https://terryli.ai/https://terryli.ai/about/https://terryli.ai/colophon/https://terryli.ai/library/https://terryli.ai/posts/https://terryli.ai/posts/2/https://terryli.ai/posts/3/https://terryli.ai/posts/4-principles-for-agent-facing-cli-design/2026-04-07T02:35:07.000Zhttps://terryli.ai/posts/4/https://terryli.ai/posts/5/https://terryli.ai/posts/6/https://terryli.ai/posts/7/https://terryli.ai/posts/8/https://terryli.ai/posts/9/https://terryli.ai/posts/10/https://terryli.ai/posts/11/https://terryli.ai/posts/12/https://terryli.ai/posts/13/https://terryli.ai/posts/14/https://terryli.ai/posts/270-agents-while-i-slept/2026-07-26T00:00:00.000Zhttps://terryli.ai/posts/2026-03-19-exoskeleton-not-colleague/2026-03-19T09:00:00.000Zhttps://terryli.ai/posts/2026-07-25-your-check-passed-on-the-proxy/2026-07-25T10:30:00.000Zhttps://terryli.ai/posts/2026-07-26-near-parity-is-mostly-a-claim-about-formatting/2026-07-26T00:00:00.000Zhttps://terryli.ai/posts/2026-07-26-the-number-checks-out-the-claim-doesnt/2026-07-26T00:00:00.000Zhttps://terryli.ai/posts/2026-07-26-your-reviewer-model-is-not-independent/2026-07-26T06:07:46.000Zhttps://terryli.ai/posts/a-garden-is-not-a-changelog/2026-05-15T11:50:39.000Zhttps://terryli.ai/posts/a-persona-is-not-a-control/2026-05-17T05:55:08.000Zhttps://terryli.ai/posts/a-skill-is-not-a-prompt/2026-06-02T22:55:00.000Zhttps://terryli.ai/posts/access-control-not-abstraction/2026-03-25T09:30:00.000Zhttps://terryli.ai/posts/act-on-receipt-the-third-task-class/2026-03-13T03:28:00.000Zhttps://terryli.ai/posts/after-automation-judgment-becomes-infrastructure/2026-05-22T09:50:53.000Zhttps://terryli.ai/posts/after-the-harness/2026-05-26T10:28:05.000Zhttps://terryli.ai/posts/agent-execution-the-missing-layer/2026-05-06T07:50:00.000Zhttps://terryli.ai/posts/agent-experience-ax/2026-03-06T11:30:00.000Zhttps://terryli.ai/posts/agent-first-cli-design/2026-03-06T17:30:00.000Zhttps://terryli.ai/posts/agent-native-onboarding-is-not-a-signup-form/2026-05-23T03:45:00.000Zhttps://terryli.ai/posts/agentic-ai-in-production-looks-like-a-workflow/2026-03-13T03:27:55.000Zhttps://terryli.ai/posts/agentic-engineering-less-is-more/2026-03-06T23:00:00.000Zhttps://terryli.ai/posts/agentic-search-ate-rag/2026-03-17T08:26:38.000Zhttps://terryli.ai/posts/agi-definitions-reliability-threshold/2026-03-06T15:00:00.000Zhttps://terryli.ai/posts/ai-agent-frameworks-enterprise-fs/2026-03-06T09:30:00.000Zhttps://terryli.ai/posts/ai-agent-teams-are-colonies-not-companies/2026-03-20T06:25:04.000Zhttps://terryli.ai/posts/ai-agents-need-notebooks-not-just-memories/2026-03-13T04:58:59.000Zhttps://terryli.ai/posts/ai-agents-need-passports/2025-01-27T00:00:00.000Zhttps://terryli.ai/posts/ai-controls-architecture/2026-04-18T13:09:00.000Zhttps://terryli.ai/posts/ai-displacement-loop/2026-03-06T19:00:00.000Zhttps://terryli.ai/posts/ai-era-business-archetypes/2026-03-06T14:00:00.000Zhttps://terryli.ai/posts/ai-evals-metrics-before-diagnosis/2026-03-06T21:00:00.000Zhttps://terryli.ai/posts/ai-fixed-my-perfectionism-sort-of/2026-03-07T04:53:00.000Zhttps://terryli.ai/posts/ai-governance-routing-vs-compliance/2026-03-06T12:00:00.000Zhttps://terryli.ai/posts/ai-job-title-illusion/2026-03-06T14:30:00.000Zhttps://terryli.ai/posts/ai-risk-tiering-banking/2026-03-19T12:00:00.000Zhttps://terryli.ai/posts/ai-vendor-values-decision/2026-03-06T20:30:00.000Zhttps://terryli.ai/posts/ai-vendors-are-not-neutral-infrastructure/2026-03-14T09:09:08.000Zhttps://terryli.ai/posts/always-latest-as-a-system-property/2026-04-04T07:18:08.000Zhttps://terryli.ai/posts/ambient-agents/2025-08-27T00:00:00.000Zhttps://terryli.ai/posts/assume-the-llm-never-ran/2026-04-11T13:26:34.000Zhttps://terryli.ai/posts/autonomy-starts-at-the-check/2026-05-16T04:59:45.000Zhttps://terryli.ai/posts/autopoiesis/2026-03-25T08:27:08.000Zhttps://terryli.ai/posts/backtest-vs-operational-validation/2026-03-20T10:00:00.000Zhttps://terryli.ai/posts/banking-ds-to-ai-consulting/2026-03-06T08:30:00.000Zhttps://terryli.ai/posts/banks-ax-problem/2026-03-06T16:00:00.000Zhttps://terryli.ai/posts/banks-dont-need-ai-strategy/2026-03-06T08:00:00.000Zhttps://terryli.ai/posts/benchmark-your-research-stack/2026-03-09T16:00:00.000Zhttps://terryli.ai/posts/bio-naming-as-design-constraint/2026-04-08T00:00:00.000Zhttps://terryli.ai/posts/bridge-or-seed/2026-03-23T11:47:47.000Zhttps://terryli.ai/posts/budding/2026-03-25T08:56:00.000Zhttps://terryli.ai/posts/building-a-bus-alert-system-in-one-session/2026-03-15T14:40:08.000Zhttps://terryli.ai/posts/building-consulting-toolkit-day-one/2026-03-06T17:00:00.000Zhttps://terryli.ai/posts/building-porin/2026-04-06T14:00:00.000Zhttps://terryli.ai/posts/case-against-knowledge-management/2026-03-11T08:10:00.000Zhttps://terryli.ai/posts/cast-the-wide-net/2026-03-16T10:32:23.000Zhttps://terryli.ai/posts/chinas-ai-stack-is-now-hardware-deep/2026-03-14T09:09:49.000Zhttps://terryli.ai/posts/choosing-a-steam-iron-with-vertical-steam/2026-03-22T01:30:54.000Zhttps://terryli.ai/posts/claude-code-analyze-my-spending/2025-08-14T10:00:00.000Zhttps://terryli.ai/posts/claude-code-mobile-evolution/2025-08-07T10:00:00.000Zhttps://terryli.ai/posts/claude-code-not-coding-agent/2025-07-26T10:00:00.000Zhttps://terryli.ai/posts/cli-enforces-structure-saves-tokens/2026-03-06T22:00:00.000Zhttps://terryli.ai/posts/cli-mcp-or-code-mode/2026-04-05T10:39:47.000Zhttps://terryli.ai/posts/composure-is-a-skill/2026-03-16T04:53:07.000Zhttps://terryli.ai/posts/compounding-the-only-mental-model/2026-03-16T06:20:00.000Zhttps://terryli.ai/posts/consulting-reducing-uncertainty/2026-03-11T08:20:00.000Zhttps://terryli.ai/posts/contract-pattern-hard-gates/2026-03-06