[{"data":1,"prerenderedAt":136},["ShallowReactive",2],{"project-10844215":3},{"id":4,"name":5,"line":6,"line_source":7,"title":8,"url":9,"domain":5,"maker":10,"posted_at":11,"year":12,"points":13,"comments":14,"rank_in_year":15,"shown_in_year":16,"top":17,"tier":18,"category":19,"subcategory":20,"format":21,"keywords":22,"reception":28,"sentiment":29,"stance_mix":30,"scored_comments":31,"aspects":33,"strip":35,"post_text":29,"what":37,"said":59,"maker_in_thread":85,"shape":90,"grid":93,"attempts":99,"same_week":100,"maker_shows":128,"maker_hidden":129,"reader_version":130,"card":131},"10844215","otoro.net","Neural Net Writes Fake Kanji in Your Web Browser, Stroke-By-Stroke","title","Show HN: Neural Net Writes Fake Kanji in Your Web Browser, Stroke-By-Stroke","http:\u002F\u002Fotoro.net\u002Fkanji\u002F","hardmaru",1452012443,2016,16,3,1342,2264,60,"shown","ai","generative model demo","web_app",[23,24,25,26,27],"neural network","kanji","lstm","stroke generation","fake characters","quiet",null,{"positive":31,"neutral":32,"mixed":32,"critical":32},1,0,{"idea":34},{"positive":31,"critical":32},[36,32,32],5,{"problem":38,"audience":29,"pricing":29,"motivation":29,"time_to_build":29,"stack":44,"openness":58,"stage":58,"team_size":58},{"value":39,"source":40,"evidence":41},"Generates new fake kanji characters drawn stroke by stroke","inferred",[42],{"id":43},"post",[45,52],{"value":46,"source":47,"evidence":48},"LSTM with MDN","maker",[49],{"id":50,"by":10,"text":51},"10849617","It's very interesting you bring this up, Canjie (倉頡) was the first Chinese input method I learned, and I didn't know it was designed to run within disk+memory limitations. To…",{"value":53,"source":47,"evidence":54},"TensorFlow",[55],{"id":56,"by":10,"text":57},"10847787","A blog post on how this algorithm works: http:\u002F\u002Fblog.otoro.net\u002F2015\u002F12\u002F28\u002Frecurrent-net-dreams-up-fake-chinese-characters-in-vector-format-with-tensorflow\u002F","unknown",{"praise":60,"objections":73,"requests":74,"questions":75,"alternatives":80},[61,69],{"value":62,"source":63,"evidence":64},"Interesting idea, relates to Cangjie input method design","commenter",[65],{"id":66,"by":67,"text":68},"10849093","tdeck","This is really interesting, and it reminds me of the Cangjie input method for Chinese characters https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FCangjie_input_method. It was designed to run within…",{"value":70,"source":47,"evidence":71},"Cangjie is still a strong input method for professionals",[72],{"id":50,"by":10,"text":51},[],[],[76],{"value":77,"source":47,"evidence":78},"Could sparse-encoder neural nets produce a more efficient Cangjie-like input method",[79],{"id":50,"by":10,"text":51},[81],{"value":82,"source":63,"evidence":83},"Cangjie input method",[84],{"id":66,"by":67,"text":68},{"comments":86,"answered":32,"key":87},2,[88,89],{"id":50,"text":51},{"id":56,"text":57},{"commenters":86,"max_depth":86,"maker_replies":86,"first_comment_mins":91,"lifespan_hours":92},530,17,[94,96,97],[66,32,36,95,67,32],"project\u002Fpositive\u002Fidea",[50,32,32,95,10,31],[56,31,32,98,10,31],"maker\u002Fneutral\u002Fdocs",[],[101,106,110,115,120,124],{"id":102,"name":103,"points":104,"category":105},"10824064","GPU text rendering with vector textures",334,"languages",{"id":107,"name":108,"points":109,"category":105},"10849460","Nodal",288,{"id":111,"name":112,"points":113,"category":114},"10830587","Saola",151,"dev_tools",{"id":116,"name":117,"points":118,"category":119},"10830799","friends",96,"productivity",{"id":121,"name":122,"points":123,"category":119},"10850811","GitBook Editor Beta",92,{"id":125,"name":126,"points":127,"category":105},"10832207","Searchkit",74,6,false,"v2",{"id":132,"name":5,"line":6,"ai":129,"cat":19,"year":12,"points":13,"comments":14,"rec":28,"s":29,"top":17,"kw":133,"maker":10,"domain":5,"tier":18,"attempts":32,"strip":134,"url":9,"mix":135,"praise":62,"objection":-1,"sub":20},10844215,[23,24,25,26,27],"500",[31,32,32,32],1793990935677]