[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"summary-274d310f289e54db-yin-yang-llm-pipeline-cuts-noise-in-code-scanning-summary":3,"summaries-facets-categories":112,"summary-related-274d310f289e54db-yin-yang-llm-pipeline-cuts-noise-in-code-scanning-summary":3697},{"id":4,"title":5,"ai":6,"body":13,"categories":76,"created_at":78,"date_modified":78,"description":69,"extension":79,"faq":78,"featured":80,"kicker_label":78,"meta":81,"navigation":93,"path":94,"published_at":95,"question":78,"scraped_at":96,"seo":97,"sitemap":98,"source_id":99,"source_name":100,"source_type":101,"source_url":102,"stem":103,"tags":104,"thumbnail_url":78,"tldr":109,"tweet":78,"unknown_tags":110,"__hash__":111},"summaries\u002Fsummaries\u002F274d310f289e54db-yin-yang-llm-pipeline-cuts-noise-in-code-scanning-summary.md","Yin-Yang LLM Pipeline Cuts Noise in Code Scanning",{"provider":7,"model":8,"input_tokens":9,"output_tokens":10,"processing_time_ms":11,"cost_usd":12},"openrouter","x-ai\u002Fgrok-4.1-fast",5484,1270,11737,0.00122495,{"type":14,"value":15,"toc":68},"minimark",[16,21,34,38,46,50,61,65],[17,18,20],"h2",{"id":19},"designed-distrust-yin-yang-agents-balance-recall-and-precision","Designed Distrust: Yin-Yang Agents Balance Recall and Precision",[22,23,24,25,29,30,33],"p",{},"Treat LLMs as stochastic machines with expected errors, not infallible oracles. Veritas pipeline uses two opposing agents: the ",[26,27,28],"strong",{},"hypothesis agent (Yang)"," maximizes recall by scanning architecture entry points, trust boundaries, and threat models to flag every plausible vulnerability candidate cheaply—prioritizing low cost of refuted hypotheses over missing real issues. The ",[26,31,32],{},"evidence agent (Yin)"," maximizes precision as a skeptical auditor, refuting claims unless backed by cited source-code evidence like sanitizers or validation logic. This tug-of-war decouples generation from judgment, letting hypothesis over-generate (reducing Type 2 false negatives) while evidence strictly verifies (reducing Type 1 false positives), avoiding single-stage scanners' internal conflicts on the precision-recall curve.",[17,35,37],{"id":36},"information-bottleneck-defeats-anchoring-bias","Information Bottleneck Defeats Anchoring Bias",[22,39,40,41,45],{},"Strip reasoning from hypotheses via ",[42,43,44],"code",{},"slim_hypotheses_for_evidence()"," before evidence review, forcing independent verification against raw code context. Deleting the \"why\" prevents anchoring bias, where prior explanations cause overfitting and hallucinated confirmations. Evidence agent must rediscover exploit paths or discard findings, making the system smarter by intentionally \"dumbing down\" the second stage—trading tokens for quality.",[17,47,49],{"id":48},"deterministic-policy-gate-overrides-ai-verdicts","Deterministic Policy Gate Overrides AI Verdicts",[22,51,52,53,56,57,60],{},"AI never decides alone: a mechanical ",[26,54,55],{},"policy gate"," applies checklists post-pipeline. Findings below 0.3 confidence score marked \"Inconclusive\"; critical\u002Fhigh-severity inconclusive ones flagged ",[42,58,59],{},"NEEDS_HUMAN"," instead of dropped. Every pre-scan finding gets confirmed, disproven, or escalated with full audit trail. This handles variation like assembly-line quality control, producing reviewable outputs for developers during manual reviews—not replacing SAST tools like CodeQL, but assisting Security Champions.",[17,62,64],{"id":63},"pipeline-outcomes-process-over-prompting","Pipeline Outcomes: Process Over Prompting",[22,66,67],{},"Statistical process control trumps prompt engineering: accuracy emerges from architectural tension (expansion then contraction), not model size or perfect prompts. Hypothesis runs high-recall\u002Flow-precision; evidence shifts to high-precision. Extra compute yields hallucination-free reports tied to evidence, with next steps in empirical calibration against known vuln\u002Fnon-vuln cases. Veritas (GitHub POC) proves treating LLM outputs as intermediates—with inspection, disposition, escalation—yields trustworthy code review aids.",{"title":69,"searchDepth":70,"depth":70,"links":71},"",2,[72,73,74,75],{"id":19,"depth":70,"text":20},{"id":36,"depth":70,"text":37},{"id":48,"depth":70,"text":49},{"id":63,"depth":70,"text":64},[77],"AI & LLMs",null,"md",false,{"content_references":82,"triage":88},[83],{"type":84,"title":85,"url":86,"context":87},"tool","Veritas","https:\u002F\u002Fgithub.com\u002Fjoshconkel\u002FVeritas-POC","mentioned",{"relevance":89,"novelty":90,"quality":90,"actionability":90,"composite":91,"reasoning":92},5,4,4.35,"Category: AI Automation. The article presents a novel approach to AI code scanning by introducing a dual-agent system that balances recall and precision, addressing a specific pain point for developers looking to improve code reliability. It offers actionable insights on implementing a deterministic policy gate and emphasizes the importance of independent verification, making it relevant and practical for the target audience.",true,"\u002Fsummaries\u002F274d310f289e54db-yin-yang-llm-pipeline-cuts-noise-in-code-scanning-summary","2026-05-03 16:01:01","2026-05-03 17:00:57",{"title":5,"description":69},{"loc":94},"274d310f289e54db","Towards AI","article","https:\u002F\u002Fpub.towardsai.net\u002Fyin-yang-and-the-llm-engineering-reliability-into-ai-code-scanning-35f4f6c222fa?source=rss----98111c9905da---4","summaries\u002F274d310f289e54db-yin-yang-llm-pipeline-cuts-noise-in-code-scanning-summary",[105,106,107,108],"llm","agents","ai-automation","software-engineering","Build reliable AI code scanners by pitting a recall-focused hypothesis agent against a precision-focused evidence agent, stripping reasoning to avoid bias, and enforcing a deterministic policy gate—treating LLMs as stochastic machines, not oracles.",[107,108],"JWz6oOrzw6FIoiG_JJucEu0m9Nr-uNWnazNK1OjyM-I",[113,116,119,121,124,127,129,131,133,135,137,139,142,144,146,148,150,152,154,156,158,160,163,166,168,170,173,175,177,180,182,184,186,188,190,192,194,196,198,200,202,204,206,208,210,212,214,216,218,220,222,224,226,228,230,232,234,236,238,240,242,244,246,248,250,252,254,256,258,260,262,264,266,268,270,272,274,276,278,280,282,284,286,288,290,292,294,296,298,300,302,304,306,308,310,312,314,316,318,320,322,324,326,328,330,332,334,336,338,340,342,344,346,348,350,352,354,356,358,360,362,364,366,368,370,372,374,376,378,380,382,384,386,388,390,392,394,396,398,400,402,404,406,408,410,412,414,416,418,420,422,424,426,428,430,432,434,437,439,441,443,445,447,449,451,453,455,457,459,461,463,465,467,469,471,473,475,477,479,481,483,485,487,489,491,493,495,497,499,501,503,505,507,509,511,513,515,517,519,521,523,525,527,529,531,533,535,537,539,541,543,545,547,549,551,553,555,557,559,561,563,565,567,569,571,573,575,577,579,581,583,585,587,589,591,593,595,597,599,601,603,605,607,609,611,613,615,617,619,621,623,625,627,629,631,633,635,637,639,641,643,645,647,649,651,653,655,657,659,661,663,665,667,669,671,673,675,677,679,681,683,685,687,689,691,693,695,697,699,701,703,705,707,709,711,713,715,717,719,721,723,725,727,729,731,733,735,737,739,741,743,745,747,749,751,753,755,757,759,761,763,765,767,769,771,773,775,777,779,781,783,785,787,789,791,793,795,797,799,801,803,805,807,809,811,813,815,817,819,821,823,825,827,829,831,833,835,837,839,841,843,845,847,849,851,853,855,857,859,861,863,865,867,869,871,873,875,877,879,881,883,885,887,889,891,893,895,897,899,901,903,905,907,909,911,913,915,917,919,921,923,925,927,929,931,933,935,937,939,941,943,945,947,949,951,953,955,957,959,961,963,965,967,969,971,973,975,977,979,981,983,985,987,989,991,993,995,997,999,1001,1003,1005,1007,1009,1011,1013,1015,1017,1019,1021,1023,1025,1027,1029,1031,1033,1035,1037,1039,1041,1043,1045,1047,1049,1051,1053,1055,1057,1059,1061,1063,1065,1067,1069,1071,1073,1075,1077,1079,1081,1083,1085,1087,1089,1091,1093,1095,1097,1099,1101,1103,1105,1107,1109,1111,1113,1115,1117,1119,1121,1123,1125,1127,1129,1131,1133,1135,1137,1139,1141,1143,1145,1147,1149,1151,1153,1155,1157,1159,1161,1163,1165,1167,1169,1171,1173,1175,1177,1179,1181,1183,1185,1187,1189,1191,1193,1195,1197,1199,1201,1203,1205,1207,1209,1211,1213,1215,1217,1219,1221,1223,1225,1227,1229,1231,1233,1235,1237,1239,1241,1243,1245,1247,1249,1251,1253,1255,1257,1259,1261,1263,1265,1267,1269,1271,1273,1275,1277,1279,1281,1283,1285,1287,1289,1291,1293,1295,1297,1299,1301,1303,1305,1307,1309,1311,1313,1315,1317,1319,1321,1323,1325,1327,1329,1331,1333,1335,1337,1339,1341,1343,1345,1347,1349,1351,1353,1355,1357,1359,1361,1363,1365,1367,1369,1371,1373,1375,1377,1379,1381,1383,1385,1387,1389,1391,1393,1395,1397,1399,1401,1403,1405,1407,1409,1411,1413,1415,1417,1419,1421,1423,1425,1427,1429,1431,1433,1435,1437,1439,1441,1443,1445,1447,1449,1451,1453,1455,1457,1459,1461,1463,1465,1467,1469,1471,1473,1475,1477,1479,1481,1483,1485,1487,1489,1491,1493,1495,1497,1499,1501,1503,1505,1507,1509,1511,1513,1515,1517,1519,1521,1523,1525,1527,1529,1531,1533,1535,1537,1539,1541,1543,1545,1547,1549,1551,1553,1555,1557,1559,1561,1563,1565,1567,1569,1571,1573,1575,1577,1579,1581,1583,1585,1587,1589,1591,1593,1595,1597,1599,1601,1603,1605,1607,1609,1611,1613,1615,1617,1619,1621,1623,1625,1627,1629,1631,1633,1635,1637,1639,1641,1643,1645,1647,1649,1651,1653,1655,1657,1659,1661,1663,1665,1667,1669,1671,1673,1675,1677,1679,1681,1683,1685,1687,1689,1691,1693,1695,1697,1699,1701,1703,1705,1707,1709,1711,1713,1715,1717,1719,1721,1723,1725,1727,1729,1731,1733,1735,1737,1739,1741,1743,1745,1747,1749,1751,1753,1755,1757,1759,1761,1763,1765,1767,1769,1771,1773,1775,1777,1779,1781,1783,1785,1787,1789,1791,1793,1795,1797,1799,1801,1803,1805,1807,1809,1811,1813,1815,1817,1819,1821,1823,1825,1827,1829,1831,1833,1835,1837,1839,1841,1843,1845,1847,1849,1851,1853,1855,1857,1859,1861,1863,1865,1867,1869,1871,1873,1875,1877,1879,1881,1883,1885,1887,1889,1891,1893,1895,1897,1899,1901,1903,1905,1907,1909,1911,1913,1915,1917,1919,1921,1923,1925,1927,1929,1931,1933,1935,1937,1939,1941,1943,1945,1947,1949,1951,1953,1955,1957,1959,1961,1963,1965,1967,1969,1971,1973,1975,1977,1979,1981,1983,1985,1987,1989,1991,1993,1995,1997,1999,2001,2003,2005,2007,2009,2011,2013,2015,2017,2019,2021,2023,2025,2027,2029,2031,2033,2035,2037,2039,2041,2043,2045,2047,2049,2051,2053,2055,2057,2059,2061,2063,2065,2067,2069,2071,2073,2075,2077,2079,2081,2083,2085,2087,2089,2091,2093,2095,2097,2099,2101,2103,2105,2107,2109,2111,2113,2115,2117,2119,2121,2123,2125,2127,2129,2131,2133,2135,2137,2139,2141,2143,2145,2147,2149,2151,2153,2155,2157,2159,2161,2163,2165,2167,2169,2171,2173,2175,2177,2179,2181,2183,2185,2187,2189,2191,2193,2195,2197,2199,2201,2203,2205,2207,2209,2211,2213,2215,2217,2219,2221,2223,2225,2227,2229,2231,2233,2235,2237,2239,2241,2243,2245,2247,2249,2251,2253,2255,2257,2259,2261,2263,2265,2267,2269,2271,2273,2275,2277,2279,2281,2283,2285,2287,2289,2291,2293,2295,2297,2299,2301,2303,2305,2307,2309,2311,2313,2315,2317,2319,2321,2323,2325,2327,2329,2331,2333,2335,2337,2339,2341,2343,2345,2347,2349,2351,2353,2355,2357,2359,2361,2363,2365,2367,2369,2371,2373,2375,2377