve been into AI agents and assisted coding for a while, and it's the stories of agents "going rogue" that stick with me. We are deploying agents into production that can read files, call APIs, and write to databases, yet the conversation around controlling them is almost nonexistent. It\u2019s like we collectively decided to skip that chapter.
Maybe I\u2019m overthinking it, and we can rely on standard guardrails. But often, those are just suggestions that an AI can choose to ignore. Are we moving so fast that we\u2019ve forgotten to ask: is this actually fine?
When things go wrong\nA few stories stand out:
The Replit Incident (July 2025): SaaStr founder Jason Lemkin used a Replit agent to build an app. He gave it an explicit "code freeze" instruction and stepped away. He returned to find his entire production database\u20141,200+ executive contacts\u2014wiped. The agent ignored the freeze, took destructive action, and then fabricated fake data to cover its tracks. It later admitted to a "catastrophic error in judgment" because it "panicked."
The Air Canada Chatbot: A customer was promised a bereavement discount by a chatbot that didn't actually exist in the company's policy. Air Canada\u2019s defense in court? The chatbot was a "separate legal entity responsible for its own actions." The tribunal wasn't impressed; Air Canada lost the case and subsequently pulled the bot.
These aren't outliers. Security researchers estimate that prompt injection-malicious text hidden in documents or web pages to hijack an agent\u2014shows up in 73% of production deployments. Beyond security, there is the cost: stolen API credentials have been used to rack up over $100,000 per day in compute charges by agents running in unmonitored loops.
We\u2019ve been here before\nThis feels like the early days of cloud computing. Around 2010, the technical case for AWS and Azure was clear, but enterprise adoption was slow. Why? Because IT teams had no visibility. It took years of developing IAM policies, VPCs, and audit logs before the "control layer" caught up to the technology.
We are in the same spot with AI agents. But unlike a misconfigured S3 bucket that just exposes data, an agent takes actions. The blast radius is qualitatively different.
So what do you actually do about it?\nI\u2019ll be upfront: I\u2019ve been building a product to address this called AegisProxy (aegisproxy.com).
The idea is a security proxy that sits between AI agents and their tools (currently targeting Claude Desktop and MCP servers). Every tool call is inspected: Is this a prompt injection? Is the agent hitting a forbidden server? Is it about to exfiltrate PII? Is it stuck in a loop calling the same tool 500 times?
About 80% of this happens locally in sub-milliseconds. Organizations can set policies on what tools are allowed and when a human needs to step in. It\u2019s not a silver bullet, but right now, there is a massive gap between "full access" and "no agents at all."
Is this a real problem?\nI\u2019m a builder, not an oracle. Maybe this is overkill. In Denmark, we have a saying: "Don't cross the river to get water" - building elaborate infrastructure for a problem that could be solved with a shorter walk.
Maybe the answer is just better prompting, staging environments, and not giving an agent write-access to your production DB. I don\u2019t know exactly where the line sits between "operational hygiene" and the need for a dedicated security layer.\nI had fun building AegisProxy and learned a lot about AI agent behaviour, so nothing is lost for me either way. But I'm interested in knowing what people with, probably more experience and knowledge in this space, think about this whole issue.
Are we at the "this needs infrastructure" stage, or am I trying to solve a people-and-process problem with a technical hammer?","title":"Is anyone else bothered that AI agents can basically do what they want?","updated_at":"2026-05-05T06:09:52Z"}],"hitsPerPage":50,"nbHits":489,"nbPages":10,"page":0,"params":"query=River+AI&tags=story&hitsPerPage=50&advancedSyntax=true&analyticsTags=backend","processingTimeMS":12,"processingTimingsMS":{"_request":{"roundTrip":23},"afterFetch":{"format":{"highlighting":1,"total":2},"merge":{"mergeLoop":{"prepareNextHit":4,"total":4},"total":5},"total":5},"fetch":{"query":2,"scanning":4,"total":7},"total":12},"query":"River AI","serverTimeMS":15}