16:[[49080905,{"by":"ingve","descendants":12,"id":49080905,"kids":[49083972,49082108,49083614,49084129,49081292,49081286,49081282,49083294],"score":12,"time":1785226678,"title":"If AI Writes All the Code, What Do the Programmers Do?","type":"story","url":"https://probablydance.com/2026/07/27/if-ai-writes-all-the-code-what-do-the-programmers-do/"}],[48942000,{"by":"haritha1313","descendants":201,"id":48942000,"kids":[48943090,48942719,48944121,48944339,48944608,48943218,48946147,48942787,48942683,48942998,48945752,48944528,48943428,48944416,48942757,48942943,48943234,48945202,48943836,48943326,48950937,48944552,48946887,48943517,48943017,48945728,48943978,48946588,48946420,48943898,48943453,48947642,48946428,48943665,48943958,48945488,48947640,48942866,48948520,48946330,48946408,48946457,48944572,48944408,48945746,48944322,48945407,48945762,48946476,48943508,48942881,48945164,48944431,48977101,48942833],"score":318,"time":1784247669,"title":"The human-in-the-loop is tired","type":"story","url":"https://pydantic.dev/articles/the-human-in-the-loop-is-tired"}],[48922706,{"by":"nkov47","descendants":25,"id":48922706,"kids":[48945397,48925709,48967936,48931932,48931581,48933379,48930682,48927259,48924123,48930801,48932059,48932333],"score":44,"text":"$17","time":1784130680,"title":"Launch HN: Coasty (YC S26) – An API for computer-use agents","type":"story","url":"https://coasty.ai/docs"}],[48930363,{"by":"cratermoon","descendants":113,"id":48930363,"kids":[48931289,48930798,48930820,48930983,48930840,48931016,48930870,48933772,48931035,48931947,48931110,48931332,48931442,48930955,48931376,48932334,48931432,48931704,48931545,48933264,48934809,48931917,48931024,48932911,48930376,48931040,48935115,48931068,48931202,48930934,48930791,48930874,48930979,48930710,48930932,48931314,48932767,48931393,48930960,48930769],"score":106,"time":1784176261,"title":"Stop saying that AI is just a tool and it only matters how it is used","type":"story","url":"https://www.frank.computer/blog/2025/05/just-a-tool.html"}],[48916975,{"by":"macleginn","descendants":294,"id":48916975,"kids":[48918331,48917262,48917641,48917802,48917841,48918774,48917314,49080309,48919590,48918395,48917456,48917293,48917437,48919473,48919445,48921759,48917540,48918683,48917265,48917781,48942264,48917531,48917380,48919258,48917950,48917324,48917687,48917639,48920606,48925715,48924776,48926943,48918217,48917471,48919016,48917961,48922090,48919181,48917316,48926993,48917768,48919285,48917579,48919525,48918735,48920286,48918549,48918283,48917493,48919868,48917419,48918568,48922899,48919266,48918315,48917748,48917090,48918317,48919332,48919471,48917423,48917366,48917398],"score":674,"time":1784096880,"title":"I tricked Claude into leaking your deepest, darkest secrets","type":"story","url":"https://www.ayush.digital/blog/the-memory-heist"}],[48771515,{"by":"yehiaabdelm","descendants":206,"id":48771515,"kids":[48779784,48780140,48779017,48782072,48771907,48779404,48780818,48771922,48774103,48772037,48772023,48779308,48772274,48780664,48771844,48771558,48771908,48780254,48772049,48771846,48772031,48773615,48780323,48781568,48772067,48779870,48774635,48813788,48782539,48779198,48772655,48772050,48779159,48779429,48780572,48771813,48865107,48801622,48779519,48780725,48776921,48780615,48772171,48780665,48771940,48779774,48780635,48780512,48781102,48774147,48773463,48861468,48773801,48779629,48772253,48779307,48780148,48780663,48781607,48772551,48814494,48779953,48772025,48779985,48781823,48780687,48771836,48771910,48779833,48771969,48773778,48840418,48779215,48780877,48781699,48772203,48771683,48804058,48781629,48790124,48772113,48771785,48772201,48786954,48772226,48782641,48771738,48790263,48779375,48785505,48772149,48780190,48780621,48779767,48779892,48772247,48786289,48771982,48773409,48772574,48831094,48806894,48782296,48780544,48785971,48794004,48782408,48772013,48771644,48772527,48800273,48772632,48788537,48792568,48791893,48771945,48812754,48793079,48799908,48780848,48787253,48787133,48781470,48862733,48784499,48780079,48780259,48772301],"score":212,"text":"I'm a bit annoyed by the feeling that we're kind of stuck when it comes to using LLMs for programming.

I use Claude Code and Codex, but I haven't been able to enter flow state like I can when I hand write code.

This is kind of ironic to me since AI should be a bicycle for the mind, but right now it feels like a bicycle that just brakes abruptly every couple minutes. I stop, wait, review, prompt again.

Is there anyone exploring something fundamentally different than the prompt response loop we have today?

I actually think the idea of a tab model is directionally better than prompt response.