,2379,2381,2383,2385,2387,2389,2391,2393,2395,2397,2399,2401,2403,2405,2407,2409,2411,2413,2415,2417,2419,2421,2423,2425,2427,2429,2431,2433,2435,2437,2439,2441,2443,2445,2447,2449,2451,2453,2455,2457,2459,2461,2463,2465,2467,2469,2471,2473,2475,2477,2479,2481,2483,2485,2487,2489,2491,2493,2495,2497,2499,2501,2503,2505,2507,2509,2511,2513,2515,2517,2519,2521,2523,2525,2527,2529,2531,2533,2535,2537,2539,2541,2543,2545,2547,2549,2551,2553,2555,2557,2559,2561,2563,2565,2567,2569,2571,2573,2575,2577,2579,2581,2583,2585,2587,2589,2591,2593,2595,2597,2599,2601,2603,2605,2607,2609,2611,2613,2615,2617,2619,2621,2623,2625,2627,2629,2631,2633,2635,2637,2639,2641,2643,2645,2647,2649,2651,2653,2655,2657,2659,2661,2663,2665,2667,2669,2671,2673,2675,2677,2679,2681,2683,2685,2687,2689,2691,2693,2695,2697,2699,2701,2703,2705,2707,2709,2711,2713,2715,2717,2719,2721,2723,2725,2727,2729,2731,2733,2735,2737,2739,2741,2743,2745,2747,2749,2751,2753,2755,2757,2759,2761,2763,2765,2767,2769,2771,2773,2775,2777,2779,2781,2783,2785,2787,2789,2791,2793,2795,2797,2799,2801,2803,2805,2807,2809,2811,2813,2815,2817,2819,2821,2823,2825,2827,2829,2831,2833,2835,2837,2839,2841,2843,2845,2847,2849,2851,2853,2855,2857,2859,2861,2863,2865,2867,2869,2871,2873,2875,2877,2879,2881,2883,2885,2887,2889,2891,2893,2895,2897,2899,2901,2903,2905,2907,2909,2911,2913,2915,2917,2919,2921,2923,2925,2927,2929,2931,2933,2935,2937,2939,2941,2943,2945,2947,2949,2951,2953,2955,2957,2959,2961,2963,2965,2967,2969,2971,2973,2975,2977,2979,2981,2983,2985,2987,2989,2991,2993,2995,2997,2999,3001,3003,3005,3007,3009,3011,3013,3015,3017,3019,3021,3023,3025,3027,3029,3031,3033,3035,3037,3039,3041,3043,3045,3047,3049,3051,3053,3055,3057,3059,3061,3063,3065,3067,3069,3071,3073,3075,3077,3079,3081,3083,3085,3087,3089,3091,3093,3095,3097,3099,3101,3103,3105,3107,3109,3111,3113,3115,3117,3119,3121,3123,3125,3127,3129,3131,3133,3135,3137,3139,3141,3143,3145,3147,3149,3151,3153,3155,3157,3159,3161,3163,3165,3167,3169,3171,3173,3175,3177,3179,3181,3183,3185,3187,3189,3191,3193,3195,3197,3199,3201,3203,3205,3207,3209,3211,3213,3215,3217,3219,3221,3223,3225,3227,3229,3231,3233,3235,3237,3239,3241,3243,3245,3247,3249,3251,3253,3255,3257,3259,3261,3263,3265,3267,3269,3271,3273,3275,3277,3279,3281,3283,3285,3287,3289,3291,3293,3295,3297,3299,3301,3303,3305,3307,3309,3311,3313,3315,3317,3319,3321,3323,3325,3327,3329,3331,3333,3335,3337,3339,3341,3343,3345,3347,3349,3351,3353,3355,3357,3359,3361,3363,3365,3367,3369,3371,3373,3375,3377,3379,3381,3383,3385,3387,3389,3391,3393,3395,3397,3399,3401,3403,3405,3407,3409,3411,3413,3415,3417,3419,3421,3423,3425,3427,3429,3431,3433,3435,3437,3439,3441,3443,3445,3447,3449,3451,3453,3455,3457,3459,3461,3463,3465,3467,3469,3471,3473,3475,3477,3479,3481,3483,3485,3487,3489,3491,3493,3495,3497,3499,3501,3503,3505,3507,3509,3511,3513,3515,3517,3519,3521,3523,3525,3527,3529,3531,3533,3535,3537,3539,3541,3543,3545,3547,3549,3551,3553,3555,3557,3559,3561,3563,3565,3567,3569,3571,3573,3575,3577,3579,3581,3583,3585,3587,3589,3591,3593,3595,3597,3599,3601,3603,3605,3607,3609,3611,3613,3615,3617,3619,3621,3623,3625,3627,3629,3631,3633,3635,3637,3639,3641,3643,3645,3647,3649,3651,3653,3655,3657,3659,3661,3663,3665,3667,3669,3671,3673,3675,3677,3679,3681,3683,3685,3687,3689,3691,3693,3695],{"categories":114},[115],"Developer Productivity",{"categories":117},[118],"Business & SaaS",{"categories":120},[77],{"categories":122},[123],"AI Automation",{"categories":125},[126],"Product Strategy",{"categories":128},[77],{"categories":130},[115],{"categories":132},[118],{"categories":134},[],{"categories":136},[77],{"categories":138},[],{"categories":140},[141],"AI News & Trends",{"categories":143},[123],{"categories":145},[141],{"categories":147},[123],{"categories":149},[123],{"categories":151},[77],{"categories":153},[77],{"categories":155},[141],{"categories":157},[77],{"categories":159},[],{"categories":161},[162],"Design & Frontend",{"categories":164},[165],"Data Science & Visualization",{"categories":167},[141],{"categories":169},[],{"categories":171},[172],"Software Engineering",{"categories":174},[77],{"categories":176},[123],{"categories":178},[179],"Marketing & Growth",{"categories":181},[77],{"categories":183},[123],{"categories":185},[],{"categories":187},[],{"categories":189},[162],{"categories":191},[123],{"categories":193},[115],{"categories":195},[162],{"categories":197},[77],{"categories":199},[123],{"categories":201},[141],{"categories":203},[],{"categories":205},[],{"categories":207},[123],{"categories":209},[172],{"categories":211},[],{"categories":213},[118],{"categories":215},[],{"categories":217},[],{"categories":219},[123],{"categories":221},[123],{"categories":223},[77],{"categories":225},[],{"categories":227},[172],{"categories":229},[],{"categories":231},[],{"categories":233},[],{"categories":235},[77],{"categories":237},[179],{"categories":239},[162],{"categories":241},[162],{"categories":243},[77],{"categories":245},[123],{"categories":247},[77],{"categories":249},[77],{"categories":251},[123],{"categories":253},[123],{"categories":255},[165],{"categories":257},[141],{"categories":259},[123],{"categories":261},[179],{"categories":263},[123],{"categories":265},[126],{"categories":267},[],{"categories":269},[123],{"categories":271},[],{"categories":273},[123],{"categories":275},[172],{"categories":277},[162],{"categories":279},[77],{"categories":281},[],{"categories":283},[],{"categories":285},[123],{"categories":287},[],{"categories":289},[77],{"categories":291},[],{"categories":293},[115],{"categories":295},[172],{"categories":297},[118],{"categories":299},[141],{"categories":301},[77],{"categories":303},[],{"categories":305},[77],{"categories":307},[],{"categories":309},[172],{"categories":311},[165],{"categories":313},[],{"categories":315},[77],{"categories":317},[162],{"categories":319},[],{"categories":321},[162],{"categories":323},[123],{"categories":325},[],{"categories":327},[123],{"categories":329},[141],{"categories":331},[118],{"categories":333},[77],{"categories":335},[],{"categories":337},[123],{"categories":339},[77],{"categories":341},[126],{"categories":343},[],{"categories":345},[77],{"categories":347},[123],{"categories":349},[123],{"categories":351},[],{"categories":353},[165],{"categories":355},[77],{"categories":357},[],{"categories":359},[115],{"categories":361},[118],{"categories":363},[77],{"categories":365},[123],{"categories":367},[172],{"categories":369},[77],{"categories":371},[],{"categories":373},[],{"categories":375},[77],{"categories":377},[],{"categories":379},[162],{"categories":381},[],{"categories":383},[77],{"categories":385},[],{"categories":387},[123],{"categories":389},[77],{"categories":391},[162],{"categories":393},[],{"categories":395},[77],{"categories":397},[77],{"categories":399},[118],{"categories":401},[123],{"categories":403},[77],{"categories":405},[162],{"categories":407},[123],{"categories":409},[],{"categories":411},[],{"categories":413},[141],{"categories":415},[],{"categories":417},[77],{"categories":419},[118,179],{"categories":421},[],{"categories":423},[77],{"categories":425},[],{"categories":427},[],{"categories":429},[77],{"categories":431},[],{"categories":433},[77],{"categories":435},[436],"DevOps & Cloud",{"categories":438},[],{"categories":440},[141],{"categories":442},[162],{"categories":444},[],{"categories":446},[141],{"categories":448},[141],{"categories":450},[77],{"categories":452},[179],{"categories":454},[],{"categories":456},[118],{"categories":458},[],{"categories":460},[77,436],{"categories":462},[77],{"categories":464},[77],{"categories":466},[123],{"categories":468},[77,172],{"categories":470},[165],{"categories":472},[77],{"categories":474},[179],{"categories":476},[123],{"categories":478},[123],{"categories":480},[],{"categories":482},[123],{"categories":484},[77,118],{"categories":486},[],{"categories":488},[162],{"categories":490},[162],{"categories":492},[],{"categories":494},[],{"categories":496},[141],{"categories":498},[],{"categories":500},[115],{"categories":502},[172],{"categories":504},[77],{"categories":506},[162],{"categories":508},[123],{"categories":510},[172],{"categories":512},[141],{"categories":514},[162],{"categories":516},[],{"categories":518},[77],{"categories":520},[77],{"categories":522},[77],{"categories":524},[141],{"categories":526},[115],{"categories":528},[77],{"categories":530},[123],{"categories":532},[436],{"categories":534},[162],{"categories":536},[123],{"categories":538},[],{"categories":540},[],{"categories":542},[162],{"categories":544},[141],{"categories":546},[165],{"categories":548},[],{"categories":550},[77],{"categories":552},[77],{"categories":554},[118],{"categories":556},[77],{"categories":558},[77],{"categories":560},[141],{"categories":562},[],{"categories":564},[123],{"categories":566},[172],{"categories":568},[],{"categories":570},[77],{"categories":572},[77],{"categories":574},[123],{"categories":576},[],{"categories":578},[],{"categories":580},[77],{"categories":582},[],{"categories":584},[118],{"categories":586},[123],{"categories":588},[],{"categories":590},[115],{"categories":592},[77],{"categories":594},[118],{"categories":596},[141],{"categories":598},[],{"categories":600},[],{"categories":602},[],{"categories":604},[141],{"categories":606},[141],{"categories":608},[],{"categories":610},[],{"categories":612},[118],{"categories":614},[],{"categories":616},[],{"categories":618},[115],{"categories":620},[],{"categories":622},[179],{"categories":624},[123],{"categories":626},[118],{"categories":628},[123],{"categories":630},[172],{"categories":632},[],{"categories":634},[126],{"categories":636},[162],{"categories":638},[172],{"categories":640},[77],{"categories":642},[123],{"categories":644},[118],{"categories":646},[77],{"categories":648},[],{"categories":650},[],{"categories":652},[172],{"categories":654},[165],{"categories":656},[126],{"categories":658},[123],{"categories":660},[77],{"categories":662},[],{"categories":664},[436],{"categories":666},[],{"categories":668},[123],{"categories":670},[],{"categories":672},[],{"categories":674},[77],{"categories":676},[162],{"categories":678},[179],{"categories":680},[123],{"categories":682},[],{"categories":684},[115],{"categories":686},[],{"categories":688},[141],{"categories":690},[77,436],{"categories":692},[141],{"categories":694},[77],{"categories":696},[118],{"categories":698},[77],{"categories":700},[],{"categories":702},[118],{"categories":704},[],{"categories":706},[172],{"categories":708},[162],{"categories":710},[141],{"categories":712},[165],{"categories":714},[115],{"categories":716},[77],{"categories":718},[172],{"categories":720},[],{"categories":722},[],{"categories":724},[126],{"categories":726},[],{"categories":728},[77],{"categories":730},[],{"categories":732},[162],{"categories":734},[162],{"categories":736},[162],{"categories":738},[],{"categories":740},[],{"categories":742},[141],{"categories":744},[123],{"categories":746},[77],{"categories":748},[77],{"categories":750},[77],{"categories":752},[118],{"categories":754},[77],{"categories":