https://ai-coding.medium.com/a-lot-of-local-llm-discussions-focus-on-model-size-but-the-real-bottleneck-is-often-the-path-4a574b51a5312026-09-25monthly0.5 https://ai-coding.medium.com/the-interesting-architectural-shift-is-separating-thinking-from-deciding-77b07973d5622026-09-25monthly0.5 https://ai-coding.medium.com/error-recovery-feels-worth-separating-out-bd20e82b89fe2026-09-24monthly0.2 https://ai-coding.medium.com/phia-affiliate-attribution-hidden-click-e9d369e2e53d2026-09-24monthly0.5 https://ai-coding.medium.com/the-pod-restart-result-made-me-slow-down-a-bit-04487f53110c2026-09-24monthly0.2 https://ai-coding.medium.com/jev-fast-adoption-automation-workflow-5ec342497fef2026-09-24monthly0.5 https://ai-coding.medium.com/im-curious-how-much-extra-state-shaping-jev-needs-in-a-real-agent-loop-before-some-of-that-speed-ee705b4d7dde2026-09-24monthly0.5 https://ai-coding.medium.com/the-sham-relief-condition-is-doing-a-lot-of-work-here-4c1d3180012e2026-09-24monthly0.5 https://ai-coding.medium.com/stopping-an-instance-only-helps-if-you-know-which-instance-is-acting-and-what-credentials-it-still-dfd82c00adfd2026-09-23monthly0.5 https://ai-coding.medium.com/the-failed-surface-level-reproduction-may-be-the-most-useful-part-here-ae94721f8d9c2026-09-23monthly0.5 https://ai-coding.medium.com/runs-as-the-user-sounds-reassuring-until-you-actually-try-the-same-question-with-two-users-who-27973e621f722026-09-22monthly0.5 https://ai-coding.medium.com/ai-api-docs-first-request-failure-bc8b02ab81862026-09-21monthly0.5 https://ai-coding.medium.com/the-maintenance-part-feels-easy-to-underestimate-db8baf4253182026-09-21monthly0.5 https://ai-coding.medium.com/a-calibrated-component-can-still-sit-inside-a-badly-calibrated-workflow-92f3a379be7c2026-09-20monthly0.5 https://ai-coding.medium.com/explicit-memory-writes-make-the-system-easier-to-trust-8040428fd8192026-09-20monthly0.5 https://ai-coding.medium.com/the-part-about-monitoring-getting-weaker-once-the-reasoning-itself-is-under-supervision-was-the-38ee43a947d02026-09-18monthly0.2 https://ai-coding.medium.com/i-liked-the-separation-between-imagined-and-observed-outcomes-507a0ba381fe2026-09-18monthly0.5 https://ai-coding.medium.com/the-part-id-watch-is-skill-selection-09b43c23d6522026-09-17monthly0.5 https://ai-coding.medium.com/routing-gets-harder-once-the-choice-isnt-just-tool-a-vs-004fb523efae2026-09-17monthly0.5 https://ai-coding.medium.com/splitting-who-is-viable-from-how-much-and-at-what-price-is-a-clean-boundary-47e764ba12982026-09-17monthly0.5 https://ai-coding.medium.com/i-like-that-the-principles-are-visible-enough-to-argue-with-8442f9b79e982026-09-17monthly0.5 https://ai-coding.medium.com/when-you-correct-ai-memory-what-actually-changes-2a572f93d38b2026-10-10daily0.2 https://ai-coding.medium.com/id-want-the-run-state-to-make-the-handoff-visible-too-dd70a5ae9ef42026-09-16monthly0.2 https://ai-coding.medium.com/the-app-only-tool-visibility-is-probably-the-part-id-watch-most-closely-74399ed0443c2026-09-16monthly0.5 https://ai-coding.medium.com/your-backup-ai-model-may-break-the-workflow-56a79a082a082026-09-16monthly0.5 https://ai-coding.medium.com/solving-the-problem-before-seeing-the-diff-gives-the-reviewer-something-independent-to-compare-646aa954d7362026-09-16monthly0.5 