Would love to hear about any startups, personal experiments, etc.","time":1783059704,"title":"Ask HN: Is anyone experimenting with different ways of using LLMs for coding?","type":"story"}],[48821500,{"by":"kirillklimuk","descendants":31,"id":48821500,"kids":[48821501,48822474,48824324,48825299,48825693,48822051,48822761,48822404,48823351,48825055,48827517,48831228,48822168,48827208,48831086,48824190,48824553],"score":70,"time":1783448340,"title":"Show HN: Docx-CLI: agents read/edit Word docs using 1/2 the time and tokens","type":"story","url":"https://github.com/kklimuk/docx-cli"}],[48818098,{"by":"brian_kuan","descendants":23,"id":48818098,"kids":[48824444,48822572,48825755,48823094,48823224,48824362,48824467,48828464,48823122,48914148,48831080,48825558],"score":37,"text":"$18","time":1783433228,"title":"Show HN: Halo – open-source, tamper-evident runtime evidence for AI agents","type":"story","url":"https://github.com/bkuan001/halo-record"}],[48773252,{"by":"steve_7890","descendants":0,"id":48773252,"kids":[48774649,48773253],"score":2,"time":1783073883,"title":"Beyond AI Prototyping: Why SSO, Audit Logs, and RBAC Matter in Production","type":"story","url":"https://geekyants.com/blog/beyond-ai-prototyping-sso-audit-logs-rbac"}],[48770319,{"by":"sollawen","descendants":53,"id":48770319,"kids":[48772794,48770958,48773125,48773528,48772685,48773804,48771999,48772239,48778768,48773914,48862742,48772314,48772221,48772575,48772403,48781296,48794080,48771956,48774363,48774238,48772103,48774938,48770429,48772706,48771049,48773788,48774644,48772435,48772310,48772479,48770775,48777038,48772857,48770534,48771518,48773535,48776934,48772241],"score":64,"text":"$19","time":1783048722,"title":"AI coding is a nightmare. Am I the only one experiencing this?","type":"story"}],[48762008,{"by":"edward","descendants":113,"id":48762008,"kids":[48765590,48767037,48765868,48766889,48765308,48765468,48766604,48767013,48768986,48765299,48767130,48768377,48765568,48774614,48770241,48767562,48765710,48766866],"score":120,"time":1783001843,"title":"No LLM Code in Dependencies","type":"story","url":"https://joeyh.name/blog/entry/no_LLM_code_in_dependencies/"}],[48741332,{"by":"arman-w-jalili","descendants":1,"id":48741332,"kids":[48757108,48747206,48741334],"score":14,"time":1782868661,"title":"Show HN: Coding agent that compiles intent into deterministic DAG before running","type":"story","url":"https://github.com/arman-jalili/rigorix-oss"}],[48714880,{"by":"tcballard","descendants":54,"id":48714880,"kids":[48715520,48716979,48715492,48715391,48716535,48716501,48715910,48715503,48718502,48715032,48734345,48715572,48720650,48715012,48716595,48728538],"score":47,"time":1782708035,"title":"Lore – Give your coding agent the decisions your team made","type":"story","url":"https://github.com/itsthelore/rac-core"}],[48712212,{"by":"chelm","descendants":12,"id":48712212,"kids":[48720295,48717253,48717763,48717401,48719202],"score":27,"time":1782684505,"title":"Why frontier LLMs can't read the hard documents without experts involved","type":"story","url":"https://idp-software.com/news/the-76-percent-wall/"}],[48643426,{"by":"ingve","descendants":110,"id":48643426,"kids":[48643871,48644574,48643958,48645922,48645275,48647872,48646125,48654438,48644370,48650140,48644071,48646031,48644456,48644968,48645408,48644493,48719193,48646018,48653946,48644229,48650510,48646656,48645137,48645180,48644232,48643798],"score":496,"time":1782214505,"title":"Unlimited OCR: One-shot long-horizon parsing","type":"story","url":"https://github.com/baidu/Unlimited-OCR"}],[48671675,{"by":"francisrafal","descendants":4,"id":48671675,"kids":[48677795,48696197,48718900],"score":9,"text":"I just released ParseHawk v0.1.0:\nApache-2.0 licensed 100% local document AI platform that extracts JSON from PDFs, images etc.\nIt builds on top of NuMind's NuExtract3 but additionally enforces a provided JSON schema with constrained decoding.\nIt works on Apple Silicon with pre-bundled vllm-metal as well as Linux + NVIDIA with vllm.\nLooking forward to your feedback!","time":1782385070,"title":"Show HN: ParseHawk – 100% Local Document AI with API, CLI, and Web UI","type":"story","url":"https://github.com/parsehawk/parsehawk"}],[48638999,{"by":"supernova85","descendants":0,"id":48638999,"kids":[48707401],"score":3,"text":"Founder here. I built this because many everyday notices are hard to understand: utility bills, medical bills, insurance letters, landlord notices, government letters, etc.

The app extracts a few practical fields from an uploaded PDF/image: who sent it, who it is for, deadlines, amount owed, what it seems to mean, and what should be checked against the original.