756},[],{"categories":758},[172],{"categories":760},[172],{"categories":762},[118],{"categories":764},[],{"categories":766},[77],{"categories":768},[77],{"categories":770},[118],{"categories":772},[141],{"categories":774},[179],{"categories":776},[123],{"categories":778},[],{"categories":780},[162],{"categories":782},[],{"categories":784},[77],{"categories":786},[],{"categories":788},[118],{"categories":790},[123],{"categories":792},[],{"categories":794},[436],{"categories":796},[165],{"categories":798},[172],{"categories":800},[179],{"categories":802},[172],{"categories":804},[123],{"categories":806},[],{"categories":808},[],{"categories":810},[123],{"categories":812},[115],{"categories":814},[123],{"categories":816},[126],{"categories":818},[118],{"categories":820},[],{"categories":822},[77],{"categories":824},[126],{"categories":826},[77],{"categories":828},[77],{"categories":830},[179],{"categories":832},[162],{"categories":834},[123],{"categories":836},[],{"categories":838},[],{"categories":840},[436],{"categories":842},[172],{"categories":844},[],{"categories":846},[123],{"categories":848},[77],{"categories":850},[162,77],{"categories":852},[115],{"categories":854},[],{"categories":856},[77],{"categories":858},[115],{"categories":860},[162],{"categories":862},[123],{"categories":864},[172],{"categories":866},[],{"categories":868},[77],{"categories":870},[],{"categories":872},[115],{"categories":874},[],{"categories":876},[123],{"categories":878},[126],{"categories":880},[77],{"categories":882},[77],{"categories":884},[162],{"categories":886},[123],{"categories":888},[436],{"categories":890},[162],{"categories":892},[123],{"categories":894},[77],{"categories":896},[77],{"categories":898},[77],{"categories":900},[141],{"categories":902},[],{"categories":904},[126],{"categories":906},[123],{"categories":908},[162],{"categories":910},[123],{"categories":912},[172],{"categories":914},[162],{"categories":916},[123],{"categories":918},[141],{"categories":920},[],{"categories":922},[77],{"categories":924},[162],{"categories":926},[77],{"categories":928},[115],{"categories":930},[141],{"categories":932},[77],{"categories":934},[179],{"categories":936},[77],{"categories":938},[77],{"categories":940},[123],{"categories":942},[123],{"categories":944},[77],{"categories":946},[123],{"categories":948},[162],{"categories":950},[77],{"categories":952},[],{"categories":954},[],{"categories":956},[172],{"categories":958},[],{"categories":960},[115],{"categories":962},[436],{"categories":964},[],{"categories":966},[115],{"categories":968},[118],{"categories":970},[179],{"categories":972},[],{"categories":974},[118],{"categories":976},[],{"categories":978},[],{"categories":980},[],{"categories":982},[],{"categories":984},[],{"categories":986},[77],{"categories":988},[123],{"categories":990},[436],{"categories":992},[115],{"categories":994},[77],{"categories":996},[172],{"categories":998},[126],{"categories":1000},[77],{"categories":1002},[179],{"categories":1004},[77],{"categories":1006},[77],{"categories":1008},[77],{"categories":1010},[77,115],{"categories":1012},[172],{"categories":1014},[172],{"categories":1016},[162],{"categories":1018},[77],{"categories":1020},[],{"categories":1022},[],{"categories":1024},[],{"categories":1026},[172],{"categories":1028},[165],{"categories":1030},[141],{"categories":1032},[162],{"categories":1034},[],{"categories":1036},[77],{"categories":1038},[77],{"categories":1040},[],{"categories":1042},[],{"categories":1044},[123],{"categories":1046},[77],{"categories":1048},[118],{"categories":1050},[],{"categories":1052},[115],{"categories":1054},[77],{"categories":1056},[115],{"categories":1058},[77],{"categories":1060},[172],{"categories":1062},[179],{"categories":1064},[77,162],{"categories":1066},[141],{"categories":1068},[162],{"categories":1070},[],{"categories":1072},[436],{"categories":1074},[162],{"categories":1076},[123],{"categories":1078},[],{"categories":1080},[],{"categories":1082},[],{"categories":1084},[],{"categories":1086},[172],{"categories":1088},[123],{"categories":1090},[123],{"categories":1092},[436],{"categories":1094},[77],{"categories":1096},[77],{"categories":1098},[77],{"categories":1100},[],{"categories":1102},[162],{"categories":1104},[],{"categories":1106},[],{"categories":1108},[123],{"categories":1110},[],{"categories":1112},[],{"categories":1114},[179],{"categories":1116},[179],{"categories":1118},[123],{"categories":1120},[],{"categories":1122},[77],{"categories":1124},[77],{"categories":1126},[172],{"categories":1128},[162],{"categories":1130},[162],{"categories":1132},[123],{"categories":1134},[115],{"categories":1136},[77],{"categories":1138},[162],{"categories":1140},[162],{"categories":1142},[123],{"categories":1144},[123],{"categories":1146},[77],{"categories":1148},[],{"categories":1150},[],{"categories":1152},[77],{"categories":1154},[123],{"categories":1156},[141],{"categories":1158},[172],{"categories":1160},[115],{"categories":1162},[77],{"categories":1164},[],{"categories":1166},[123],{"categories":1168},[123],{"categories":1170},[],{"categories":1172},[115],{"categories":1174},[77],{"categories":1176},[115],{"categories":1178},[115],{"categories":1180},[],{"categories":1182},[],{"categories":1184},[123],{"categories":1186},[123],{"categories":1188},[77],{"categories":1190},[77],{"categories":1192},[141],{"categories":1194},[165],{"categories":1196},[126],{"categories":1198},[141],{"categories":1200},[162],{"categories":1202},[],{"categories":1204},[141],{"categories":1206},[],{"categories":1208},[],{"categories":1210},[],{"categories":1212},[],{"categories":1214},[172],{"categories":1216},[165],{"categories":1218},[],{"categories":1220},[77],{"categories":1222},[77],{"categories":1224},[165],{"categories":1226},[172],{"categories":1228},[],{"categories":1230},[],{"categories":1232},[123],{"categories":1234},[141],{"categories":1236},[141],{"categories":1238},[123],{"categories":1240},[115],{"categories":1242},[77,436],{"categories":1244},[],{"categories":1246},[162],{"categories":1248},[115],{"categories":1250},[123],{"categories":1252},[162],{"categories":1254},[],{"categories":1256},[123],{"categories":1258},[123],{"categories":1260},[77],{"categories":1262},[179],{"categories":1264},[172],{"categories":1266},[162],{"categories":1268},[],{"categories":1270},[123],{"categories":1272},[77],{"categories":1274},[123],{"categories":1276},[123],{"categories":1278},[123],{"categories":1280},[179],{"categories":1282},[123],{"categories":1284},[77],{"categories":1286},[],{"categories":1288},[179],{"categories":1290},[141],{"categories":1292},[123],{"categories":1294},[],{"categories":1296},[],{"categories":1298},[77],{"categories":1300},[123],{"categories":1302},[141],{"categories":1304},[123],{"categories":1306},[],{"categories":1308},[],{"categories":1310},[],{"categories":1312},[123],{"categories":1314},[],{"categories":1316},[],{"categories":1318},[165],{"categories":1320},[77],{"categories":1322},[165],{"categories":1324},[141],{"categories":1326},[77],{"categories":1328},[77],{"categories":1330},[123],{"categories":1332},[77],{"categories":1334},[],{"categories":1336},[],{"categories":1338},[436],{"categories":1340},[],{"categories":1342},[],{"categories":1344},[115],{"categories":1346},[],{"categories":1348},[],{"categories":1350},[],{"categories":1352},[],{"categories":1354},[172],{"categories":1356},[141],{"categories":1358},[179],{"categories":1360},[118],{"categories":1362},[77],{"categories":1364},[77],{"categories":1366},[118],{"categories":1368},[],{"categories":1370},[162],{"categories":1372},[123],{"categories":1374},[118],{"categories":1376},[77],{"categories":1378},[77],{"categories":1380},[115],{"categories":1382},[],{"categories":1384},[115],{"categories":1386},[77],{"categories":1388},[179],{"categories":1390},[123],{"categories":1392},[141],{"categories":1394},[118],{"categories":1396},[77],{"categories":1398},[123],{"categories":1400},[],{"categories":1402},[77],{"categories":1404},[115],{"categories":1406},[77],{"categories":1408},[],{"categories":1410},[141],{"categories":1412},[77],{"categories":1414},[],{"categories":1416},[118],{"categories":1418},[77],{"categories":1420},[],{"categories":1422},[],{"categories":1424},[],{"categories":1426},[77],{"categories":1428},[],{"categories":1430},[436],{"categories":1432},[77],{"categories":1434},[],{"categories":1436},[77],{"categories":1438},[77],{"categories":1440},[77],{"categories":1442},[77,436],{"categories":1444},[77],{"categories":1446},[77],{"categories":1448},[162],{"categories":1450},[123],{"categories":1452},[],{"categories":1454},[123],{"categories":1456},[77],{"categories":1458},[77],{"categories":1460},[77],{"categories":1462},[115],{"categories":1464},[115],{"categories":1466},[172],{"categories":1468},[162],{"categories":1470},[123],{"categories":1472},[],{"categories":1474},[77],{"categories":1476},[141],{"categories":1478},[77],{"categories":1480},[118],{"categories":1482},[],{"categories":1484},[436],{"categories":1486},[162],{"categories":1488},[162],{"categories":1490},[123],{"categories":1492},[141],{"categories":1494},[123],{"categories":1496},[77],{"categories":1498},[],{"categories":1500},[77],{"categories":1502},[],{"categories":1504},[],{"categories":1506},[77],{"categories":1508},[77],{"categories":1510},[77],{"categories":1512},[123],{"categories":1514},[77],{"categories":1516},[],{"categories":1518},[165],{"categories":1520},[123],{"categories":1522},[],{"categories":1524},[],{"categories":1526},[77],{"categories":1528},[141],{"categories":1530},[],{"categories":1532},[162],{"categories":1534},[436],{"categories":1536},[141],{"categories":1538},[172],{"categories":1540},[172],{"categories":1542},[141],{"categories":1544},[141],{"categories":1546},[436],{"categories":1548},[],{"categories":1550},[141],{"categories":1552},[77],{"categories":1554},[115],{"categories":1556},[141],{"categories":1558},[],{"categories":1560},[165],{"categories":1562},[141],{"categories":1564},[172],{"categories":1566},[141],{"categories":1568},[436],{"categories":1570},[77],{"categories":1572},[77],{"categories":1574},[],{"categories":1576},[118],{"categories":1578},[],{"categories":1580},[],{"categories":1582},[77],{"categories":1584},[77],{"categories":1586},[77],{"categories":1588},[77],{"categories":1590},[],{"categories":1592},[165],{"categories":1594},[115],{"categories":1596},[],{"categories":1598},[77],{"categories":1600},[77],{"categories":1602},[436],{"categories":1604},[436],{"categories":1606},[],{"categories":1608},[123],{"categories":1610},[141],{"categories":1612},[141],{"categories":1614},[77],{"categories":1616},[123],{"categories":1618},[],{"categories":1620},[162],{"categories":1622},[77],{"categories":1624},[77],{"categories":1626},[],{"categories":1628},[],{"categories":1630},[436],{"categories":1632},[77],{"categories":1634},[172],{"categories":1636},[118],{"categories":1638},[77],{"categories":1640},[],{"categories":1642},[123],{"categories":1644},[115],{"categories":1646},[115],{"categories":1648},[],{"categories":1650},[77],{"categories":1652},