https://ai-coding.medium.com/the-step-3-failure-is-actually-the-useful-part-of-this-demo-4994cfc8ca5e2026-09-16monthly0.5 https://ai-coding.medium.com/narrow-tools-with-boring-contracts-are-probably-easier-to-trust-at-scale-7726865b5ec12026-09-16monthly0.2 https://ai-coding.medium.com/a-failed-run-is-usually-more-informative-than-a-benchmark-bump-46bb514099552026-09-16monthly0.5 https://ai-coding.medium.com/a-12-hour-agent-run-is-a-very-different-problem-from-a-good-single-response-187069acd4dc2026-09-16monthly0.7 https://ai-coding.medium.com/the-ugly-part-is-when-both-databases-are-healthy-and-still-disagree-about-the-same-user-e4a47490cbae2026-10-15daily0.2 https://ai-coding.medium.com/four-manually-checked-rows-finding-a-wrong-minute-a-nonexistent-var-review-and-a-merged-match-is-f98ef510cd722026-10-15daily0.5 https://ai-coding.medium.com/ten-input-failures-and-only-one-fixed-by-higher-effort-is-a-useful-ratio-8ed42f040a352026-10-15daily0.2 https://ai-coding.medium.com/the-47-second-worst-case-fallback-is-a-good-reminder-that-a-safety-net-can-become-its-own-latency-5a30e73ef51a2026-10-14daily0.5 https://ai-coding.medium.com/a-response-can-look-randomly-truncated-when-the-real-break-happened-earlier-at-a-tool-call-or-think-dba1f287e8472026-10-13daily0.5 https://ai-coding.medium.com/63-tokens-per-second-sounds-fast-until-a-cold-128k-prompt-sits-there-for-170-seconds-before-the-badc37189ea32026-10-13daily0.5 https://ai-coding.medium.com/a-wrong-digit-can-survive-because-the-field-still-looks-perfectly-valid-88cb0788ab892026-10-12daily0.2 https://ai-coding.medium.com/the-lease-can-expire-while-the-old-worker-still-thinks-the-job-is-its-own-f3be0e34afcd2026-10-10daily0.2 https://ai-coding.medium.com/the-counter-only-broke-after-scrolling-away-and-back-cfe6d80998222026-10-10daily0.5 https://ai-coding.medium.com/ssl-pinning-fixes-one-trust-problem-and-creates-another-operational-one-688f50c311282026-10-09daily0.5 https://ai-coding.medium.com/catching-a-wrong-assumption-in-the-plan-is-much-cheaper-than-finding-it-after-forty-files-change-88021b4dee342026-10-09daily0.5 https://ai-coding.medium.com/a-metric-can-stay-technically-healthy-long-after-its-definition-stops-being-useful-b29ce351245e2026-10-09daily0.5 https://ai-coding.medium.com/a-seamless-clip-doesnt-tell-you-whether-the-guest-game-stayed-live-f6e8391e724a2026-10-06daily0.5 https://ai-coding.medium.com/openai-12h-prompt-caching-pricing-473aec98fb682026-10-23daily0.5 https://ai-coding.medium.com/the-browsing-without-browsing-training-detail-was-the-most-surprising-part-to-me-5dbb590f531b2026-10-22daily0.5 https://ai-coding.medium.com/your-2020-example-makes-the-inflection-point-distinction-stick-daily-cases-can-peak-while-total-ea7bc3ed34232026-10-22daily0.5 https://ai-coding.medium.com/i-like-the-split-between-letting-the-agent-understand-the-question-and-having-a-deterministic-tool-bb9fc5aed7412026-10-22daily0.5 https://ai-coding.medium.com/the-frozen-model-detail-is-what-makes-harness²-interesting-the-improvement-loop-works-on-the-setup-5ff301b219f72026-10-22daily0.5 https://ai-coding.medium.com/the-final-caveat-is-the-part-id-keep-one-flag-isn-t-proof-someone-lied-it-s-a-reason-to-check-88022d6dab302026-10-22daily0.5 