I know this is a sensitive area because people are uploading private documents. Current design choices:\n- users control retention/deletion\n- summaries are explicitly not legal/financial advice\n- failed or low-confidence model output should fail visibly rather than show fake certainty\n- duplicate uploads reuse prior successful results without reprocessing

The technical parts I’ve been working through are queue reliability, provider failures, document retention, quota accounting, and making the UI avoid “confident nonsense” when extraction fails.

I’m the builder and can answer implementation/privacy questions.","time":1782178157,"title":"Show HN: ExplainNotice: AI summary for bills, notices, and letters","type":"story","url":"https://explainnotice.com"}],[48704373,{"by":"handfuloflight","descendants":56,"id":48704373,"kids":[48705423,48705143,48760765,48705919,48704672,48705752,48706592,48705785,48705453,48706914,48707362,48706152,48709560,48705071,48706912,48705250,48705059,48706785,48706782,48706720,48707395,48708000,48706108,48709345,48704878,48707116,48705219,48704890,48705749,48705417],"score":123,"time":1782621078,"title":"Wayfinder Router: deterministic routing of queries between local and hosted LLM","type":"story","url":"https://github.com/itsthelore/wayfinder-router"}],[48689028,{"by":"minimaxir","descendants":744,"id":48689028,"kids":[48690510,48689959,48689193,48692734,48689256,48691353,48689338,48691244,48690399,48690497,48689874,48690077,48696432,48690722,48689806,48689107,48689480,48689541,48689498,48689226,48689402,48689183,48690813,48691041,48689140,48689104,48690861,48689131,48693522,48689152,48693377,48692383,48694700,48691231,48690472,48689111,48689474,48689653,48689080,48690678,48695183,48689857,48696397,48689061,48689968,48689563,48692941,48705968,48693237,48694505,48705969,48695212,48689141,48690009,48689386,48689941,48689529,48696232,48690262,48692024,48697149,48689293,48697896,48690713,48691238,48692124,48689281,48694597,48695589,48690810,48722002,48689844,48698999,48689227,48689384,48692334,48695699,48694422,48689408,48689816,48689660,48689114,48701592,48690215,48692195,48696827,48695707,48692264,48698973,48690220,48689116,48696460,48690150,48690485,48692842,48689584,48689572,48689225,48694985,48691592,48690767,48694508,48694506,48691601,48689323,48690254,48698273,48698106,48694283,48695168,48689742,48694516,48696173,48690387,48707522,48707390,48689473,48690665,48690512,48697026,48689594,48698385,48689123,48690619,48690451,48689510,48690649,48689222,48689713,48695909,48690837,48690747],"score":1139,"text":"System card: https://deploymentsafety.openai.com/gpt-5-6-preview","time":1782493615,"title":"Previewing GPT‑5.6 Sol: a next-generation model","type":"story","url":"https://openai.com/index/previewing-gpt-5-6-sol/"}],[48696916,{"by":"bonigv","descendants":5,"id":48696916,"kids":[48697025,48697019,48701239,48705369,48698266,48717010],"score":1,"text":"$1a","time":1782555524,"title":"Ask HN: How do we measure software in LLM era?","type":"story"}],[48686050,{"by":"RicoElectrico","descendants":55,"id":48686050,"kids":[48686431,48686292,48686252,48686260,48686589,48686558,48686445,48686273,48686791,48686623,48686256,48686563,48686331,48686588,48686970,48686428,48698262,48686251],"score":63,"time":1782478270,"title":"Google hallucinated that I am sponsored by Ground News","type":"story","url":"https://www.youtube.com/post/UgkxsAIUDMQTYc15yQqURsGy0Kg7O1x2TtnB"}],[48665700,{"by":"gabert","descendants":1,"id":48665700,"kids":[48666889,48680943,48665701],"score":3,"time":1782335524,"title":"Show HN: An LLM agent that emits typed intent","type":"story","url":"https://github.com/gabert/ontocortex"}],[48631726,{"by":"zdgeier","descendants":189,"id":48631726,"kids":[48635235,48633378,48633789,48634278,48634005,48634727,48633590,48638979,48632283,48632097,48637166,48643995,48634670,48632088,48638897,48635091,48634099,48635246,48639197,48635998,48632096,48639310,48632163,48639948,48632118,48632198,48633224,48635343,48635486,48638809,48636535,48632427,48632364,48634290,48633870,48636024,48632961,48633871,48642703,48638779,48640076,48644168,48633355,48633734,48636387,48632170,48636831,48642992,48646703,48644206,48641215,48632216,48632325,48642341,48636162,48645192,48636999,48641253,48725462,48642738,48642334,48639069,48642574,48641434],"score":216,"text":"$1b","time":1782142668,"title":"Show HN: Oak – Git alternative designed for agents","type":"story","url":"https://oak.space/oak/oak"}]]

user:	madikz
created:	October 20, 2022
karma:	3
about:	Building AI-powered document processing for US tax preparation. MIT Sloan MBA. 16+ years in enterprise IT & digital asset management (pharma). Interested in OCR, document AI, workflow automation, regulated industries. https://taxscout.ai