[162],{"categories":1654},[123],{"categories":1656},[],{"categories":1658},[77],{"categories":1660},[77],{"categories":1662},[123],{"categories":1664},[],{"categories":1666},[123],{"categories":1668},[172],{"categories":1670},[],{"categories":1672},[77],{"categories":1674},[],{"categories":1676},[77],{"categories":1678},[],{"categories":1680},[77],{"categories":1682},[77],{"categories":1684},[],{"categories":1686},[77],{"categories":1688},[141],{"categories":1690},[77],{"categories":1692},[77],{"categories":1694},[115],{"categories":1696},[77],{"categories":1698},[141],{"categories":1700},[123],{"categories":1702},[],{"categories":1704},[77],{"categories":1706},[179],{"categories":1708},[],{"categories":1710},[],{"categories":1712},[],{"categories":1714},[115],{"categories":1716},[141],{"categories":1718},[123],{"categories":1720},[77],{"categories":1722},[162],{"categories":1724},[123],{"categories":1726},[],{"categories":1728},[123],{"categories":1730},[],{"categories":1732},[77],{"categories":1734},[123],{"categories":1736},[77],{"categories":1738},[],{"categories":1740},[77],{"categories":1742},[77],{"categories":1744},[141],{"categories":1746},[162],{"categories":1748},[123],{"categories":1750},[162],{"categories":1752},[118],{"categories":1754},[],{"categories":1756},[],{"categories":1758},[77],{"categories":1760},[115],{"categories":1762},[141],{"categories":1764},[],{"categories":1766},[],{"categories":1768},[172],{"categories":1770},[162],{"categories":1772},[],{"categories":1774},[77],{"categories":1776},[],{"categories":1778},[179],{"categories":1780},[77],{"categories":1782},[436],{"categories":1784},[172],{"categories":1786},[],{"categories":1788},[123],{"categories":1790},[77],{"categories":1792},[123],{"categories":1794},[123],{"categories":1796},[77],{"categories":1798},[],{"categories":1800},[115],{"categories":1802},[77],{"categories":1804},[118],{"categories":1806},[172],{"categories":1808},[162],{"categories":1810},[],{"categories":1812},[],{"categories":1814},[],{"categories":1816},[123],{"categories":1818},[162],{"categories":1820},[141],{"categories":1822},[77],{"categories":1824},[141],{"categories":1826},[162],{"categories":1828},[],{"categories":1830},[162],{"categories":1832},[141],{"categories":1834},[118],{"categories":1836},[77],{"categories":1838},[141],{"categories":1840},[179],{"categories":1842},[],{"categories":1844},[],{"categories":1846},[165],{"categories":1848},[77,172],{"categories":1850},[141],{"categories":1852},[77],{"categories":1854},[123],{"categories":1856},[123],{"categories":1858},[77],{"categories":1860},[],{"categories":1862},[172],{"categories":1864},[77],{"categories":1866},[165],{"categories":1868},[123],{"categories":1870},[179],{"categories":1872},[436],{"categories":1874},[],{"categories":1876},[115],{"categories":1878},[123],{"categories":1880},[123],{"categories":1882},[172],{"categories":1884},[77],{"categories":1886},[77],{"categories":1888},[],{"categories":1890},[],{"categories":1892},[],{"categories":1894},[436],{"categories":1896},[141],{"categories":1898},[77],{"categories":1900},[77],{"categories":1902},[77],{"categories":1904},[],{"categories":1906},[165],{"categories":1908},[118],{"categories":1910},[],{"categories":1912},[123],{"categories":1914},[436],{"categories":1916},[],{"categories":1918},[162],{"categories":1920},[162],{"categories":1922},[],{"categories":1924},[172],{"categories":1926},[162],{"categories":1928},[77],{"categories":1930},[],{"categories":1932},[141],{"categories":1934},[77],{"categories":1936},[162],{"categories":1938},[123],{"categories":1940},[141],{"categories":1942},[],{"categories":1944},[123],{"categories":1946},[162],{"categories":1948},[77],{"categories":1950},[],{"categories":1952},[77],{"categories":1954},[77],{"categories":1956},[436],{"categories":1958},[141],{"categories":1960},[165],{"categories":1962},[165],{"categories":1964},[],{"categories":1966},[],{"categories":1968},[],{"categories":1970},[123],{"categories":1972},[172],{"categories":1974},[172],{"categories":1976},[],{"categories":1978},[],{"categories":1980},[77],{"categories":1982},[],{"categories":1984},[123],{"categories":1986},[77],{"categories":1988},[],{"categories":1990},[77],{"categories":1992},[118],{"categories":1994},[77],{"categories":1996},[179],{"categories":1998},[123],{"categories":2000},[77],{"categories":2002},[172],{"categories":2004},[141],{"categories":2006},[123],{"categories":2008},[],{"categories":2010},[141],{"categories":2012},[123],{"categories":2014},[123],{"categories":2016},[],{"categories":2018},[118],{"categories":2020},[123],{"categories":2022},[],{"categories":2024},[77],{"categories":2026},[115],{"categories":2028},[141],{"categories":2030},[436],{"categories":2032},[123],{"categories":2034},[123],{"categories":2036},[115],{"categories":2038},[77],{"categories":2040},[],{"categories":2042},[],{"categories":2044},[162],{"categories":2046},[77,118],{"categories":2048},[],{"categories":2050},[115],{"categories":2052},[165],{"categories":2054},[77],{"categories":2056},[172],{"categories":2058},[77],{"categories":2060},[123],{"categories":2062},[77],{"categories":2064},[77],{"categories":2066},[141],{"categories":2068},[123],{"categories":2070},[],{"categories":2072},[],{"categories":2074},[123],{"categories":2076},[77],{"categories":2078},[436],{"categories":2080},[],{"categories":2082},[77],{"categories":2084},[123],{"categories":2086},[],{"categories":2088},[77],{"categories":2090},[179],{"categories":2092},[165],{"categories":2094},[123],{"categories":2096},[77],{"categories":2098},[436],{"categories":2100},[],{"categories":2102},[77],{"categories":2104},[179],{"categories":2106},[162],{"categories":2108},[77],{"categories":2110},[],{"categories":2112},[179],{"categories":2114},[141],{"categories":2116},[77],{"categories":2118},[77],{"categories":2120},[115],{"categories":2122},[],{"categories":2124},[],{"categories":2126},[162],{"categories":2128},[77],{"categories":2130},[165],{"categories":2132},[179],{"categories":2134},[179],{"categories":2136},[141],{"categories":2138},[],{"categories":2140},[],{"categories":2142},[77],{"categories":2144},[],{"categories":2146},[77,172],{"categories":2148},[141],{"categories":2150},[123],{"categories":2152},[172],{"categories":2154},[77],{"categories":2156},[115],{"categories":2158},[],{"categories":2160},[],{"categories":2162},[115],{"categories":2164},[179],{"categories":2166},[77],{"categories":2168},[],{"categories":2170},[162,77],{"categories":2172},[436],{"categories":2174},[115],{"categories":2176},[],{"categories":2178},[118],{"categories":2180},[118],{"categories":2182},[77],{"categories":2184},[172],{"categories":2186},[123],{"categories":2188},[141],{"categories":2190},[179],{"categories":2192},[162],{"categories":2194},[77],{"categories":2196},[77],{"categories":2198},[77],{"categories":2200},[115],{"categories":2202},[77],{"categories":2204},[123],{"categories":2206},[141],{"categories":2208},[],{"categories":2210},[],{"categories":2212},[165],{"categories":2214},[172],{"categories":2216},[77],{"categories":2218},[162],{"categories":2220},[165],{"categories":2222},[77],{"categories":2224},[77],{"categories":2226},[123],{"categories":2228},[123],{"categories":2230},[77,118],{"categories":2232},[],{"categories":2234},[162],{"categories":2236},[],{"categories":2238},[77],{"categories":2240},[141],{"categories":2242},[115],{"categories":2244},[115],{"categories":2246},[123],{"categories":2248},[77],{"categories":2250},[118],{"categories":2252},[172],{"categories":2254},[179],{"categories":2256},[],{"categories":2258},[141],{"categories":2260},[77],{"categories":2262},[77],{"categories":2264},[141],{"categories":2266},[172],{"categories":2268},[77],{"categories":2270},[123],{"categories":2272},[141],{"categories":2274},[77],{"categories":2276},[162],{"categories":2278},[77],{"categories":2280},[77],{"categories":2282},[436],{"categories":2284},[126],{"categories":2286},[123],{"categories":2288},[77],{"categories":2290},[141],{"categories":2292},[123],{"categories":2294},[179],{"categories":2296},[77],{"categories":2298},[],{"categories":2300},[77],{"categories":2302},[],{"categories":2304},[],{"categories":2306},[],{"categories":2308},[118],{"categories":2310},[77],{"categories":2312},[123],{"categories":2314},[141],{"categories":2316},[141],{"categories":2318},[141],{"categories":2320},[141],{"categories":2322},[],{"categories":2324},[115],{"categories":2326},[123],{"categories":2328},[141],{"categories":2330},[115],{"categories":2332},[123],{"categories":2334},[77],{"categories":2336},[77,123],{"categories":2338},[123],{"categories":2340},[436],{"categories":2342},[141],{"categories":2344},[141],{"categories":2346},[123],{"categories":2348},[77],{"categories":2350},[],{"categories":2352},[141],{"categories":2354},[179],{"categories":2356},[115],{"categories":2358},[77],{"categories":2360},[77],{"categories":2362},[],{"categories":2364},[172],{"categories":2366},[],{"categories":2368},[115],{"categories":2370},[123],{"categories":2372},[141],{"categories":2374},[77],{"categories":2376},[141],{"categories":2378},[115],{"categories":2380},[141],{"categories":2382},[141],{"categories":2384},[],{"categories":2386},[118],{"categories":2388},[123],{"categories":2390},[141],{"categories":2392},[141],{"categories":2394},[141],{"categories":2396},[141],{"categories":2398},[141],{"categories":2400},[141],{"categories":2402},[141],{"categories":2404},[141],{"categories":2406},[141],{"categories":2408},[141],{"categories":2410},[165],{"categories":2412},[115],{"categories":2414},[77],{"categories":2416},[77],{"categories":2418},[],{"categories":2420},[77,115],{"categories":2422},[],{"categories":2424},[123],{"categories":2426},[141],{"categories":2428},[123],{"categories":2430},[77],{"categories":2432},[77],{"categories":2434},[77],{"categories":2436},[77],{"categories":2438},[77],{"categories":2440},[123],{"categories":2442},[118],{"categories":2444},[162],{"categories":2446},[141],{"categories":2448},[77],{"categories":2450},[],{"categories":2452},[],{"categories":2454},[123],{"categories":2456},[162],{"categories":2458},[77],{"categories":2460},[],{"categories":2462},[],{"categories":2464},[179],{"categories":2466},[77],{"categories":2468},[],{"categories":2470},[],{"categories":2472},[115],{"categories":2474},[118],{"categories":2476},[77],{"categories":2478},[118],{"categories":2480},[162],{"categories":2482},[],{"categories":2484},[141],{"categories":2486},[],{"categories":2488},[162],{"categories":2490},[77],{"categories":2492},[179],{"categories":2494},[],{"categories":2496},[179],{"categories":2498},[],{"categories":2500},[],{"categories":2502},[123],{"categories":2504},[],{"categories":2506},[118],{"categories":2508},[115],{"categories":2510},[162],{"categories":2512},[172],{"categories":2514},[],{"categories":2516},[],{"categories":2518},[77],{"categories":2520},[115],{"categories":2522},[179],{"categories":2524},[],{"categories":2526},[123],{"categories":2528},[123],{"categories":2530},[141],{"categories":2532},[77],{"categories":2534},[123],{"categories