T13:00:00.000Zhttps://terryli.ai/posts/conversation-is-metabolism/2026-03-21T20:36:59.000Zhttps://terryli.ai/posts/correctness-is-model-determined/2026-07-19T02:17:13.000Zhttps://terryli.ai/posts/cross-cutting-is-just-another-word-for-optional/2026-03-12T09:14:32.000Zhttps://terryli.ai/posts/cross-model-review-why-model-diversity-beats-model-capability/2026-03-20T04:47:27.000Zhttps://terryli.ai/posts/default-to-the-whole-conversation/2026-03-17T04:52:05.000Zhttps://terryli.ai/posts/delegation-is-delegation/2026-03-17T02:31:00.000Zhttps://terryli.ai/posts/design-actions-not-actors/2026-03-21T00:16:40.000Zhttps://terryli.ai/posts/detect-and-degrade/2026-05-10T08:20:17.000Zhttps://terryli.ai/posts/deterministic-over-judgment/2026-03-22T10:15:00.000Zhttps://terryli.ai/posts/division-of-labour-five-categories-for-human-ai-work/2026-03-19T09:51:21.000Zhttps://terryli.ai/posts/dont-ask-ai-to-find-problems/2026-03-06T11:00:00.000Zhttps://terryli.ai/posts/dont-be-impressed-by-fluency/2026-03-14T00:00:00.000Zhttps://terryli.ai/posts/dont-optimise-for-the-proxy/2026-03-17T08:16:19.000Zhttps://terryli.ai/posts/dual-interfaces-human-ai/2025-08-28T00:00:00.000Zhttps://terryli.ai/posts/eliminate-the-reminder-dont-schedule-it/2026-03-09T03:45:14.000Zhttps://terryli.ai/posts/enterprise-agents-transformation-not-technology/2026-03-16T06:40:00.000Zhttps://terryli.ai/posts/enterprise-ai-has-a-plumbing-problem-not-a-model-problem/2026-03-11T09:29:28.000Zhttps://terryli.ai/posts/enzyme-receptor-cell-type/2026-03-25T08:33:25.000Zhttps://terryli.ai/posts/everyone-becomes-middle-management/2025-01-27T10:00:00.000Zhttps://terryli.ai/posts/everything-is-energy/2026-03-21T13:59:57.000Zhttps://terryli.ai/posts/exa-indexes-wechat/2026-03-07T03:46:52.000Zhttps://terryli.ai/posts/expansion-not-speedup/2026-03-07T01:12:16.000Zhttps://terryli.ai/posts/force-the-level/2026-03-25T02:55:00.000Zhttps://terryli.ai/posts/format-is-thinking-discipline/2026-05-04T09:00:00.000Zhttps://terryli.ai/posts/framework-rejection-primitive-harvest/2026-05-02T09:00:00.000Zhttps://terryli.ai/posts/from-chatbots-to-event-loops/2026-03-11T09:00:00.000Zhttps://terryli.ai/posts/garp-rai-exam-surprises/2026-03-06T15:30:00.000Zhttps://terryli.ai/posts/good-enough-parrots/2026-03-17T06:16:04.000Zhttps://terryli.ai/posts/govern-the-workflow-not-the-model/2026-05-16T07:29:22.000Zhttps://terryli.ai/posts/governance-is-a-design-problem/2026-04-03T10:22:00.000Zhttps://terryli.ai/posts/governance-is-a-tax/2026-03-18T05:07:03.000Zhttps://terryli.ai/posts/governing-agents-the-way-cells-govern-themselves/2026-04-18T12:00:00.000Zhttps://terryli.ai/posts/gpt-54-pro-responses-api-silent-stall/2026-03-07T14:00:00.000Zhttps://terryli.ai/posts/growing-up/2026-03-25T02:22:00.000Zhttps://terryli.ai/posts/guardrails-as-rivers/2026-03-16T07:00:00.000Zhttps://