[{"data":1,"prerenderedAt":859},["ShallowReactive",2],{"project-49026810":3},{"id":4,"name":5,"line":6,"line_source":7,"title":8,"url":9,"domain":9,"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":35,"aspects":36,"strip":51,"post_text":53,"what":54,"said":74,"maker_in_thread":191,"shape":201,"grid":205,"attempts":819,"same_week":820,"maker_shows":17,"maker_hidden":850,"reader_version":851,"card":852},"49026810","Echo","Fable-level results at 1\u002F3 the cost using open-weight models","title","Show HN: Echo – Fable-level results at 1\u002F3 the cost using open-weight models",null,"adam_rida",1784834761,2026,484,228,55,4366,2,"legend","ai","model routing and ensembling","web_app",[23,24,25,26,27],"open-weight models","model routing","ensemble","llm cost","openai compatible api","mixed",-0.383,{"positive":31,"neutral":32,"mixed":33,"critical":34},10,35,6,47,98,{"design":37,"privacy":39,"execution":41,"pricing":43,"docs":45,"performance":46,"idea":48},{"positive":38,"critical":31},0,{"positive":38,"critical":40},1,{"positive":38,"critical":42},4,{"positive":38,"critical":44},7,{"positive":40,"critical":17},{"positive":38,"critical":47},3,{"positive":49,"critical":50},8,14,[40,40,40,40,40,40,38,38,38,38,38,38,40,40,40,17,42,17,40,52,17,40,47,40,52,17,40,40,40,40,17,47,38,40,40,17,40,40,47,42],5,"I’ve been building Echo (https:\u002F\u002Fecho.tracerml.ai\u002F), an experiment in making one AI system out of a pool of open-weight models rather than choosing a single model and using it for every task.\n\nIt started with a simple experiment. I took a group of models, including GLM-5.2, Kimi K2.7 and others, and ran them on the same evaluations. Then I measured what would happen if, for each problem, you somehow knew in advance which models would be useful and how their outputs should be combined.\n\nThat hypothetical system performed substantially better than any individual model in the pool. Of course, it ",{"problem":55,"audience":9,"pricing":61,"motivation":67,"time_to_build":9,"stack":71,"openness":72,"stage":73,"team_size":72},{"value":56,"source":57,"evidence":58},"High cost of frontier models for each request","maker",[59],{"id":60},"post",{"value":62,"source":57,"evidence":63},"Free credits of $10 per account, no credit card required",[64],{"id":65,"by":10,"text":66},"49030253","thanks to everyone for taking the time to try Echo and share feedback, this is precisely why i wanted to launch early. i am going to try to address a couple of topics that came…",{"value":68,"source":57,"evidence":69},"Recover the advantage of an ensemble of models without knowing outcomes in advance",[70],{"id":60},[],"unknown","beta",{"praise":75,"objections":88,"requests":139,"questions":151,"alternatives":174},[76],{"value":77,"source":78,"evidence":79},"Model complementarity and ensembling idea is interesting","commenter",[80,84],{"id":81,"by":82,"text":83},"49036741","smokeeaasd","I think one of the most interesting takeaways is that model selection may matter more than model size. The industry has largely focused on building larger models, but your…",{"id":85,"by":86,"text":87},"49027374","bnjemian","I don’t find the recent spate of blog posts and systems delegating and combining LLMs to get better performance particularly interesting. Especially given that anyone who’s taken…",[89,104,119,131],{"value":90,"source":78,"evidence":91},"Claimed cost savings are unclear versus subscriptions and caching",[92,96,100],{"id":93,"by":94,"text":95},"49027790","jacobgold","> Fable-level results at 1\u002F3 the cost using open-weight models But we get ~$2500\u002Fmo worth of Fable credits for $200\u002Fmo on Anthropic pan? I'm still confused why people (who don't…",{"id":97,"by":98,"text":99},"49029078","subygan","This does not really work well, if you don't know the complexity of the