":2536},[77],{"categories":2538},[123],{"categories":2540},[77],{"categories":2542},[126],{"categories":2544},[141],{"categories":2546},[],{"categories":2548},[179],{"categories":2550},[172],{"categories":2552},[123],{"categories":2554},[],{"categories":2556},[77],{"categories":2558},[123],{"categories":2560},[118],{"categories":2562},[115],{"categories":2564},[77],{"categories":2566},[162],{"categories":2568},[172],{"categories":2570},[172],{"categories":2572},[77],{"categories":2574},[165],{"categories":2576},[77],{"categories":2578},[123],{"categories":2580},[118],{"categories":2582},[123],{"categories":2584},[77],{"categories":2586},[77],{"categories":2588},[123],{"categories":2590},[141],{"categories":2592},[],{"categories":2594},[115],{"categories":2596},[77],{"categories":2598},[123],{"categories":2600},[77],{"categories":2602},[77],{"categories":2604},[],{"categories":2606},[162],{"categories":2608},[118],{"categories":2610},[141],{"categories":2612},[77],{"categories":2614},[77],{"categories":2616},[162],{"categories":2618},[179],{"categories":2620},[165],{"categories":2622},[77],{"categories":2624},[141],{"categories":2626},[77],{"categories":2628},[123],{"categories":2630},[436],{"categories":2632},[77],{"categories":2634},[123],{"categories":2636},[165],{"categories":2638},[],{"categories":2640},[123],{"categories":2642},[172],{"categories":2644},[162],{"categories":2646},[77],{"categories":2648},[115],{"categories":2650},[118],{"categories":2652},[172],{"categories":2654},[],{"categories":2656},[123],{"categories":2658},[77],{"categories":2660},[],{"categories":2662},[141],{"categories":2664},[],{"categories":2666},[141],{"categories":2668},[77],{"categories":2670},[123],{"categories":2672},[123],{"categories":2674},[123],{"categories":2676},[],{"categories":2678},[],{"categories":2680},[77],{"categories":2682},[77],{"categories":2684},[],{"categories":2686},[162],{"categories":2688},[123],{"categories":2690},[179],{"categories":2692},[115],{"categories":2694},[],{"categories":2696},[],{"categories":2698},[141],{"categories":2700},[172],{"categories":2702},[77],{"categories":2704},[77],{"categories":2706},[77],{"categories":2708},[172],{"categories":2710},[141],{"categories":2712},[162],{"categories":2714},[77],{"categories":2716},[77],{"categories":2718},[77],{"categories":2720},[141],{"categories":2722},[77],{"categories":2724},[141],{"categories":2726},[123],{"categories":2728},[123],{"categories":2730},[172],{"categories":2732},[123],{"categories":2734},[77],{"categories":2736},[172],{"categories":2738},[162],{"categories":2740},[],{"categories":2742},[123],{"categories":2744},[],{"categories":2746},[],{"categories":2748},[],{"categories":2750},[118],{"categories":2752},[77],{"categories":2754},[123],{"categories":2756},[115],{"categories":2758},[123],{"categories":2760},[179],{"categories":2762},[],{"categories":2764},[123],{"categories":2766},[],{"categories":2768},[115],{"categories":2770},[123],{"categories":2772},[],{"categories":2774},[123],{"categories":2776},[77],{"categories":2778},[141],{"categories":2780},[77],{"categories":2782},[123],{"categories":2784},[141],{"categories":2786},[123],{"categories":2788},[172],{"categories":2790},[162],{"categories":2792},[115],{"categories":2794},[],{"categories":2796},[123],{"categories":2798},[162],{"categories":2800},[436],{"categories":2802},[141],{"categories":2804},[77],{"categories":2806},[162],{"categories":2808},[115],{"categories":2810},[],{"categories":2812},[123],{"categories":2814},[123],{"categories":2816},[77],{"categories":2818},[],{"categories":2820},[123],{"categories":2822},[126],{"categories":2824},[141],{"categories":2826},[123],{"categories":2828},[118],{"categories":2830},[],{"categories":2832},[77],{"categories":2834},[126],{"categories":2836},[77],{"categories":2838},[123],{"categories":2840},[141],{"categories":2842},[115],{"categories":2844},[436],{"categories":2846},[77],{"categories":2848},[77],{"categories":2850},[77],{"categories":2852},[141],{"categories":2854},[118],{"categories":2856},[77],{"categories":2858},[162],{"categories":2860},[141],{"categories":2862},[436],{"categories":2864},[77],{"categories":2866},[],{"categories":2868},[],{"categories":2870},[436],{"categories":2872},[165],{"categories":2874},[123],{"categories":2876},[123],{"categories":2878},[141],{"categories":2880},[77],{"categories":2882},[115],{"categories":2884},[162],{"categories":2886},[123],{"categories":2888},[77],{"categories":2890},[179],{"categories":2892},[77],{"categories":2894},[123],{"categories":2896},[],{"categories":2898},[77],{"categories":2900},[77],{"categories":2902},[141],{"categories":2904},[115],{"categories":2906},[],{"categories":2908},[77],{"categories":2910},[77],{"categories":2912},[172],{"categories":2914},[162],{"categories":2916},[77,123],{"categories":2918},[179,118],{"categories":2920},[77],{"categories":2922},[],{"categories":2924},[123],{"categories":2926},[],{"categories":2928},[172],{"categories":2930},[77],{"categories":2932},[141],{"categories":2934},[],{"categories":2936},[123],{"categories":2938},[],{"categories":2940},[162],{"categories":2942},[123],{"categories":2944},[115],{"categories":2946},[123],{"categories":2948},[77],{"categories":2950},[436],{"categories":2952},[179],{"categories":2954},[118],{"categories":2956},[118],{"categories":2958},[115],{"categories":2960},[115],{"categories":2962},[77],{"categories":2964},[123],{"categories":2966},[77],{"categories":2968},[77],{"categories":2970},[115],{"categories":2972},[77],{"categories":2974},[179],{"categories":2976},[141],{"categories":2978},[77],{"categories":2980},[123],{"categories":2982},[77],{"categories":2984},[],{"categories":2986},[172],{"categories":2988},[],{"categories":2990},[123],{"categories":2992},[115],{"categories":2994},[],{"categories":2996},[436],{"categories":2998},[77],{"categories":3000},[],{"categories":3002},[141],{"categories":3004},[123],{"categories":3006},[172],{"categories":3008},[77],{"categories":3010},[123],{"categories":3012},[172],{"categories":3014},[123],{"categories":3016},[141],{"categories":3018},[115],{"categories":3020},[141],{"categories":3022},[172],{"categories":3024},[77],{"categories":3026},[162],{"categories":3028},[77],{"categories":3030},[77],{"categories":3032},[77],{"categories":3034},[77],{"categories":3036},[123],{"categories":3038},[77],{"categories":3040},[123],{"categories":3042},[77],{"categories":3044},[115],{"categories":3046},[77],{"categories":3048},[123],{"categories":3050},[162],{"categories":3052},[115],{"categories":3054},[123],{"categories":3056},[162],{"categories":3058},[],{"categories":3060},[77],{"categories":3062},[77],{"categories":3064},[172],{"categories":3066},[],{"categories":3068},[123],{"categories":3070},[179],{"categories":3072},[77],{"categories":3074},[141],{"categories":3076},[179],{"categories":3078},[123],{"categories":3080},[118],{"categories":3082},[118],{"categories":3084},[77],{"categories":3086},[115],{"categories":3088},[],{"categories":3090},[77],{"categories":3092},[],{"categories":3094},[115],{"categories":3096},[77],{"categories":3098},[123],{"categories":3100},[123],{"categories":3102},[],{"categories":3104},[172],{"categories":3106},[172],{"categories":3108},[179],{"categories":3110},[162],{"categories":3112},[],{"categories":3114},[77],{"categories":3116},[115],{"categories":3118},[77],{"categories":3120},[172],{"categories":3122},[115],{"categories":3124},[141],{"categories":3126},[141],{"categories":3128},[],{"categories":3130},[141],{"categories":3132},[123],{"categories":3134},[162],{"categories":3136},[165],{"categories":3138},[77],{"categories":3140},[],{"categories":3142},[141],{"categories":3144},[172],{"categories":3146},[118],{"categories":3148},[77],{"categories":3150},[115],{"categories":3152},[436],{"categories":3154},[115],{"categories":3156},[],{"categories":3158},[],{"categories":3160},[141],{"categories":3162},[],{"categories":3164},[123],{"categories":3166},[123],{"categories":3168},[123],{"categories":3170},[],{"categories":3172},[77],{"categories":3174},[],{"categories":3176},[141],{"categories":3178},[115],{"categories":3180},[162],{"categories":3182},[77],{"categories":3184},[141],{"categories":3186},[141],{"categories":3188},[],{"categories":3190},[141],{"categories":3192},[115],{"categories":3194},[77],{"categories":3196},[],{"categories":3198},[123],{"categories":3200},[123],{"categories":3202},[115],{"categories":3204},[],{"categories":3206},[],{"categories":3208},[],{"categories":3210},[162],{"categories":3212},[123],{"categories":3214},[77],{"categories":3216},[],{"categories":3218},[],{"categories":3220},[],{"categories":3222},[162],{"categories":3224},[],{"categories":3226},[115],{"categories":3228},[],{"categories":3230},[],{"categories":3232},[162],{"categories":3234},[77],{"categories":3236},[141],{"categories":3238},[],{"categories":3240},[179],{"categories":3242},[141],{"categories":3244},[179],{"categories":3246},[77],{"categories":3248},[],{"categories":3250},[],{"categories":3252},[123],{"categories":3254},[],{"categories":3256},[],{"categories":3258},[123],{"categories":3260},[77],{"categories":3262},[],{"categories":3264},[123],{"categories":3266},[141],{"categories":3268},[179],{"categories":3270},[165],{"categories":3272},[123],{"categories":3274},[123],{"categories":3276},[],{"categories":3278},[],{"categories":3280},[],{"categories":3282},[141],{"categories":3284},[],{"categories":3286},[],{"categories":3288},[162],{"categories":3290},[115],{"categories":3292},[],{"categories":3294},[118],{"categories":3296},[179],{"categories":3298},[77],{"categories":3300},[172],{"categories":3302},[115],{"categories":3304},[165],{"categories":3306},[118],{"categories":3308},[172],{"categories":3310},[],{"categories":3312},[],{"categories":3314},[123],{"categories":3316},[115],{"categories":3318},[162],{"categories":3320},[115],{"categories":3322},[123],{"categories":3324},[436],{"categories":3326},[123],{"categories":3328},[],{"categories":3330},[77],{"categories":3332},[141],{"categories":3334},[172],{"categories":3336},[],{"categories":3338},[162],{"categories":3340},[141],{"categories":3342},[115],{"categories":3344},[123],{"categories":3346},[77],{"categories":3348},[118],{"categories":3350},[123,436],{"categories":3352},[123],{"categories":3354},[172],{"categories":3356},[77],{"categories":3358},[165],{"categories":3360},[179],{"categories":3362},[123],{"categories":3364},[],{"categories":3366},[123],{"categories":3368},[77],{"categories":3370},[118],{"categories":3372},[],{"categories":3374},[],{"categories":3376},[77],{"categories":3378},[165],{"categories":3380},[77],{"categories":3382},[],{"categories":3384},[141],{"categories":3386},[],{"categories":3388},[141],{"categories":3390},[172],{"categories":3392},[123],{"categories":3394},[77],{"categories":3396},[179],{"categories":3398},[172],{"categories":3400},[],{"categories":3402},[141],{"categories":3404},[77],{"categories":3406},[],{"categories":3408},[77],{"categories":3410},[123],{"categories":3412},[77],{"categories":3414},[123],{"categories":3416