https://ai-coding.medium.com/i-appreciate-that-you-left-the-tradeoff-in-the-dynamic-workflow-caught-more-planted-bugs-but-cost-e627e95fcf852026-10-22daily0.5 https://ai-coding.medium.com/the-repeated-6-km-commute-example-made-the-singular-matrix-error-click-for-me-if-every-trip-has-2f9fdf846db82026-10-22daily0.2 https://ai-coding.medium.com/i-like-that-the-privacy-and-customization-points-are-separate-92c916a3575d2026-10-21daily0.5 https://ai-coding.medium.com/the-jump-from-code-search-to-an-agent-owned-inbox-and-smart-home-controls-caught-my-attention-5982711fd5f82026-10-21daily0.5 https://ai-coding.medium.com/the-row-order-caveat-is-the-detail-that-made-this-benchmark-feel-practical-to-me-e84d58f8fe512026-10-21daily0.5 https://ai-coding.medium.com/i-liked-the-distinction-between-torch-compile-9b1dd03a30052026-10-21daily0.2 https://ai-coding.medium.com/the-timing-matters-here-864f54137a752026-10-19daily0.2 https://ai-coding.medium.com/requiring-two-successful-writers-gives-the-workflow-room-to-recover-from-partial-failures-757d30d6ef922026-10-19daily0.5 https://ai-coding.medium.com/i-hadnt-thought-about-someone-logging-in-halfway-through-a-streamed-reply-d788678645c12026-10-16daily0.5 https://ai-coding.medium.com/how-anyps5-works-7ab01188ee972026-10-16daily0.2 https://ai-coding.medium.com/the-six-hour-boundary-check-is-a-nice-detail-cbb025f3ed6c2026-10-16daily0.5 https://ai-coding.medium.com/the-model-names-will-change-but-the-closing-criteria-are-reusable-check-memory-fit-chip-5a6f4dcc01b62026-11-02daily0.5 https://ai-coding.medium.com/the-goal-section-makes-an-important-distinction-an-evaluator-can-judge-the-evidence-it-sees-but-ad70cff220eb2026-11-02daily0.5 https://ai-coding.medium.com/the-distinction-between-phps-runtime-security-issues-and-the-legacy-practices-built-around-it-is-0c815278f5842026-10-30daily0.2 https://ai-coding.medium.com/should-you-build-on-genie-4-6b46e94cd7902026-10-29daily0.5 https://ai-coding.medium.com/your-parallel-pathways-example-makes-the-identifiability-problem-concrete-an-intervention-can-show-ca6d7b9ef2612026-10-26daily0.5 https://ai-coding.medium.com/the-handoff-files-give-people-and-agents-one-shared-starting-point-11b28486afd52026-10-26daily0.5 https://ai-coding.medium.com/just-upgrade-the-embedding-model-gets-a-lot-less-casual-once-theres-a-live-index-behind-it-88b8a825934a2026-08-26monthly0.2 https://ai-coding.medium.com/the-abstention-problem-feels-more-important-than-the-generation-problem-here-a921410d32522026-08-26monthly0.5 https://ai-coding.medium.com/composition-gets-especially-interesting-once-the-underlying-facts-stop-being-static-d95b15010a0e2026-08-26monthly0.2 https://ai-coding.medium.com/a-lot-of-what-gets-labeled-as-an-llm-hallucination-in-rag-is-really-an-upstream-retrieval-failure-f64d907306a52026-08-26monthly0.5 https://ai-coding.medium.com/pure-c-is-a-pretty-unforgiving-teacher-a74595f7522f2026-08-26monthly0.5 https://ai-coding.medium.com/one-of-the-most-convincing-choices-here-is-making-the-correlate-agent-deliberately-weaker-07eded869d272026-08-25monthly0.5 https://ai-coding.medium.com/the-idle-is-not-enough-section-is-the-one-ill-probably-remember-b5f513e2f38f2026-08-25monthly0.2 https://ai-coding.medium.com/zero-retention-ai-still-needs-cross-session-safety-signals-534f6953978c2026-08-25monthly0.5 