problem ahead of time and ensure all future conversations go to the same model. Else, you break the cache…",{"id":101,"by":102,"text":103},"49028780","purplecats","but u wouldnt get caching savings",{"value":105,"source":78,"evidence":106},"Sign-up flow and login issues",[107,111,115],{"id":108,"by":109,"text":110},"49029638","qainsights","Can't sign up: `too many authentication attempts`",{"id":112,"by":113,"text":114},"49026989","kamranjon","No benchmarks, no info on which models are used, ai generated video, just a signup page with nothing else. Anyhow, this kinda reminds me of that quote about architecture: \"We…",{"id":116,"by":117,"text":118},"49027029","tj800x","No single signin. Privacy policy allows training. No try it first without credit card. It's a good idea, but this looks premature.",{"value":120,"source":78,"evidence":121},"Ensembling is not new",[122,123,127],{"id":85,"by":86,"text":87},{"id":124,"by":125,"text":126},"49028999","fgoose180","enjoying how people are re-discovering ensemble methods.",{"id":128,"by":129,"text":130},"49030008","Art9681","The \"ensemble of models\" is nothing new. It's just that it's not really a moat that can be monetized. In the end you're always paying for something. You can engineer an elaborate…",{"value":132,"source":78,"evidence":133},"Benchmarks thin or hard to verify",[134,135],{"id":112,"by":113,"text":114},{"id":136,"by":137,"text":138},"49027755","ninjahawk1","I’m very confused on what this is, my initial thought was “oh nice, open source router.” I go to the website…and it’s a sign up. I expected a repo. Otherwise how do I use it? As…",[140,144],{"value":141,"source":57,"evidence":142},"Benchmarks for coding and agentic tasks",[143],{"id":65,"by":10,"text":66},{"value":145,"source":78,"evidence":146},"Clearer eval page with more graphs",[147],{"id":148,"by":149,"text":150},"49031597","sudo_cowsay","Hi this is very cool. As a student (read: more dumb than adults), I find the eval page too loaded with words. Maybe more graphs could help?",[152,167],{"value":153,"source":78,"evidence":154},"How does this differ from OpenRouter Fusion",[155,159,163],{"id":156,"by":157,"text":158},"49027336","wizche","how does this differs from OpenRouter fusion?",{"id":160,"by":161,"text":162},"49031540","thatxliner","How is this different from what that feature OpenRouter does?",{"id":164,"by":165,"text":166},"49028256","meander_water","Seems similar to Openrouter Fusion - https:\u002F\u002Fopenrouter.ai\u002Fdocs\u002Fguides\u002Frouting\u002Frouters\u002Ffusion-router",{"value":168,"source":78,"evidence":169},"What signals decide which models participate",[170],{"id":171,"by":172,"text":173},"49029630","islambaraka","What signals does Echo use to decide which models should participate before it has seen their outputs?",[175,180],{"value":176,"source":78,"evidence":177},"OpenRouter Fusion",[178,179],{"id":164,"by":165,"text":166},{"id":156,"by":157,"text":158},{"value":181,"source":78,"evidence":182},"Sakana Fugu",[183,187],{"id":184,"by":185,"text":186},"49029508","alightsoul","There's so many people reimplementing Sakana fugu from its two ICLR papers but no open source version of it.",{"id":188,"by":189,"text":190},"49027514","ototot","Is this yet another Sakana Fugu \u002F OpenRouter Fusion?",{"comments":192,"answered":47,"key":193},11,[194,197,200],{"id":195,"text":196},"49035601","removing it now",{"id":198,"text":199},"49039870","good catch, on it too