[{"data":1,"prerenderedAt":584},["ShallowReactive",2],{"project-44022353":3},{"id":4,"name":5,"line":6,"line_source":7,"title":8,"url":9,"domain":10,"maker":11,"posted_at":12,"year":13,"points":14,"comments":15,"rank_in_year":16,"shown_in_year":17,"top":18,"tier":19,"category":20,"subcategory":21,"format":22,"keywords":23,"reception":29,"sentiment":30,"stance_mix":31,"scored_comments":35,"aspects":36,"strip":41,"post_text":42,"what":43,"said":75,"maker_in_thread":188,"shape":199,"grid":203,"attempts":549,"same_week":550,"maker_shows":34,"maker_hidden":577,"reader_version":578,"card":579},"44022353","Voynich NLP analysis","I modeled the Voynich Manuscript with SBERT to test for structure","title","Show HN: I modeled the Voynich Manuscript with SBERT to test for structure","https:\u002F\u002Fgithub.com\u002Fbrianmg\u002Fvoynich-nlp-analysis","github.com","brig90",1747584541,2025,381,132,124,3837,4,"legend","learning_science","linguistics research","other",[24,25,26,27,28],"voynich manuscript","sbert embeddings","nlp","kmeans clustering","markov transition","curious",0,{"positive":32,"neutral":33,"mixed":34,"critical":32},5,17,1,28,{"execution":37,"idea":38,"docs":40},{"positive":34,"critical":34},{"positive":39,"critical":18},3,{"positive":34,"critical":30},[39,39,34,34,39,39,30,39,30,30,30,30,30,39,30,32,39,30,30,32,30,30,30,39,30,39,30,34,39,30,30,30,30,30,39,30,32,30,30,30],"I built this project as a way to learn more about NLP by applying it to something weird and unsolved.\n\nThe Voynich Manuscript is a 15th-century book written in an unknown script. No one’s been able to translate it, and many think it’s a hoax, a cipher, or a constructed language. I wasn’t trying to decode it — I just wanted to see: does it behave like a structured language?\n\nI stripped a handful of common suffix-like endings (aiin, dy, etc.) to isolate what looked like root forms. I know that’s a strong assumption — I call it out directly in the repo — but it helped clarify the clustering. From",{"problem":44,"audience":50,"pricing":50,"motivation":51,"time_to_build":50,"stack":58,"openness":73,"stage":73,"team_size":74},{"value":45,"source":46,"evidence":47},"Tests whether the Voynich Manuscript behaves like a structured language","maker",[48],{"id":49},"post",null,{"value":52,"source":46,"evidence":53},"Learning NLP by applying it to an unsolved, weird problem",[54,55],{"id":49},{"id":56,"by":11,"text":57},"44026234","Honestly, I had never even heard of the manuscript before this weekend. I’ve been looking for interesting ways to strengthen my understanding of NLPs, and thought: 1) maybe this…",[59,65,69],{"value":60,"source":46,"evidence":61},"SBERT (paraphrase-multilingual-MiniLM-L12-v2)",[62],{"id":63,"by":11,"text":64},"44022749","Totally fair — I defaulted to paraphrase-multilingual-MiniLM-L12-v2 mostly for speed and wide compatibility, but you’re right that it’s long in the tooth by today’s standards.…",{"value":66,"source":46,"evidence":67},"KMeans clustering",[68],{"id":49},{"value":70,"source":46,"evidence":71},"Markov transition matrix",[72],{"id":49},"unknown","solo",{"praise":76,"objections":92,"requests":126,"questions":150,"alternatives":172},[77,85],{"value":78,"source":79,"evidence":80},"Interesting, nerdy deep-dive into the Voynich","commenter",[81],{"id":82,"by":83,"text":84},"44026036","pawanjswal","This is hands-down the nerdiest and coolest deep-dive into the Voynich I’ve seen.",{"value":86,"source":79,"evidence":87},"Fascinating approach to an unsolved puzzle",[88],{"id":89,"by":90,"text":91},"44031333","thearn4","Voynich is one of my favorite unsolved puzzles. This approach looks fascinating, so thanks for sharing your work here!",[93,100,108,115],{"value":94,"source":79,"evidence":95},"Embedding model is outdated and may not suit an unknown language",[96],{"id":97,"by":98,"text":99},"44022718","minimaxir","A point of note is that the text embeddings model used here is paraphrase-multilingual-MiniLM-L12-v2…",{"value":101,"source":79,"evidence":102},"Stripping suffixes removes information and may harm results",[103,104],{"id":97,"by":98,"text":99},{"id":105,"by":106,"text":107},"44029277","DonaldFisk","This is very interesting. You should post a link to https:\u002F\u002Fwww.voynich.ninja\u002Findex.php I'm not familiar with SBERT, or with modern statistical NLP in general, but SBERT works on…",{"value":109,"source":79,"evidence":110},"Trained-on-known-languages models cannot extract meaning from an unknown script",[111],{"id":112,"by":113,"text":114},"44117794","theRealEros","I feel like we are missing an important point...the transformer model that has been used here is trained on known languages. This means it cannot extract meaningful embeddings…",{"value":116,"source":79,"evidence":117},"Manuscript is likely a hoax or gibberish",[118,122],{"id":119,"by":120,"text":121},"44022686","glimshe","I strongly believe the manuscript is undecipherable in the sense thats it's all gibberish. I can't prove it, but at this point I think it's more likely than not to be hoax.",{"id":123,"by":124,"text":125},"44028862","mach5","theres no need to do any of this, its fake, its a forgery",[127,134,142],{"value":128,"source":79,"evidence":129},"Try UMAP or t-SNE instead of PCA",[130],{"id":131,"by":132,"text":133},"44022597","tetris11","UMAP or TSNE would be nice, even if PCA already shows nice separation. Reference mapping each cluster to all the others would be a nice way to indicate that there's no…",{"value":135,"source":79,"evidence":136},"Run a control test on known or meaningless text",[137,141],{"id":138,"by":139,"text":140},"44023668","thih9","(I know nothing about NLP) Does it make sense to check the process with a control group? E.g. if we ask a human to write something that resembles a language but isn’t, then…",{"id":105,"by":106,"text":107},{"value":143,"source":79,"evidence":144},"Try newer embedding models or keep suffixes",[145,146],{"id":97,"by":98,"text":99},{"id":147,"by":148,"text":149},"44026068","gthompson512","Sorry if I missed it, but what about keeping the suffixes and trying to do some finetuning on the source then clustering sentences or at least pages which given the media should…",[151,158,165],{"value":152,"source":79,"evidence":153},"Was the write-up link broken",[154],{"id":155,"by":156,"text":157},"44023944","GTP","The link to the write-up seems broken, can you write the correct one?",{"value":159,"source":79,"evidence":160},"Was the text processed left to right when stripping suffixes",[161],{"id":162,"by":163,"text":164},"44022560","andyjohnson0","This looks very interesting - nice work! I have no background in NLP or linguistics, but I do have a question about this: > I stripped a set of recurring suffix-like endings from…",{"value":166,"source":79,"evidence":167},"Has the manuscript already been deciphered",[168],{"id":169,"by":170,"text":171},"44023233","ablanton","Wasn't it already deciphered, though? https:\u002F\u002Fwww.researchgate.net\u002Fpublication\u002F368991190_The_Voynich_Manuscript_Decoded",[173,180],{"value":174,"source":79,"evidence":175},"Prior Voynich analyses by Zattera, Roe, Feaster and Coons",[176],{"id":177,"by":178,"text":179},"44022868","veqq","The best work on Voynich has been done by Emma Smith, Coons and Patrick Feaster, about loops and QOKEDAR and CHOLDAIIN cycles. Here's a good presentation:…",{"value":181,"source":79,"evidence":182},"voynich.nu transcriptions and voynich.ninja forum",[183,184],{"id":177,"by":178,"text":179},{"id":185,"by":186,"text":187},"44029822","andrewla","Although I skimmed the methodology out of curiosity, what really drew my eye was the transcription in the repository of the manuscript. This