},[77],{"categories":3418},[77],{"categories":3420},[77],{"categories":3422},[77],{"categories":3424},[118],{"categories":3426},[],{"categories":3428},[126],{"categories":3430},[141],{"categories":3432},[77],{"categories":3434},[],{"categories":3436},[172],{"categories":3438},[77],{"categories":3440},[77],{"categories":3442},[123],{"categories":3444},[141],{"categories":3446},[77],{"categories":3448},[77],{"categories":3450},[118],{"categories":3452},[123],{"categories":3454},[162],{"categories":3456},[],{"categories":3458},[165],{"categories":3460},[77],{"categories":3462},[],{"categories":3464},[141],{"categories":3466},[179],{"categories":3468},[],{"categories":3470},[],{"categories":3472},[141],{"categories":3474},[141],{"categories":3476},[179],{"categories":3478},[115],{"categories":3480},[123],{"categories":3482},[123],{"categories":3484},[77],{"categories":3486},[118],{"categories":3488},[],{"categories":3490},[],{"categories":3492},[141],{"categories":3494},[165],{"categories":3496},[172],{"categories":3498},[123],{"categories":3500},[162],{"categories":3502},[165],{"categories":3504},[165],{"categories":3506},[],{"categories":3508},[141],{"categories":3510},[77],{"categories":3512},[77],{"categories":3514},[172],{"categories":3516},[],{"categories":3518},[141],{"categories":3520},[141],{"categories":3522},[141],{"categories":3524},[],{"categories":3526},[123],{"categories":3528},[77],{"categories":3530},[],{"categories":3532},[115],{"categories":3534},[118],{"categories":3536},[],{"categories":3538},[77],{"categories":3540},[77],{"categories":3542},[],{"categories":3544},[172],{"categories":3546},[],{"categories":3548},[],{"categories":3550},[],{"categories":3552},[],{"categories":3554},[77],{"categories":3556},[141],{"categories":3558},[],{"categories":3560},[],{"categories":3562},[77],{"categories":3564},[77],{"categories":3566},[77],{"categories":3568},[165],{"categories":3570},[77],{"categories":3572},[165],{"categories":3574},[],{"categories":3576},[165],{"categories":3578},[165],{"categories":3580},[436],{"categories":3582},[123],{"categories":3584},[172],{"categories":3586},[],{"categories":3588},[],{"categories":3590},[165],{"categories":3592},[172],{"categories":3594},[172],{"categories":3596},[172],{"categories":3598},[],{"categories":3600},[115],{"categories":3602},[172],{"categories":3604},[172],{"categories":3606},[115],{"categories":3608},[172],{"categories":3610},[118],{"categories":3612},[172],{"categories":3614},[172],{"categories":3616},[172],{"categories":3618},[165],{"categories":3620},[141],{"categories":3622},[141],{"categories":3624},[77],{"categories":3626},[172],{"categories":3628},[165],{"categories":3630},[436],{"categories":3632},[165],{"categories":3634},[165],{"categories":3636},[165],{"categories":3638},[],{"categories":3640},[118],{"categories":3642},[],{"categories":3644},[436],{"categories":3646},[172],{"categories":3648},[172],{"categories":3650},[172],{"categories":3652},[123],{"categories":3654},[141,118],{"categories":3656},[165],{"categories":3658},[],{"categories":3660},[],{"categories":3662},[165],{"categories":3664},[],{"categories":3666},[165],{"categories":3668},[141],{"categories":3670},[123],{"categories":3672},[],{"categories":3674},[172],{"categories":3676},[77],{"categories":3678},[162],{"categories":3680},[],{"categories":3682},[77],{"categories":3684},[],{"categories":3686},[141],{"categories":3688},[115],{"categories":3690},[165],{"categories":3692},[],{"categories":3694},[172],{"categories":3696},[141],[3698,3865,3997,4087],{"id":3699,"title":3700,"ai":3701,"body":3706,"categories":3843,"created_at":78,"date_modified":78,"description":69,"extension":79,"faq":78,"featured":80,"kicker_label":78,"meta":3844,"navigation":93,"path":3852,"published_at":3853,"question":78,"scraped_at":3854,"seo":3855,"sitemap":3856,"source_id":3857,"source_name":3858,"source_type":101,"source_url":3859,"stem":3860,"tags":3861,"thumbnail_url":78,"tldr":3862,"tweet":78,"unknown_tags":3863,"__hash__":3864},"summaries\u002Fsummaries\u002Ffe837463ae43e90d-context-engines-fix-agent-context-to-cut-tokens-50-summary.md","Context Engines: Fix Agent Context to Cut Tokens 50%",{"provider":7,"model":8,"input_tokens":3702,"output_tokens":3703,"processing_time_ms":3704,"cost_usd":3705},8656,2106,27045,0.00276155,{"type":14,"value":3707,"toc":3836},[3708,3712,3715,3718,3721,3724,3728,3731,3734,3737,3740,3744,3747,3750,3778,3781,3784,3788,3791,3794,3797,3800,3803,3806,3810],[17,3709,3711],{"id":3710},"agents-need-org-context-to-avoid-doom-loops","Agents Need Org Context to Avoid Doom Loops",[22,3713,3714],{},"Peter Werry from Unblocked explains that AI agents start at \"ground zero\" with no knowledge of your codebase, conventions, or decisions. Without targeted context, they rip through repos inefficiently, leading to wrong code, long reviews, and \"doom loops\"—endless iterations fixing misguided outputs. Humans used to be the context engine, manually feeding tickets and correcting agents, but scaling to background agents demands automation.",[22,3716,3717],{},"The goal mirrors human onboarding: accumulate \"battle scars\" from incidents, mentors, and history to know \"why\" things are built a certain way. Werry references an adoption curve (lifted from Vimath) showing progression from autocomplete (2022, 8k tokens) to parallel agents with MCP servers, toward cloud-based YOLO agents. Humans become the bottleneck in context-switching; context engines enable agentic freedom by supplying \"all the context you need and most importantly none you don't\" in optimized form.",[22,3719,3720],{},"\"Context engineering is kind of the art of supplying uh all the context that you need and most importantly none of the context that you don't need in a highly optimized way so that when the agent starts to run it executes the task uh in a streamlined way that's in line with your organization's best practices.\" (Peter Werry defining context engines—core to avoiding waste.)",[22,3722,3723],{},"A demo contrasts MCP-only vs. context engine: without it, an agent missed legacy Anthropic token budget deps, deleting code; with it, the agent preserved backwards compatibility after reasoning over history.",[17,3725,3727],{"id":3726},"myths-rag-mcp-and-big-windows-fall-short","Myths: RAG, MCP, and Big Windows Fall Short",[22,3729,3730],{},"Naive RAG over docs causes \"satisfaction of search\"—agents grab the first plausible match (e.g., from Notion) and stop, missing golden nuggets in Slack incidents or deleted PRs. In large orgs, it pulls irrelevant code, creates conflicts, wastes tokens on compaction, and ignores tasks\u002Fpersonalization.",[22,3732,3733],{},"Connecting MCP servers gives access but no relationships or \"why.\" Bigger windows (e.g., Gemini's 1M tokens) excel at needle-in-haystack but fail reasoning across sources or truth selection. Most orgs exceed 1M tokens anyway; even 50M won't resolve conflicts or focus retrieval.",[22,3735,3736],{},"\"Access doesn't equal understanding.\" (Werry on why MCP\u002FRAG alone leads to doom loops—emphasizes need for reasoning layer.)",[22,3738,3739],{},"The iceberg below compiling code: user intent, past rejections, failures, deletions. Agents need history to infer absences.",[17,3741,3743],{"id":3742},"building-blocks-graphs-resolution-personalization","Building Blocks: Graphs, Resolution, Personalization",[22,3745,3746],{},"Inputs: planning tools, docs, Slack\u002FTeams, code, PRs. Outputs: agents, CLI, code review in SCM, messaging.",[22,3748,3749],{},"Key requirements:",[3751,3752,3753,3760,3766,3772],"ul",{},[3754,3755,3756,3759],"li",{},[26,3757,3758],{},"Unified relationships via social engineering graph",": Link data deeply—e.g., distill repeated PR comments into \"memories\" of best practices. Easy: PR-Slack links. Hard: infer decisions from patterns\u002Fincidents. (Workshop builds this graph.)",[3754,3761,3762,3765],{},[26,3763,3764],{},"Conflict resolution",": Reject naive recency (docs\u002Fchats lag code). Evolved to: main branch as truth + expert convos (forward-looking) + history (what not to repeat). Surface unresolvable conflicts for human input.",[3754,3767,3768,3771],{},[26,3769,3770],{},"Permissions",": Flow access controls—e.g., Slack private channels only for authorized users; answers stay private.",[3754,3773,3774,3777],{},[26,3775,3776],{},"Targeted retrieval",": Bias to user's repos (via PR counts): deep vector search there, shallow elsewhere. Task-focused to save tokens.",[22,3779,3780],{},"High-level flow: Data sources → graph\u002Freasoning → personalized context → agents.",[22,3782,3783],{},"\"Satisfaction of search is a term that actually comes out of uh the medical field in radiology... they might find something... and then they stop.\" (Werry borrowing radiology concept—explains why agents halt prematurely, missing key org history.)",[17,3785,3787],{"id":3786},"production-lessons-what-unblocked-got-wrong","Production Lessons: What Unblocked Got Wrong",[22,3789,3790],{},"Initially optimized for access: knowledge graph + retrieval tools. Failed—agents couldn't traverse meaningfully.",[22,3792,3793],{},"Hid conflicts by forcing naive resolution (recency\u002Fcode bias). Better: surface them to learn from feedback.",[22,3795,3796],{},"Don't cache answers—code\u002Fdocs change constantly; prior answers regress to mean, polluting context.",[22,3798,3799],{},"Experiment: Larger task without engine repeated past failures (missed implementation details); with it, nailed it. Time: 2.5 hours → 25 minutes. Tokens: 21M → 10M. (Claude-estimated; vibe is dramatic efficiency.)",[22,3801,3802],{},"Use in planning (biggest wins via MCP skill) and review (understands motivations beyond code).",[22,3804,3805],{},"\"Initially we optimized for access not understanding... that does not work.