now!",{"id":65,"text":66},{"commenters":202,"max_depth":44,"maker_replies":192,"first_comment_mins":203,"lifespan_hours":204},153,16,184,[206,210,214,217,220,223,226,229,231,235,236,238,241,244,247,250,253,255,257,261,264,267,270,272,275,278,280,284,288,291,294,298,301,305,309,312,315,317,319,322,325,328,331,334,337,340,343,345,348,350,353,356,359,361,364,367,370,372,375,378,381,384,386,389,392,395,397,400,403,405,407,410,413,417,419,423,426,429,431,435,438,440,443,446,449,452,455,458,462,464,467,470,473,476,479,482,485,488,491,493,496,499,502,504,507,510,513,516,519,522,525,527,530,533,535,536,539,542,545,546,548,550,553,557,560,563,565,569,571,573,577,580,583,586,589,592,595,598,602,604,607,611,614,616,618,622,624,627,629,631,633,637,641,643,647,650,652,654,656,660,662,664,668,672,676,680,682,686,690,694,697,699,702,705,708,711,714,717,720,723,726,728,731,733,737,739,741,744,748,750,752,755,757,760,763,767,769,773,774,777,781,784,786,790,792,794,796,799,801,804,807,810,814,817],[207,38,40,208,209,38],"49032948","project\u002Fcritical\u002Fdesign","fmx",[211,38,38,212,213,38],"49033577","commenter\u002Fneutral\u002Fnone","pizzao",[215,38,40,208,216,38],"49033657","lukan",[218,38,40,208,219,38],"49033788","boesboes",[221,38,40,208,222,38],"49033912","idonotknowwhy",[224,38,38,212,225,38],"49045894","7bit",[227,38,40,208,228,38],"49034721","ciefa",[195,38,38,230,10,40],"maker\u002Fneutral\u002Fdesign",[232,38,47,233,234,38],"49035870","maker\u002Fneutral\u002Fprivacy","rufasterisco",[198,38,38,230,10,40],[65,40,38,237,10,40],"project\u002Fneutral\u002Fnone",[239,40,40,208,240,38],"49031151","user_7832",[242,40,40,208,243,38],"49031744","troupo",[245,40,38,212,246,38],"49033993","RugnirViking",[248,40,38,212,249,38],"49035017","bitexploder",[251,40,38,212,252,38],"49034170","jorisw",[254,40,38,212,240,38],"49034479",[256,40,38,230,10,40],"49039896",[258,40,40,259,260,38],"49032768","project\u002Fcritical\u002Fexecution","rahulroy",[262,40,40,259,263,38],"49032926","peterlopen",[265,40,47,266,263,38],"49035921","project\u002Fneutral\u002Fexecution",[268,40,47,269,260,38],"49054893","project\u002Fneutral\u002Fdesign",[271,40,47,237,113,38],"49034249",[273,40,38,212,274,38],"49041644","schwetim",[276,40,40,208,277,38],"49034596","afzalive",[279,40,40,208,228,38],"49034736",[281,17,40,282,283,38],"49029061","project\u002Fcritical\u002Fpricing","cheema33",[285,17,38,286,287,38],"49029438","commenter\u002Fneutral\u002Fpricing","killingtime74",[289,17,38,212,290,38],"49029595","nickthegreek",[292,17,38,286,293,38],"49029878","rodrodrod",[295,17,38,296,297,38],"49032253","commenter\u002Fneutral\u002Fprivacy","lnrd",[299,17,38,286,300,38],"49029993","monk_grilla",[302,17,38,303,304,38],"49030687","other\u002Fneutral\u002Fpricing","byzantinegene",[306,17,38,307,308,38],"49030800","other\u002Fneutral\u002Fnone","bensyverson",[310,17,38,307,311,38],"49031397","nl",[313,17,38,307,314,38],"49033138","lelanthran",[316,17,38,307,311,38],"49033325",[318,17,38,307,300,38],"49078173",[320,17,38,303,321,38],"49030357","stilesja",[323,17,38,307,324,38],"49030852","good8675309",[326,17,38,307,327,38],"49033684","sunaookami",[329,17,38,307,330,38],"49031298","jambalaya8",[332,17,38,307,333,38],"49031343","DougN7",[335,17,38,307,336,38],"49031742","Foobar8568",[338,17,38,303,339,38],"49032378","stkdump",[341,17,38,307,342,38],"49032834","cube00",[344,17,38,307,330,38],"49039232",[346,17,38,307,347,38],"49034500","mattjoyce",[349,17,38,307,330,38],"49116595",[351,17,38,303,352,38],"49031315","hmottestad",[354,17,38,307,355,38],"49031732","sheepscreek",[357,17,38,307,358,38],"49033695","Bombthecat",[360,17,38,303,304,38],"49030680",[362,17,38,303,363,38],"49030949","suby",[365,17,38,303,366,38],"49030964","orsorna",[368,17,38,303,369,38],"49031090","victorbjorklund",[371,17,38,303,311,38],"49031415",[373,17,38,303,374,38],"49032197","kelnos",[376,17,38,307,377,38],"49031129","aroman",[379,17,38,307,380,38],"49031161","somenameforme",[382,17,38,212,383,38],"49032026","davedx",[385,17,38,303,297,38],"49032276",[387,17,38,303,388,38],"49031240","neonstatic",[390,17,38,303,391,38],"49031317","MertsA",[393,17,38,303,394,38],"49031372","agar",[396,17,38,303,311,38