led me down a rabbit hole leading…",{"comments":189,"answered":190,"key":191},14,7,[192,195,196],{"id":193,"text":194},"44023585","Thanks for pointing those out — I hadn’t seen PaCMAP or LocalMAP before, but that definitely looks like the kind of structure-preserving approach that would fit this data better…",{"id":63,"text":64},{"id":197,"text":198},"44022618","Great points — thank you. PCA gave me surprisingly clean separation early on, so I stuck with it for the initial run. But you’re right — throwing UMAP or t-SNE at it would…",{"commenters":200,"max_depth":201,"maker_replies":189,"first_comment_mins":35,"lifespan_hours":202},81,9,240,[204,208,210,213,216,219,221,225,227,228,231,233,237,240,244,247,248,249,252,255,257,259,262,264,266,271,274,277,279,281,284,287,290,292,294,297,300,302,304,307,310,312,316,319,321,325,329,333,336,339,342,344,347,351,353,357,359,362,364,367,370,372,375,380,383,385,387,390,392,394,396,398,400,404,407,410,412,414,418,420,423,426,428,430,432,436,439,441,444,448,449,452,455,457,459,462,465,468,471,474,477,479,482,485,488,490,492,494,497,500,503,505,507,509,512,514,516,519,523,526,529,531,533,535,537,541,544,546],[205,30,39,206,207,30],"44023494","project\u002Fneutral\u002Fexecution","patcon",[193,30,30,209,11,34],"maker\u002Fpositive\u002Fexecution",[211,30,39,206,212,30],"44023946","staticautomatic",[214,30,30,215,207,30],"44024851","commenter\u002Fneutral\u002Fnone",[217,30,39,206,218,30],"44026285","loxias",[220,30,39,206,207,30],"44031487",[222,30,39,223,224,30],"44028466","project\u002Fneutral\u002Fidea","khafra",[97,34,34,226,98,30],"project\u002Fcritical\u002Fexecution",[63,34,30,206,11,34],[229,34,30,215,230,30],"44070184","Ey7NFZ3P0nzAe",[138,232,39,223,139,30],2,[234,232,34,235,236,30],"44025019","project\u002Fcritical\u002Fidea","awinter-py",[238,232,39,223,239,30],"44026162","flir",[241,39,30,242,243,30],"44029730","other\u002Fneutral\u002Fnone","cedws",[245,39,30,242,246,30],"44033819","IAmBroom",[131,18,39,206,132,30],[197,18,30,209,11,34],[250,18,39,206,251,30],"44022643","jszymborski",[253,18,30,215,254,30],"44023148","tomrod",[256,18,30,215,98,30],"44023791",[258,18,30,215,251,30],"44024267",[260,18,30,215,261,30],"44023034","lukeinator42",[263,18,30,215,132,30],"44023606",[105,32,39,265,106,30],"project\u002Fmixed\u002Fidea",[267,268,39,269,270,30],"44022579",6,"project\u002Fneutral\u002Fdocs","Avicebron",[272,268,30,273,11,34],"44022606","maker\u002Fneutral\u002Fdocs",[275,190,30,242,276,30],"44022679","us-merul",[278,190,30,242,178,30],"44023053",[280,190,30,215,276,30],"44023095",[282,190,30,242,283,30],"44025108","GolfPopper",[285,190,30,242,286,30],"44025739","renhanxue",[288,190,30,242,289,30],"44025814","emmelaich",[291,190,30,242,286,30],"44025853",[293,190,30,242,289,30],"44026130",[295,190,30,242,296,30],"44026049","philistine",[298,190,30,242,299,30],"44026859","riffraff",[301,190,30,242,296,30],"44030015",[303,190,30,242,299,30],"44059262",[305,190,30,242,306,30],"44027311","cronopios",[308,190,30,242,309,30],"44028995","poulpy123",[311,190,30,242,106,30],"44029056",[313,314,39,223,315,30],"44024155",8,"quantadev",[317,201,39,223,318,30],"44024081","codesnik",[320,201,30,242,178,30],"44024467",[322,323,30,242,324,30],"44029571",10,"empath75",[326,327,30,242,328,30],"44031094",11,"bunderbunder",[330,331,30,242,332,30],"44030013",12,"gwillen",[334,331,30,242,335,30],"44030526","tough",[337,331,30,338,11,34],"44030730","maker\u002Fneutral\u002Fnone",[82,340,32,341,83,30],13,"project\u002Fpositive\u002Fexecution",[56,340,30,343,11,34],"maker\u002Fpositive\u002Fnone",[345,189,39,223,346,30],"44023658","user32489318",[348,349,39,223,350,30],"44026405",15,"frozenseven",[352,349,30,343,11,34],"44037327",[354,355,30,242,356,30],"44023644",16,"ck2",[358,355,30,242,270,30],"44023736",[360,33,30,242,361,30],"44025734","bdbenton5255