https://ai-coding.medium.com/voice-is-a-modality-not-the-agent-architecture-is-the-line-that-made-the-whole-stack-click-for-f585ec309d2b2026-08-25monthly0.2 https://ai-coding.medium.com/the-provenance-and-confidence-layer-makes-this-much-more-convincing-than-just-put-the-repo-in-a-17c80f686d0b2026-08-24monthly0.5 https://ai-coding.medium.com/there-was-something-genuinely-fun-about-the-early-vibe-coding-phase-4699e9b287b32026-08-24monthly0.2 https://ai-coding.medium.com/ai-making-the-job-easier-and-raising-the-bar-at-the-same-time-is-probably-the-part-that-feels-most-88e5ddc0590f2026-08-24monthly0.5 https://ai-coding.medium.com/i-think-this-finally-gave-me-the-right-mental-model-for-browser-agents-ea94cf4961c72026-08-24monthly0.5 https://ai-coding.medium.com/the-stock-picks-are-almost-the-least-interesting-part-here-1eb9e8242deb2026-08-24monthly0.5 https://ai-coding.medium.com/its-interesting-how-ai-sre-can-mean-two-very-different-things-an-agent-that-investigates-after-d1b7b99e4d882026-08-24monthly0.5 https://ai-coding.medium.com/the-more-i-read-about-agents-the-more-it-feels-like-the-model-is-becoming-the-least-surprising-5c5880e089822026-08-24monthly0.7 https://ai-coding.medium.com/really-like-the-closed-loop-framing-here-a572e84512142026-08-21monthly0.5 https://ai-coding.medium.com/the-skipped-verification-call-feels-like-the-most-important-failure-here-2b7d94dcbeae2026-08-21monthly0.5 https://ai-coding.medium.com/really-liked-how-clearly-this-separates-model-behavior-from-execution-control-f314e1d142172026-08-21monthly0.2 https://ai-coding.medium.com/the-skipped-verification-call-feels-like-the-most-important-failure-here-6dc73c324a4d2026-08-21monthly0.5 https://ai-coding.medium.com/really-like-that-this-treats-skills-mcp-and-multiple-models-as-parts-of-one-testing-workflow-952ca937e7af2026-08-21monthly0.5 https://ai-coding.medium.com/really-like-this-distinction-87b46845f84a2026-08-21monthly0.2 https://ai-coding.medium.com/the-read-only-boundary-is-a-good-start-but-live-data-only-solves-one-layer-of-the-trust-problem-29b64436e3062026-08-20monthly0.5 https://ai-coding.medium.com/the-factory-framing-gets-much-more-useful-once-throughput-is-measured-across-the-whole-path-not-e06c710cecf22026-08-20monthly0.5 https://ai-coding.medium.com/the-ratchet-is-the-part-that-feels-especially-dangerous-d6a2ad2b60512026-08-20monthly0.5 https://ai-coding.medium.com/stateful-sandboxes-seem-like-the-harder-design-problem-here-320fa9b86d942026-08-20monthly0.2 https://ai-coding.medium.com/using-graphrag-before-extraction-is-the-part-i-find-most-interesting-53f7b61ff7442026-08-19monthly0.5 https://ai-coding.medium.com/the-interruption-handling-stood-out-to-me-ed7a092cab202026-08-19monthly0.5 https://ai-coding.medium.com/the-competency-question-framing-is-useful-but-the-maintenance-loop-feels-even-more-important-fb55b24f70a02026-08-19monthly0.5 https://ai-coding.medium.com/one-thing-i-keep-coming-back-to-with-the-capability-layer-idea-is-where-the-mcp-server-boundary-d2dfd5afa5b12026-08-18monthly0.5 https://ai-coding.medium.com/once-an-agent-can-call-tools-and-change-state-what-does-this-entity-mean-2152eb92c16b2026-08-18monthly0.5 