],"49031444",[398,17,38,303,399,38],"49031828","dd8601fn",[401,17,38,303,402,38],"49031607","hahahaa",[116,47,40,404,117,38],"project\u002Fcritical\u002Fprivacy",[406,47,38,233,10,40],"49027297",[408,47,40,282,409,38],"49027731","cynicalsecurity",[411,47,38,307,412,38],"49027986","seizethecheese",[414,47,40,415,416,38],"49030671","project\u002Fcritical\u002Fnone","DatCodeMania",[112,42,40,418,113,38],"project\u002Fcritical\u002Fdocs",[420,42,47,421,422,38],"49027017","project\u002Fmixed\u002Fdocs","guessmyname",[424,42,40,425,412,38],"49027669","project\u002Fcritical\u002Fperformance",[427,42,38,307,428,38],"49029132","yorwba",[430,42,38,307,412,38],"49029159",[432,42,40,433,434,38],"49027212","project\u002Fcritical\u002Fidea","j45",[436,42,38,437,10,40],"49027324","maker\u002Fneutral\u002Fdocs",[439,42,47,237,412,38],"49029028",[441,42,47,442,300,38],"49030006","project\u002Fneutral\u002Fdocs",[444,42,47,421,445,38],"49030462","dannyw",[447,42,47,442,448,38],"49033655","lionkor",[450,42,40,433,451,38],"49027436","codekansas",[453,42,52,454,412,38],"49027963","project\u002Fpositive\u002Fidea",[456,42,47,237,457,38],"49028525","vunderba",[459,42,47,460,461,38],"49029938","project\u002Fneutral\u002Fidea","throw10920",[463,42,38,307,457,38],"49029977",[465,42,38,307,466,38],"49028051","dimitrios1",[468,42,38,307,469,38],"49028230","WD-42",[471,42,40,415,472,38],"49031282","XCSme",[474,52,40,433,475,38],"49027950","dluan",[477,52,52,454,478,38],"49028275","glasss",[480,52,52,454,481,38],"49028968","tiffanyh",[483,52,40,433,484,38],"49029064","Avicebron",[486,52,47,460,487,38],"49029072","onlyrealcuzzo",[489,52,38,212,490,38],"49029250","Terretta",[492,52,38,307,240,38],"49031109",[494,52,38,307,495,38],"49031185","rablackburn",[497,52,40,433,498,38],"49030572","ai-x",[500,52,38,307,501,38],"49031218","joe_the_user",[503,52,38,307,487,38],"49034430",[505,52,38,307,506,38],"49028283","alizaki",[508,52,38,307,509,38],"49028299","cdelsolar",[511,52,52,454,512,38],"49028446","ignoramous",[514,52,47,460,515,38],"49028700","fmajid",[517,52,40,433,518,38],"49029480","johnvanommen",[520,52,52,454,521,38],"49030027","ljlolel",[523,52,38,524,512,38],"49031210","other\u002Fneutral\u002Fdocs",[526,52,38,307,521,38],"49070506",[528,52,40,433,529,38],"49030285","moralestapia",[531,52,40,433,532,38],"49030364","gchamonlive",[534,52,47,460,402,38],"49031632",[97,33,40,425,98,38],[537,33,47,538,402,38],"49031616","project\u002Fneutral\u002Fperformance",[540,33,38,307,541,38],"49032758","lbriner",[543,44,40,415,544,38],"49028294","slashdave",[164,49,47,460,165,38],[547,49,47,460,412,38],"49028740",[81,549,52,454,82,38],9,[551,549,38,552,10,40],"49037342","maker\u002Fneutral\u002Fidea",[554,31,47,555,556,38],"49031934","project\u002Fmixed\u002Fidea","Arshad-Talpur",[558,192,40,415,559,38],"49028551","jmspring",[561,192,38,212,562,38],"49028746","dijit",[564,192,38,212,559,38],"49030993",[566,567,40,433,568,38],"49030429",12,"springtimesun",[570,567,38,307,472,38],"49031287",[572,567,38,307,568,38],"49121976",[574,575,52,454,576,38],"49027140",13,"Alifatisk",[578,575,38,579,412,38],"49027623","other\u002Fneutral\u002Fidea",[581,575,38,307,582,38],"49028069","hmokiguess",[584,575,38,307,585,38],"49028493","NetOpWibby",[587,575,38,579,588,38],"49029101","tintor",[590,50,47,460,591,38],"49027097","jmaw",[593,594,38,307,521,38],"49030022",15,[596,203,40,415,597,38],"49031722","jdthedisciple",[599,203,52,600,601,38],"49031740","project\u002Fpositive\u002Fdocs","elnatro",[603,203,40,425,597,38],"49032054",[605,203,40,282,606,38],"49036362","saberience",[608,609,38,579,610,38],"49027554",17,"yonatan8070",[612,609,38,579,613,38],"49027581","janalsncm",[615,609,38,579,216,38],"49028215",[617,609,38,579,613,38],"49028774",[619,620,47,460,621,38],"49030301",18,"blobbers",[623,620,38,552,10,40],"49030