\" (Werry on first failure—shift to reasoning was key pivot.)",[17,3807,3809],{"id":3808},"key-takeaways","Key Takeaways",[3751,3811,3812,3815,3818,3821,3824,3827,3830,3833],{},[3754,3813,3814],{},"Build social graphs to link code\u002FPRs\u002FSlack into memories of decisions\u002Fbest practices, not just surface links.",[3754,3816,3817],{},"Resolve conflicts with multi-signal scoring (recency + code truth + expert forward-look); surface irresolvables.",[3754,3819,3820],{},"Personalize via user PR history: deep retrieval on their repos, shallow elsewhere.",[3754,3822,3823],{},"Enforce permissions end-to-end—e.g., private Slack only for access holders.",[3754,3825,3826],{},"Integrate early in agent flows (planning\u002Freview) via MCP skills for 50%+ token\u002Ftime savings.",[3754,3828,3829],{},"Avoid caching answers or feeding prior outputs—dynamic orgs demand fresh retrieval.",[3754,3831,3832],{},"Target 'why' via history\u002Fincidents, beating RAG's 'what' for production code.",[3754,3834,3835],{},"Prototype with graphs before full engine; workshop-style builds reveal gaps fast.",{"title":69,"searchDepth":70,"depth":70,"links":3837},[3838,3839,3840,3841,3842],{"id":3710,"depth":70,"text":3711},{"id":3726,"depth":70,"text":3727},{"id":3742,"depth":70,"text":3743},{"id":3786,"depth":70,"text":3787},{"id":3808,"depth":70,"text":3809},[77],{"content_references":3845,"triage":3850},[3846],{"type":3847,"title":3848,"author":3849,"context":87},"other","Adoption curve","Vimath",{"relevance":89,"novelty":90,"quality":90,"actionability":90,"composite":91,"reasoning":3851},"Category: AI & LLMs. The article discusses the importance of context in AI agents, addressing a specific pain point for developers integrating AI into their workflows. It provides actionable insights on building a reasoning layer to optimize agent performance, which is directly applicable to product builders.","\u002Fsummaries\u002Ffe837463ae43e90d-context-engines-fix-agent-context-to-cut-tokens-50-summary","2026-05-03 16:00:06","2026-05-04 16:07:41",{"title":3700,"description":69},{"loc":3852},"b0a932deb93b0851","AI Engineer","https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=5ID22ACI7IM","summaries\u002Ffe837463ae43e90d-context-engines-fix-agent-context-to-cut-tokens-50-summary",[106,105,107,108],"Agents fail without org-specific context; build a reasoning layer that personalizes retrieval, resolves conflicts, and respects permissions to deliver task-focused info, reducing task time from 2.5hrs\u002F21M tokens to 25min\u002F10M.",[107,108],"LZtC3JN5Z5WtQmxN4NyOtG03I0YxgtekncwOC8lwGuY",{"id":3866,"title":3867,"ai":3868,"body":3873,"categories":3963,"created_at":78,"date_modified":78,"description":69,"extension":79,"faq":78,"featured":80,"kicker_label":78,"meta":3964,"navigation":93,"path":3985,"published_at":78,"question":78,"scraped_at":3986,"seo":3987,"sitemap":3988,"source_id":3989,"source_name":3990,"source_type":101,"source_url":3991,"stem":3992,"tags":3993,"thumbnail_url":78,"tldr":3994,"tweet":78,"unknown_tags":3995,"__hash__":3996},"summaries\u002Fsummaries\u002F176b270d6050d162-ai-agents-speed-up-gpu-kernels-1-81x-with-scaffold-summary.md","AI Agents Speed Up GPU Kernels 1.81x with Scaffolding",{"provider":7,"model":8,"input_tokens":3869,"output_tokens":3870,"processing_time_ms":3871,"cost_usd":3872},8277,2360,15037,0.00281275,{"type":14,"value":3874,"toc":3956},[3875,3879,3882,3885,3889,3892,3902,3905,3909,3912,3915,3919,3922,3925,3927],[17,3876,3878],{"id":3877},"benchmark-refinements-unlock-realistic-kernel-optimization-measurement","Benchmark Refinements Unlock Realistic Kernel Optimization Measurement",[22,3880,3881],{},"METR addressed KernelBench's limitations—naive PyTorch baselines, outdated tasks from mid-2010s architectures, and noisy measurements—by filtering 45 low-quality tasks and adding Level 5 with 14 frontier workloads (DeepSeek-V3, Llama 3, Hunyuan Video, State Space Models like Mamba2). Filters targeted issues like low signal-to-noise outputs (-0.01 to 0.01 range), uniform tensors, constant functions (e.g., mean(softmax(x))), and seed-independent computations. They modified tasks with randomized weights or inputs to prevent caching cheats and switched to triton.testing.do_bench for reliable timings. Excluded Level 4 due to HuggingFace dependencies requiring codebase parsing. Result: 225 tasks across Levels 1-5, all single-H100 inference, emphasizing fusions like attention and convolutions in higher levels.",[22,3883,3884],{},"This setup better mirrors small research teams' needs, where naive PyTorch is common, but highlights gaps like no multi-GPU comms, fixed shapes, or training backprop. \"KernelBench initial implementations are naive PyTorch, rather than highly optimized CUDA. This makes KernelBench much easier, more representative of small research groups, and not representative of large scale applications.\"",[17,3886,3888],{"id":3887},"kernelagents-parallel-search-drives-cost-effective-speedups","KernelAgent's Parallel Search Drives Cost-Effective Speedups",[22,3890,3891],{},"KernelAgent uses parallel tree search: 8 initial attempts (PyTorch\u002FTriton\u002FCUDA), resampling from best-so-far (weighted by speedup) for 300 total attempts\u002Fproblem (40 for costly o1). Includes Triton\u002FCUDA docs, ground-truth verification. Best-of-k across GPT-4o, Claude 3.5 Sonnet, o1 yields 1.81x geometric mean speedup vs. original KernelBench's 1.05x—15x better—attributed to scaffolding, prompt tuning, $20\u002Ftask spend (vs. \u003C$1 originally). o3-mini-high hits 1.81x alone; best across all models: 2.01x, doubling in 6 months.",[22,3893,3894,3895,3901],{},"Per-cost\u002Fper-attempt: o3-mini excels (power-law gains to 300 attempts), breakeven at 27 H100-hours\u002Fworkload. Kernels: 80% Triton\u002FCUDA (outperform PyTorch), Level 5 averages 500 LOC. Fine-tuning GPT-4o on 1200 solutions nearly matches o3-mini on Level 5 (except DeepSeek MOE). Released best solutions: ",[3896,3897,3898],"a",{"href":3898,"rel":3899},"https:\u002F\u002Fgithub.com\u002FMETR\u002FKernelBenchFiltered",[3900],"nofollow",".",[22,3903,3904],{},"\"Using our KernelAgent, o3-mini-high sped up code by 1.81x, and taking the best of all models per problem using KernelAgent achieved 2.01x speedup, representing a large increase from models released only 3 months earlier.\"",[17,3906,3908],{"id":3907},"tradeoffs-strong-on-niche-tasks-far-from-frontier-expertise","Tradeoffs: Strong on Niche Tasks, Far from Frontier Expertise",[22,3910,3911],{},"Agents succeed on 69-95% tasks (>2% speedup), 93% >5% with best-of-all, peaks at 30x. But vs. humans\u002Ffrontier labs: 4 engineer-weeks vs. 5 engineer-years\u002Fmodel; agents can't adapt FlashAttention-2 or match expert novelty (token limits, no context factoring). Torch.compile wrappers lag (best-of-k flags \u003C1.81x). Economic: Models beat humans for \u003C$500 niches (hundreds compute dollars), not billions-scale.",[22,3913,3914],{},"Limitations persist: no human baselines, inference-only, 0.01 tolerance (vs. end-to-end stability), flawed tasks remain. Original KernelBench elicited fraction of capabilities sans scaffolding. \"Our results do not imply that current LM agents can automate kernel engineering... kernels produced by our agents are never as novel and sophisticated as the best open source kernels produced by top experts.\"",[17,3916,3918],{"id":3917},"broader-implications-for-ai-rd-automation-risks","Broader Implications for AI R&D Automation Risks",[22,3920,3921],{},"1.81-2.01x speedups show agents automate kernel bottlenecks (hundreds millions $\u002Fyear savings), accelerating non-frontier ML (open-source research). Withheld agent code due to proliferation; sharing emphasizes evaluation rigor for safety policy. Capabilities advance fast; proper elicitation critical. \"We believe our results highlight some challenges of doing evaluations, especially the importance of proper capability elicitation, commensurate to the economic value that models can actually provide.\"",[22,3923,3924],{},"\"Model written kernels could fill the underserved niche of accelerating machine learning projects that use only hundreds of dollars of compute.\"",[17,3926,3809],{"id":3808},[3751,3928,3929,3932,3935,3938,3941,3944,3947,3950,3953],{},[3754,3930,3931],{},"Filter benchmarks ruthlessly: Remove low-SNR, cheat-prone tasks (e.g., constants, uniform outputs) to isolate real optimization signal.",[3754,3933,3934],{},"Invest in scaffolding: Parallel tree search + high compute ($20\u002Ftask) unlocks 15x better elicitation than zero-shot.",[3754,3936,3937],{},"Scale attempts power-law: o3-mini gains to 300+; breakeven at 27 H100-hours for PyTorch models.",[3754,3939,3940],{},"Prioritize Triton\u002FCUDA: Agents favor them (80%), outperforming PyTorch for complex fusions.",[3754,3942,3943],{},"Benchmark small-team realities: Naive PyTorch + single GPU reveals niche where AI > humans economically.",[3754,3945,3946],{},"Add frontier tasks: Level 5 (DeepSeek-V3 etc.) stresses agents; fine-tuning closes gaps fast.",[3754,3948,3949],{},"Context matters: 1.81x vs. 1.05x shows eval pitfalls; always compare to baselines like torch.compile.",[3754,3951,3952],{},"Safety via sharing: Measure R&D automation transparently for policy, despite capability risks.",[3754,3954,3955],{},"Humans still win frontiers: Agents lag expert novelty by orders of magnitude in effort\u002Fquality.",{"title":69,"searchDepth":70,"depth":70,"links":3957},[3958,3959,3960,3961,3962],{"id":3877,"depth":70,"text":3878},{"id":3887,"depth":70,"text":3888},{"id":3907,"depth":70,"text":3908},{"id":3917,"depth":70,"text":3918},{"id":3808,"depth":70,"text":3809},[77,172],{"content_references":3965,"triage":3983},[3966,3971,3974,3977,3980],{"type":3967,"title":3968,"url":3969,"context":3970},"paper","DeepSeek-V3","https:\u002F\u002Farxiv.org\u002Fhtml\u002F2412.19437v1","cited",{"type":84,"title":3972,"url":3973,"context":87},"KernelBench","https:\u002F\u002Fscalingintelligence.stanford.edu\u002Fblogs\u002Fkernelbench\u002F",{"type":84,"title":3975,"url":3976,"context":87},"FlashAttention 2","https:\u002F\u002Fgithub.com\u002FDao-AILab\u002Fflash-attention",{"type":3847,"title":3978,"url":3979,"context":3970},"Thoughts on sharing information about language model","https:\u002F\u002Fwww.alignmentforum.org\u002Fposts\u002FfRSj2W4Fjje8rQWm9\u002Fthoughts-on-sharing-information-about-language-model",{"type":84,"title":3981,"url":3898,"context":3982},"KernelBenchFiltered","recommended",{"relevance":89,"novelty":90,"quality":90,"actionability":90,"composite":91,"reasoning":3984},"Category: AI Automation. The article provides a detailed analysis of how AI agents can optimize GPU kernels, addressing a specific pain point for developers looking to improve performance in AI applications. It offers actionable insights on using KernelAgent for practical speedups, making it relevant for product builders.","\u002Fsummaries\u002F176b270d6050d162-ai-agents-speed-up-gpu-kernels-1-81x-with-scaffold-summary","2026-04-15 15:30:30",{"title":3867,"description":69},{"loc":3985},"176b270d6050d162","__oneoff__","https:\u002F\u002Fmetr.org\u002Fblog\u002F2025-02-14-measuring-automated-kernel-engineering\u002F","summaries\u002F176b270d6050d162-ai-agents-speed-up-gpu-kernels-1-81x-with-scaffold-summary",[105,106,107,108],"METR's KernelAgent, using o3-mini and others, achieves 1.81x average speedup on filtered KernelBench tasks via parallel tree search and high test-time compute, costing ~$20\u002Ftask—far below human engineers for small ML projects.",[107,108],"EfpvZbC0IB4JJA4YtygyC1UVDynNSAkAV5xl9_YPWfs",{"id":3998,"title":3999,"ai":4000,"body":4005,"categories":4058,"created_at":78,"date_modified":78,"description":69,"extension":79,"faq":78,"featured":80,"kicker_label":78,"meta":4059,"navigation":93,"path":4075,"published_at":4076,"question":78,"scraped_at":4077,"seo":4078,"sitemap":4079,"source_id":4080,"source_name":3858,"source_type":101,"source_url":4081,"stem":4082,"tags":4083,"thumbnail_url":78,"tldr":4084,"tweet":78,"unknown_tags":4085,"__hash__":4086},"summaries\u002Fsummaries\u002F8d4ddaba94ddcd2c-code-mode-ai-agents-generate-executable-js-over-js-summary.md","Code Mode: AI Agents Generate Executable JS Over JSON Tools",{"provider":7,"model":8,"input_tokens":4001,"output_tokens":4002,"processing_time_ms":4003,"cost_usd":4004},7705,1763,10403,0.00213165,{"type":14,"value":4006,"toc":4052},[4007,4011,4014,4025,4029,4032,4036,4039,4045,4049],[17,4008,4010],{"id":4009},"scale-ai-agents-with-code-generation-not-json-tool-calls","Scale AI Agents with Code Generation, Not JSON Tool Calls",[22,4012,4013],{},"Traditional tool calling fails at scale: stuffing hundreds of tools (e.g., Google services, Jira, wiki) into context causes breakage, poor composition, and slow back-and-forth rounds. For Cloudflare's 2,600 API endpoints, exposing each as a tool requires 1.2 million tokens in the first call—impossible. Instead, use \"Code Mode\": prompt LLMs to output executable JavaScript that runs in one shot against an environment. Benefits include typed APIs with syntax\u002Ftype checking (leveraging models' training on terabytes of code), state holding, looping, sequencing, and parallelization—capabilities JSON lacks.",[22,4015,4016,4017,4020,4021,4024],{},"Expose minimal tools like ",[42,4018,4019],{},"search"," (searches full OpenAPI JSON spec) and ",[42,4022,4023],{},"execute"," (runs code against discovered endpoints). This shrinks 1.2M tokens to ~1,000 (99.9% reduction). Example prompt: \"We're getting DDoS'd—find and block offending IPs.