https://ai-coding.medium.com/the-harder-memory-problem-for-me-isnt-retrieval-it-s-authority-2ec658333a242026-08-18monthly0.5 https://ai-coding.medium.com/id-be-curious-how-that-design-changes-when-the-action-is-irreversible-because-a-failed-code-e9aed0c340bc2026-08-18monthly0.2 https://ai-coding.medium.com/what-i-kept-thinking-about-here-is-that-the-framework-question-probably-only-becomes-important-once-820c2f4ca19a2026-08-17monthly0.5 https://ai-coding.medium.com/the-three-revenue-numbers-example-makes-the-problem-feel-much-more-concrete-a06c2a7469702026-08-17monthly0.7 https://ai-coding.medium.com/i-really-liked-this-piece-especially-the-argument-around-document-based-workflows-6b9f62e8d17c2026-08-17monthly0.5 https://ai-coding.medium.com/the-funny-thing-is-that-removing-the-retriever-doesnt-really-remove-the-retrieval-problem-it-just-bdd9a2deab442026-08-17monthly0.2 https://ai-coding.medium.com/this-made-me-think-that-the-moment-an-agent-can-take-an-irreversible-action-observability-stops-295778895bf92026-08-17monthly0.5 https://ai-coding.medium.com2026-11-04weekly1.0 https://ai-coding.medium.com/about2026-11-04weekly1.0 https://ai-coding.medium.com/the-architecture-makes-sense-to-me-but-id-separate-deterministic-retrieval-from-deterministic-a3f3cd51afd72026-08-13monthly0.5 https://ai-coding.medium.com/the-ownership-part-feels-more-important-than-the-schema-9bf78b2e7c042026-08-13monthly0.5 https://ai-coding.medium.com/i-like-the-idea-of-giving-failed-hypotheses-a-reopen-condition-instead-of-just-deleting-them-f38c6be2018d2026-08-13monthly0.5 https://ai-coding.medium.com/the-irreversible-decision-boundary-is-the-part-i-keep-coming-back-to-a2d555e5e3732026-08-12monthly0.5 https://ai-coding.medium.com/grok-bot-has-its-own-computer-your-approval-boundary-just-moved-9f1043294e8c2026-08-12monthly0.5 https://ai-coding.medium.com/the-plant-getting-worse-is-probably-the-most-useful-result-here-93d1a388789d2026-08-12monthly0.5 https://ai-coding.medium.com/the-harness-distinction-is-the-useful-one-for-me-6c90af44b7392026-08-12monthly0.5 https://ai-coding.medium.com/the-more-interesting-boundary-is-between-fitting-and-serving-f924bdfe46df2026-08-11monthly0.5 https://ai-coding.medium.com/the-interface-vs-architecture-distinction-makes-this-much-clearer-57d1ce64ce082026-08-11monthly0.2 https://ai-coding.medium.com/7-ai-agent-evaluation-tools-every-developer-should-know-before-2027-33bbebce48ad2026-08-10monthly0.5 https://ai-coding.medium.com/your-agent-has-50-skills-which-ones-should-be-prompts-564a9b1e6e412026-08-06monthly0.2 https://ai-coding.medium.com/human-in-the-loop-only-works-if-the-human-can-still-independently-judge-the-output-ec931ba423072026-09-04monthly0.5 https://ai-coding.medium.com/this-feels-less-like-improving-the-chatbox-and-more-like-adding-an-intent-layer-before-the-agent-f1cdf71533012026-09-04monthly0.5 https://ai-coding.medium.com/the-agent-friendly-api-point-matters-more-than-it-gets-credit-for-9a4aeb779a872026-09-04monthly0.2 https://ai-coding.medium.com/for-llm-council-id-care-less-about-the-number-of-models-and-more-about-error-correlation-2ff833ed9f202026-09-04monthly0.5 https://ai-coding.medium.com/id-be-careful-with-the-use-all-three-stack-7a62e9785f252026-09-04monthly0.5 https://ai-coding.medium.com/the-backlog-debt-is-the-part-id-treat-as-a-first-class-system-metric-0d34e7660ee82026-09-04monthly0.5 