",[363,33,30,242,11,34],"44025757",[365,33,30,242,366,30],"44025869","PaulDavisThe1st",[368,33,30,242,369,30],"44027844","psychoslave",[371,33,30,242,309,30],"44029009",[185,373,32,374,186,30],18,"project\u002Fpositive\u002Fdocs",[376,377,30,378,379,30],"44024059",19,"other\u002Fneutral\u002Fidea","marcodiego",[381,377,30,382,11,34],"44024087","maker\u002Fneutral\u002Fidea",[384,377,30,378,379,30],"44024101",[386,377,30,242,315,30],"44024212",[388,377,30,242,389,30],"44025304","edoceo",[391,377,30,242,315,30],"44025436",[393,377,30,242,335,30],"44029671",[395,377,30,242,315,30],"44030851",[397,377,30,242,335,30],"44031237",[399,377,30,242,315,30],"44031493",[401,377,30,402,403,30],"44024589","commenter\u002Fneutral\u002Fidea","munchler",[405,377,30,402,406,30],"44027438","raverbashing",[408,377,30,402,409,30],"44027312","mellow_observer",[147,411,39,223,148,30],20,[413,411,30,338,11,34],"44026220",[415,416,30,242,417,30],"44025086",21,"bpiroman",[419,416,30,242,417,30],"44025101",[421,416,30,242,422,30],"44025876","Nursie",[89,424,32,425,90,30],22,"project\u002Fpositive\u002Fidea",[112,427,34,235,113,30],23,[155,429,39,269,156,30],24,[431,429,30,273,11,34],"44024006",[433,434,30,242,435,30],"44023866",25,"rossant",[437,434,30,242,438,30],"44024387","adzm",[123,440,34,235,124,30],26,[442,440,30,378,443,30],"44029390","Tade0",[445,446,30,242,447,30],"44031501",27,"AStonesThrow",[119,35,34,235,120,30],[450,35,32,425,451,30],"44022805","lolinder",[453,35,30,215,454,30],"44023017","andoando",[456,35,30,215,389,30],"44025322",[458,35,30,215,178,30],"44023192",[460,35,30,402,461,30],"44023219","InsideOutSanta",[463,35,30,242,464,30],"44023256","neom",[466,35,30,242,467,30],"44023349","CamperBob2",[469,35,30,242,470,30],"44024910","tonymillion",[472,35,30,242,473,30],"44025743","buildsjets",[475,35,30,402,476,30],"44023309","cubefox",[478,35,30,215,461,30],"44023385",[480,35,30,402,481,30],"44024041","vehemenz",[483,35,30,402,484,30],"44025908","userbinator",[486,35,30,242,487,30],"44023343","ahmedfromtunis",[489,35,30,402,481,30],"44024027",[491,35,30,402,451,30],"44024184",[493,35,30,242,335,30],"44029696",[495,35,30,378,496,30],"44024367","Loughla",[498,35,30,402,499,30],"44024433","int_19h",[501,35,30,378,502,30],"44023408","himinlomax",[504,35,30,378,309,30],"44029152",[162,506,39,269,163,30],29,[508,506,30,382,11,34],"44022576",[510,506,30,242,511,30],"44030312","fader",[177,513,30,378,178,30],30,[515,513,30,338,11,34],"44022901",[517,513,30,242,518,30],"44023173","akomtu",[520,521,32,425,522,30],"44022550",31,"nine_k",[524,521,30,525,11,34],"44022572","maker\u002Fpositive\u002Fidea",[527,521,30,378,528,30],"44022652","gchamonlive",[530,521,30,215,461,30],"44023241",[532,521,30,215,522,30],"44023535",[534,521,30,215,461,30],"44023622",[169,536,30,242,170,30],32,[538,536,30,539,540,30],"44023426","other\u002Fcritical\u002Fnone","Reubend",[542,536,30,242,543,30],"44024136","krick",[545,536,30,242,409,30],"44027432",[547,536,30,242,548,30],"44024579","cookiengineer",[],[551,556,560,565,569,573],{"id":552,"name":553,"points":554,"category":555},"44042371","90s.dev",352,"dev_tools",{"id":557,"name":558,"points":559,"category":555},"44042490","Jwno",297,{"id":561,"name":562,"points":563,"category":564},"44009321","KVSplit",272,"ai",{"id":566,"name":567,"points":568,"category":555},"44006231","Erlang-RED",243,{"id":570,"name":571,"points":572,"category":555},"44023144","Vaev",233,{"id":574,"name":575,"points":576,"category":555},"44020734","Hardtime.nvim",214,false,"v2",{"id":580,"name":5,"line":6,"ai":577,"cat":20,"year":13,"points":14,"comments":15,"rec":29,"s":30,"top":18,"kw":581,"maker":11,"domain":10,"tier":19,"attempts":30,"strip":582,"url":9,"mix":583,"praise":78,"objection":94,"sub":21},44022353,[24,25,26,27,28],"3311330300000305300500030301300000305000",[32,33,34,32],1794036675065]