\" Model generates JS to search APIs, paginate workers, identify attackers, and block them in one execution, avoiding 8 round trips. Live demo on Mythical server lists Workers via read-only access, searching endpoints and paginating results.",[17,4026,4028],{"id":4027},"emergent-behaviors-from-inhabiting-system-state","Emergent Behaviors from Inhabiting System State",[22,4030,4031],{},"Code Mode lets agents \"inhabit the state machine\" rather than generate separate programs. In Kenton’s (Cloudflare Workers creator) canvas (TLDraw\u002FExcalidraw-style), user draws tic-tac-toe board. Agent inspects raw stroke array (grid lines + X), recognizes game state, and draws O in center—no tic-tac-toe code exists. Reasoning traces show Opus let user win, hinting at alignment quirks. This breaks programmer\u002Fnon-programmer divide: anyone prompts \"rename 200 photos by date\u002Flocation,\" agent generates\u002Fruns script using vision models—bypassing clunky $7\u002Fmonth apps.",[17,4033,4035],{"id":4034},"build-harnesses-capability-based-sandboxes-for-safe-execution","Build Harnesses: Capability-Based Sandboxes for Safe Execution",[22,4037,4038],{},"Core architecture is the \"Harness\": fast-starting sandbox (e.g., V8 isolates for 10 years of security hardening, quick spin-up; alternatives: WASM, custom JS interpreters) with zero initial capabilities (no fetches, APIs). Explicitly grant APIs as functions, control all outgoing network (prefer no fetches for speed\u002Fobservability). Run ephemerally near APIs\u002Fuser (e.g., iPhone for mashing services) or backend.",[22,4040,4041,4042,3901],{},"Enables long-running workflows (days\u002Fmonths\u002Fyears with persistent state), generative UIs tailored per user (e.g., e-commerce: custom return shoes\u002Ffind similar \u003C$100 or track delayed order interfaces from user context\u002Fcart). Design for agent DX: Markdown docs, searchable errors, discoverability. Revives capability-based security: start empty, add powers explicitly. UI shifts—React devs excel as they're user-proximate; rethink 30-year tech tree now with safe ",[42,4043,4044],{},"eval",[17,4046,4048],{"id":4047},"new-era-code-as-universal-interface-for-agents","New Era: Code as Universal Interface for Agents",[22,4050,4051],{},"Programmers always used code for power; others got buttons\u002Fforms. LLMs democratize this—agents are next billion users, dreaming in types\u002Fsyntax. Build systems exposing code-friendly APIs (not just human UIs). Harness runs general-purpose computing cheaply (no $400 Mac Mini for API calls). Outcomes: custom per-user programs, stitched services, full observability (trace why agent traded $2.3M on \"llama poop\").",{"title":69,"searchDepth":70,"depth":70,"links":4053},[4054,4055,4056,4057],{"id":4009,"depth":70,"text":4010},{"id":4027,"depth":70,"text":4028},{"id":4034,"depth":70,"text":4035},{"id":4047,"depth":70,"text":4048},[],{"content_references":4060,"triage":4073},[4061,4065,4067,4070],{"type":84,"title":4062,"author":4063,"url":4064,"context":87},"PartyKit","Sunil Pai","https:\u002F\u002Fsunilpai.dev\u002F",{"type":84,"title":4066,"context":87},"Mythical Server",{"type":3847,"title":4068,"author":4069,"context":87},"Cloudflare Agents SDK","Sunil Pai (Cloudflare)",{"type":3847,"title":4071,"author":4072,"context":87},"Dynamic Workers","Cloudflare",{"relevance":89,"novelty":90,"quality":90,"actionability":90,"composite":91,"reasoning":4074},"Category: AI Automation. The article discusses a novel approach to using AI agents for code generation instead of traditional JSON tool calls, addressing a specific pain point of scaling API interactions effectively. It provides concrete examples of how to implement this approach, making it actionable for developers looking to optimize their workflows.","\u002Fsummaries\u002F8d4ddaba94ddcd2c-code-mode-ai-agents-generate-executable-js-over-js-summary","2026-04-19 17:19:35","2026-04-21 15:13:22",{"title":3999,"description":69},{"loc":4075},"534bb415205efb9b","https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=8txf05vVVl4","summaries\u002F8d4ddaba94ddcd2c-code-mode-ai-agents-generate-executable-js-over-js-summary",[106,105,107,108],"Replace JSON tool calling with AI-generated JavaScript code execution in sandboxes to handle massive APIs (e.g., Cloudflare's 2600 endpoints, 1.2M tokens reduced to 1K), enable stateful loops\u002Fparallelism, and unlock emergent behaviors like inspecting canvas strokes for tic-tac-toe.",[107,108],"wMXZdezHnVCsIFJHpj6gXjDDI_lLycof6RjmorTHeEI",{"id":4088,"title":4089,"ai":4090,"body":4095,"categories":4139,"created_at":78,"date_modified":78,"description":69,"extension":79,"faq":78,"featured":80,"kicker_label":78,"meta":4140,"navigation":93,"path":4153,"published_at":78,"question":78,"scraped_at":4154,"seo":4155,"sitemap":4156,"source_id":4157,"source_name":3990,"source_type":101,"source_url":4158,"stem":4159,"tags":4160,"thumbnail_url":78,"tldr":4161,"tweet":78,"unknown_tags":4162,"__hash__":4163},"summaries\u002Fsummaries\u002F1eba12a4a8fa384c-parallel-claude-agents-build-linux-compiling-c-com-summary.md","Parallel Claude Agents Build Linux-Compiling C Compiler",{"provider":7,"model":8,"input_tokens":4091,"output_tokens":4092,"processing_time_ms":4093,"cost_usd":4094},6703,1741,10168,0.00170425,{"type":14,"value":4096,"toc":4134},[4097,4101,4112,4116,4127,4131],[17,4098,4100],{"id":4099},"agent-team-harness-unlocks-autonomous-long-running-development","Agent Team Harness Unlocks Autonomous Long-Running Development",[22,4102,4103,4104,4107,4108,4111],{},"Run multiple Claude instances in parallel Docker containers on a shared git repo to tackle complex projects without human input. Each agent loops indefinitely via a bash script: clone repo to workspace, claim tasks by creating lock files in ",[42,4105,4106],{},"current_tasks\u002F"," (e.g., ",[42,4109,4110],{},"parse_if_statement.txt","), work on it, merge upstream changes, push, and release lock. Git handles conflicts; agents self-specialize—some fix bugs, others dedupe code, optimize performance, critique Rust design, or update docs. This parallelism fixes single-agent limits: sequential tasking and lack of specialization. For giant tasks like Linux kernel compilation, use GCC as oracle—randomly compile subsets with Claude's compiler vs. GCC; agents parallelize bug fixes across files, applying delta debugging for interacting failures.",[17,4113,4115],{"id":4114},"tailor-tests-and-environment-to-llm-constraints","Tailor Tests and Environment to LLM Constraints",[22,4117,4118,4119,4122,4123,4126],{},"Design verifiers that are nearly perfect since agents chase whatever passes tests. Source high-quality suites (e.g., GCC torture tests at 99% pass rate), add CI to block regressions, and build scripts for benchmarks like SQLite, Redis, libjpeg, QuickJS, Lua, QEMU, FFmpeg, Postgres. Counter context pollution: limit test output to few lines with ",[42,4120,4121],{},"ERROR: reason"," format for grep, log details to files, precompute stats. Address time blindness: default ",[42,4124,4125],{},"--fast"," mode samples 1-10% of tests deterministically per agent for quick regressions. Maintain READMEs\u002Fprogress files for orientation in fresh containers. Result: agents sustain progress, hitting 99% test pass rates and compiling real projects.",[17,4128,4130],{"id":4129},"benchmarks-push-llms-to-limits-expose-gaps","Benchmarks Push LLMs to Limits, Expose Gaps",[22,4132,4133],{},"Opus 4.6 crossed thresholds prior models couldn't: clean-room 100k-line Rust compiler (std lib only), SSA IR for optimizations, GCC-compatible, compiles bootable Linux 6.9 (x86\u002FARM\u002FRISC-V, except 16-bit x86 real mode via GCC cheat), QEMU\u002FFFmpeg\u002FSQLite\u002FPostgres\u002FRedis, runs Doom. Cost: 2B input\u002F140M output tokens over 2 weeks. Limits: no full assembler\u002Flinker (uses GCC), inefficient code (worse than -O0 GCC), mediocre Rust quality, incomplete project support. New features often regress; 16-bit x86 gen bloats past 32k limit. Study repo to see breakdowns—pushes like this benchmark future capabilities.",{"title":69,"searchDepth":70,"depth":70,"links":4135},[4136,4137,4138],{"id":4099,"depth":70,"text":4100},{"id":4114,"depth":70,"text":4115},{"id":4129,"depth":70,"text":4130},[],{"content_references":4141,"triage":4151},[4142,4145,4148],{"type":3847,"title":4143,"url":4144,"context":87},"Claude's C Compiler","https:\u002F\u002Fgithub.com\u002Fanthropics\u002Fclaudes-c-compiler",{"type":84,"title":4146,"url":4147,"context":87},"GCC","https:\u002F\u002Fgcc.gnu.org\u002F",{"type":3847,"title":4149,"url":4150,"context":87},"GCC torture test suite","https:\u002F\u002Fgcc.gnu.org\u002Fonlinedocs\u002Fgccint\u002FTorture-Tests.html",{"relevance":89,"novelty":90,"quality":90,"actionability":90,"composite":91,"reasoning":4152},"Category: AI Automation. The article provides a detailed account of using parallel AI agents to autonomously build a complex software project, addressing practical applications of AI in software engineering. It outlines specific methodologies for agent collaboration and testing, which can be directly applied by developers looking to implement similar systems.","\u002Fsummaries\u002F1eba12a4a8fa384c-parallel-claude-agents-build-linux-compiling-c-com-summary","2026-04-15 15:30:27",{"title":4089,"description":69},{"loc":4153},"1eba12a4a8fa384c","https:\u002F\u002Fwww.anthropic.com\u002Fengineering\u002Fbuilding-c-compiler","summaries\u002F1eba12a4a8fa384c-parallel-claude-agents-build-linux-compiling-c-com-summary",[106,105,107,108],"16 Opus 4.6 agents in parallel autonomously produced a 100k-line Rust C compiler that builds Linux 6.9 on x86\u002FARM\u002FRISC-V after 2,000 sessions and $20k API cost, revealing harness designs for long-running LLM teams.",[107,108],"ZQZym3mE8YCtCAFuNWCTyAQB52CjkFgTp7Yon0PcdzY"]