https://ai-coding.medium.com/the-retrieval-vs-0394d6d43d4b2026-09-04monthly0.2 https://ai-coding.medium.com/the-tricky-part-may-be-keeping-all-those-memory-layers-in-sync-cda20641f2e72026-09-03monthly0.5 https://ai-coding.medium.com/getting-a-125b-model-to-fit-on-consumer-hardware-is-impressive-3fcb77de2c662026-09-03monthly0.5 https://ai-coding.medium.com/once-the-plugin-becomes-the-distribution-boundary-skill-to-mcp-compatibility-becomes-part-of-the-6aceef3d87532026-09-03monthly0.5 https://ai-coding.medium.com/impossible-probably-shouldnt-be-a-model-opinion-72db70d96a512026-09-03monthly0.5 https://ai-coding.medium.com/parallel-agents-only-help-until-human-review-becomes-the-bottleneck-c6c90cdb669e2026-09-03monthly0.7 https://ai-coding.medium.com/the-unassisted-calibration-step-is-interesting-7de667347aaa2026-09-03monthly0.5 https://ai-coding.medium.com/separating-prefill-from-decode-feels-especially-important-when-reporting-inference-gains-c6ed7e87ebaf2026-09-02monthly0.5 https://ai-coding.medium.com/building-one-agent-from-raw-api-calls-before-using-a-framework-is-probably-one-of-the-58179603752a2026-09-02monthly0.5 https://ai-coding.medium.com/verify-the-world-not-the-self-report-is-probably-the-rule-id-steal-first-788f8ec49bef2026-09-02monthly0.2 https://ai-coding.medium.com/i-think-the-hidden-metric-here-is-how-much-glue-code-appears-once-you-push-a-framework-outside-its-ed7f14bcb3922026-09-02monthly0.5 https://ai-coding.medium.com/i-like-using-reversibility-as-the-filter-here-1f993d90be2f2026-09-02monthly0.5 https://ai-coding.medium.com/the-eval-set-itself-probably-needs-versioning-once-the-system-gets-this-dynamic-bcb20aa84f042026-09-02monthly0.2 https://ai-coding.medium.com/token-cost-gets-much-more-useful-once-it-has-a-reason-attached-to-it-667a8eb082f42026-09-01monthly0.5 https://ai-coding.medium.com/counting-timeouts-and-rejected-documents-as-failures-feels-important-cb1ff91d38a32026-09-01monthly0.5 https://ai-coding.medium.com/icl-explains-a-lot-of-the-model-side-behavior-but-im-not-sure-it-explains-the-whole-agent-runtime-4308c25b00ce2026-09-01monthly0.5 https://ai-coding.medium.com/the-stopping-condition-feels-just-as-important-as-the-loop-itself-01ccb5bff7ad2026-09-01monthly0.5 https://ai-coding.medium.com/self-correction-only-works-as-well-as-the-grader-does-11ae02b8eedc2026-09-01monthly0.2 https://ai-coding.medium.com/local-inference-could-change-the-trust-boundary-for-agents-quite-a-bit-d857446a36742026-09-01monthly0.5 https://ai-coding.medium.com/self-correction-only-works-as-well-as-the-grader-does-96e4c9a94a622026-09-01monthly0.5 https://ai-coding.medium.com/the-weather-example-is-a-good-reminder-that-removing-a-tool-doesnt-necessarily-remove-the-04b0995166b32026-08-31monthly0.5 https://ai-coding.medium.com/schema-valid-is-not-the-same-as-source-correct-3a189746a31d2026-08-31monthly0.5 https://ai-coding.medium.com/for-agents-id-want-the-semantic-layer-to-expose-one-thing-very-clearly-what-was-asserted-what-86597c078b2b2026-08-31monthly0.5 https://ai-coding.medium.com/context-graphs-get-even-more-useful-when-the-edges-carry-provenance-too-e3e37fd6bb8a2026-08-31monthly0.5 