361",[625,620,38,626,621,38],"49030606","commenter\u002Fneutral\u002Fidea",[628,620,38,552,10,40],"49044659",[630,620,40,433,314,38],"49033187",[148,632,47,421,149,38],19,[634,635,40,208,636,38],"49037306",20,"NoNameAditya",[638,639,40,433,640,38],"49032258",21,"stevefan1999",[184,642,38,307,185,38],22,[644,645,38,307,646,38],"49036560",23,"motbus3",[648,649,38,303,613,38],"49027730",24,[651,649,38,303,412,38],"49028873",[653,649,38,303,613,38],"49028996",[655,649,38,303,412,38],"49029288",[657,658,40,415,659,38],"49044423",25,"boveyking",[160,661,47,460,161,38],26,[663,661,47,460,402,38],"49031647",[665,666,47,460,667,38],"49031726",27,"haris599",[669,670,40,433,671,38],"49030297",28,"zhonglin",[673,674,47,555,675,38],"49036294",29,"datadrivenangel",[677,678,47,679,434,38],"49027230",30,"project\u002Fneutral\u002Fpricing",[171,681,47,460,172,38],31,[683,684,38,307,685,38],"49028334",32,"indiantinker",[687,688,40,282,689,38],"49032173",33,"SubiculumCode",[691,692,38,307,693,38],"49027682",34,"bbstats",[695,32,52,454,696,38],"49031951","alex-moon",[93,698,38,303,94,38],36,[700,698,38,303,701,38],"49027822","teruakohatu",[703,698,38,303,704,38],"49027932","sscaryterry",[706,698,38,303,707,38],"49028074","recursivegirth",[709,698,38,303,710,38],"49028339","trollbridge",[712,698,38,303,713,38],"49030276","anonzzzies",[715,698,38,303,716,38],"49027876","dberg",[718,698,38,303,719,38],"49027946","combyn8tor",[721,698,38,303,722,38],"49027970","dbbk",[724,698,40,282,725,38],"49027983","gruez",[727,698,38,212,719,38],"49030448",[729,698,38,303,730,38],"49028237","eikenberry",[128,732,40,433,129,38],37,[734,735,47,460,736,38],"49027074",38,"fneddy",[738,735,47,460,412,38],"49028908",[108,740,40,259,109,38],39,[742,740,38,743,10,40],"49029792","maker\u002Fneutral\u002Fnone",[745,746,40,259,747,38],"49029626",40,"raver1975",[749,746,38,743,10,40],"49029791",[156,751,47,460,157,38],41,[753,751,47,460,754,38],"49027492","cawingcrow",[85,756,47,555,86,38],42,[758,756,38,212,759,38],"49027608","abernard1",[761,762,47,237,451,38],"49027299",43,[764,765,40,282,766,38],"49032356",44,"zuzululu",[136,768,40,418,137,38],45,[770,771,47,460,772,38],"49054550",46,"senderista",[124,34,40,433,125,38],[775,776,47,237,330,38],"49027302",48,[778,779,38,307,780,38],"49029762",49,"cantalopes",[782,779,38,307,783,38],"49029801","meowface",[785,779,38,212,713,38],"49030265",[787,788,40,282,789,38],"49027727",50,"maxdo",[791,788,38,212,412,38],"49028895",[101,793,47,679,102,38],51,[795,793,38,286,412,38],"49028859",[797,793,38,286,798,38],"49030249","retinaros",[188,800,47,460,189,38],52,[802,800,47,460,803,38],"49027888","kachnuv_ocasek",[805,800,47,460,806,38],"49035284","code0igx",[808,809,47,743,582,38],"49028008",53,[811,809,52,812,813,38],"49030108","maker\u002Fpositive\u002Fnone","zachdotai",[815,809,47,743,816,38],"49030318","dang",[818,809,38,212,582,38],"49035463",[],[821,825,830,835,840,845],{"id":822,"name":823,"points":824,"category":19},"49098510","TurboFieldfare",919,{"id":826,"name":827,"points":828,"category":829},"49021270","blackhole.plav.in",477,"learning_science",{"id":831,"name":832,"points":833,"category":834},"49090607","HNewhere",433,"dev_tools",{"id":836,"name":837,"points":838,"category":839},"49100043","CheapFoodMap",287,"consumer_social",{"id":841,"name":842,"points":843,"category":844},"49057666","Reverse Minesweeper",267,"games",{"id":846,"name":847,"points":848,"category":849},"49020545","Global Oil Network",244,"data",false,"v2",{"id":853,"name":5,"line":6,"ai":850,"cat":19,"year":12,"points":13,"comments":14,"rec":28,"s":854,"top":17,"kw":855,"maker":10,"domain":856,"tier":18,"attempts":38,"strip":857,"url":-1,"mix":858,"praise":77,"objection":90,"sub":20},49026810,-0.38,[23,24,25,26,27],"","1111110000001112421521315211112301121134",[31,32,33,34],1794036705872]