https://ai-coding.medium.com/ontorank-is-probably-the-most-interesting-piece-here-c2efae31d4892026-08-31monthly0.5 https://ai-coding.medium.com/automating-schema-and-prompt-creation-is-useful-but-the-evaluation-loop-is-what-makes-this-feel-4bef0c198f872026-08-28monthly0.5 https://ai-coding.medium.com/compiled-context-only-works-if-freshness-is-part-of-the-build-path-93c00901b6fe2026-08-28monthly0.2 https://ai-coding.medium.com/a-shared-agent-layer-seems-like-a-huge-compounding-advantage-4383ea37dcf82026-08-28monthly0.5 https://ai-coding.medium.com/the-interesting-boundary-is-that-exploration-can-be-agentic-but-correctness-still-needs-an-18b88cd1d0cb2026-08-28monthly0.5 https://ai-coding.medium.com/a-memory-system-that-only-learns-how-to-save-will-eventually-create-its-own-context-noise-eedaaf4144662026-08-28monthly0.5 https://ai-coding.medium.com/the-graph-had-the-facts-8df13070545f2026-08-27monthly0.2 https://ai-coding.medium.com/if-the-ontology-is-incomplete-or-the-real-world-concept-does-not-fit-the-formalism-cleanly-the-1bc5e8cb06872026-08-27monthly0.5 https://ai-coding.medium.com/once-tenant-acl-time-and-shard-routing-become-first-class-concerns-rag-starts-looking-much-more-e0b1eb75000b2026-08-27monthly0.5 https://ai-coding.medium.com/the-runtime-context-boundary-is-easy-to-miss-and-really-important-9075f90cf8582026-08-27monthly0.5 https://ai-coding.medium.com/the-works-at-works-for-employed-by-example-is-such-a-small-mistake-but-it-can-make-every-83e9e1ea7ee22026-08-27monthly0.5 https://ai-coding.medium.com/trimming-the-tool-list-feels-like-more-than-a-token-optimization-here-8914fbacea9a2026-09-15monthly0.2 https://ai-coding.medium.com/the-memory-example-gets-overlooked-655f0091040a2026-09-15monthly0.5 https://ai-coding.medium.com/adding-tools-only-helps-when-each-one-closes-a-real-capability-gap-d0cfc74e93eb2026-09-15monthly0.5 https://ai-coding.medium.com/a-watermark-shows-that-claude-was-likely-involved-ebd4e2dbe7e62026-09-11monthly0.2 https://ai-coding.medium.com/the-no-silent-cloud-fallback-rule-is-easy-to-underestimate-a845475babae2026-09-11monthly0.5 https://ai-coding.medium.com/poor-data-structure-can-make-good-reasoning-fail-0382cdf97b342026-09-11monthly0.5 https://ai-coding.medium.com/once-an-agent-starts-taking-action-ontology-becomes-more-than-a-way-to-name-things-e282b91a6e992026-09-11monthly0.5 https://ai-coding.medium.com/perplexity-measures-uncertainty-not-correctness-b4dc388a9e512026-09-11monthly0.5 https://ai-coding.medium.com/an-early-weak-source-can-make-later-evidence-look-redundant-even-when-it-would-correct-the-pool-834649c605cd2026-09-10monthly0.5 https://ai-coding.medium.com/with-each-rollout-able-to-read-what-earlier-runs-left-in-artifactory-the-real-communication-89d43cd68fc02026-09-10monthly0.2 https://ai-coding.medium.com/async-memory-writes-solve-the-latency-problem-but-they-open-a-consistency-gap-6e45814944af2026-09-10monthly0.2 https://ai-coding.medium.com/a-dead-smart-button-and-a-missing-import-path-say-a-lot-about-what-finished-means-for-coding-7eba65d0bd102026-09-10monthly0.5 https://ai-coding.medium.com/copilot-wont-read-your-late-issue-comment-move-it-to-the-pr-ac87fe687e902026-09-09monthly0.2 https://ai-coding.medium.com/the-move-from-prompt-rules-to-code-checks-is-the-right-part-to-focus-on-f1a2b6f4c4e72026-09-09monthly0.5 https://ai-coding.medium.com/three-retries-sounds-small-until-three-different-layers-sdk-queue-and-workflow-layer-each-make-1a685287a1362026-09-09monthly0.2 https://ai-coding.medium.com/recall-is-probably-the-hardest-of-the-four-fields-to-standardize-acafb290f1d02026-09-09monthly0.5 https://ai-coding.medium.com/shared-features-fix-one-consistency-problem-and-create-another-dependency-risk-9c3db9052e5e2026-09-09monthly0.2 https://ai-coding.medium.com/the-point-about-duplicate-delivery-and-idempotent-consumers-stood-out-8744d826860f2026-09-09monthly0.5 https://ai-coding.medium.com/the-point-about-duplicate-delivery-and-idempotent-consumers-stood-out-4c247477de532026-09-09monthly0.5 https://ai-coding.medium.com/fewer-tool-calls-sound-like-a-win-8cbca1cca1f02026-09-08monthly0.5 https://ai-coding.medium.com/verification-is-only-useful-when-it-measures-the-right-thing-d2bc52430d032026-09-07monthly0.7 https://ai-coding.medium.com/authorization-becomes-harder-when-agents-can-act-over-time-aa07d4240cd22026-09-07monthly0.5 https://ai-coding.medium.com/the-model-is-only-one-part-of-the-system-c4ebce8ad8972026-09-07monthly0.5 https://ai-coding.medium.com/persistent-agents-shift-reliability-from-model-quality-to-state-management-1311546de7e62026-09-07monthly0.2 https://ai-coding.medium.com/coding-agents-can-accelerate-implementation-but-losing-context-around-the-codebase-is-where-the-b091b39f06882026-09-07monthly0.5 https://ai-coding.medium.com/a-5-95-gb-pack-isnt-a-ready-to-run-budget-86a4a8e19eb32026-10-05daily0.5 https://ai-coding.medium.com/thats-a-useful-smoke-test-but-8-15-examples-per-task-means-one-or-two-labels-can-swing-a-task-43787b0246832026-10-05daily0.5 https://ai-coding.medium.com/the-first-section-gets-at-a-dependency-that-code-demos-hide-the-model-isnt-just-a-package-choice-421cfddfc90d2026-10-05daily0.5 https://ai-coding.medium.com/two-nearby-objects-swapping-identities-is-a-nasty-failure-because-the-tracker-can-keep-producing-31388bfadc9e2026-10-02monthly0.5 https://ai-coding.medium.com/claude-code-memory-claude-mem-4393d72199522026-10-02monthly0.5 https://ai-coding.medium.com/forty-dollars-and-zero-working-code-makes-the-coordination-cost-hard-to-ignore-84a722e7eee02026-10-02monthly0.5 https://ai-coding.medium.com/speech-onset-and-turn-completion-are-solving-different-timing-problems-2c4f16f07fd62026-10-02monthly0.5 https://ai-coding.medium.com/a-1-0-bug-probability-sitting-next-to-0-53-030d9e22fc352026-09-30monthly0.5 https://ai-coding.medium.com/the-model-swap-test-is-the-part-worth-running-first-2c51b66e0e7b2026-09-30monthly0.5 https://ai-coding.medium.com/shuffle-the-labels-and-the-same-case-can-cross-the-automation-threshold-2b488b17856c2026-09-29monthly0.5 https://ai-coding.medium.com/746-gefns-made-me-wonder-how-many-still-have-the-original-reason-attached-5c08c90e515f2026-09-28monthly0.5 https://ai-coding.medium.com/dropping-the-old-compacted-block-makes-the-loss-explicit-ba97fde4ce3b2026-09-28monthly0.2 https://ai-coding.medium.com/two-stable-weeks-before-retries-jumped-past-60-is-a-nasty-failure-case-c57aeb1f675f2026-09-28monthly0.5 https://ai-coding.medium.com/id-still-want-the-original-failed-call-logged-somewhere-outside-the-model-context-a3c39c6427762026-09-28monthly0.2