terryli.ai/posts/guardrails-beat-guidance/2026-03-13T04:19:14.000Zhttps://terryli.ai/posts/hidden-violence-vague-instructions/2025-08-27T00:00:00.000Zhttps://terryli.ai/posts/hk-apac-ai-hub-financial-services/2026-03-06T10:00:00.000Zhttps://terryli.ai/posts/honesty-as-default/2026-03-16T06:30:00.000Zhttps://terryli.ai/posts/hooks-are-life-infrastructure/2026-03-16T06:02:40.000Zhttps://terryli.ai/posts/how-claude-code-helps-you-think/2025-08-01T10:00:00.000Zhttps://terryli.ai/posts/how-i-used-6-llms-to-write-one-word-doc-comment/2026-04-18T22:46:28.000Zhttps://terryli.ai/posts/how-to-think-with-ai/2026-03-16T07:20:00.000Zhttps://terryli.ai/posts/human-in-the-loop-is-an-architecture-decision/2026-03-17T16:00:00.000Zhttps://terryli.ai/posts/human-oversight-doesnt-scale/2026-03-18T05:14:00.000Zhttps://terryli.ai/posts/hygiene/2026-03-21T22:37:40.000Zhttps://terryli.ai/posts/i-built-200-clis-for-my-ai-heres-what-actually-matters/2026-04-12T09:00:00.000Zhttps://terryli.ai/posts/i-didnt-mean-to-kill-my-todo-app/2026-03-07T08:05:20.000Zhttps://terryli.ai/posts/i-dont-read-documentation-anymore/2025-08-20T10:00:00.000Zhttps://terryli.ai/posts/i-made-ai-remember-to-remember/2026-03-17T05:29:11.000Zhttps://terryli.ai/posts/i-made-ai-remind-me-blind-spots/2026-03-06T22:30:00.000Zhttps://terryli.ai/posts/immune-system-of-ai-autonomy/2026-03-18T08:07:51.000Zhttps://terryli.ai/posts/impossibility-theorems-as-consulting-tools/2026-03-17T16:00:00.000Zhttps://terryli.ai/posts/inference-cost-collapse-is-a-governance-liability/2026-03-18T09:47:12.000Zhttps://terryli.ai/posts/inline-beats-reference-for-llm-attention/2026-03-18T10:33:10.000Zhttps://terryli.ai/posts/instructions-dont-enforce-behavior-templates-do/2026-03-08T00:02:36.000Zhttps://terryli.ai/posts/intelligence-on-tap/2025-07-15T10:00:00.000Zhttps://terryli.ai/posts/is-insight-an-illusion/2026-03-13T08:15:18.000Zhttps://terryli.ai/posts/judgment-is-a-moving-boundary/2026-03-24T07:47:14.000Zhttps://terryli.ai/posts/kutta-condition-of-ai/2026-03-06T18:00:00.000Zhttps://terryli.ai/posts/language-medium-not-purpose/2026-03-11T08:40:00.000Zhttps://terryli.ai/posts/latest-is-a-race-condition/2026-05-15T11:50:39.000Zhttps://terryli.ai/posts/legibility-is-the-bottleneck/2026-03-18T14:50:00.000Zhttps://terryli.ai/posts/legibility-precedes-ai/2026-05-04T22:50:47.000Zhttps://terryli.ai/posts/let-the-os-schedule-let-your-tool-dispatch/2026-03-10T01:05:59.000Zhttps://terryli.ai/posts/lint-as-cartography/2026-05-10T08:24:37.000Zhttps://terryli.ai/posts/llm-evals-arent-data-science/2026-03-11T14:00:00.000Zhttps://terryli.ai/posts/llms-are-enzymes/2026-03-21T20:37:03.000Zhttps://terryli.ai/posts/llms-are-primers/2026-03-25T01:45:00.000Zhttps://terryli.ai/posts/llms-better-editing-than-writing/2026-03-11T08:30:00.000Zhttps://terryli.ai/posts/local-first-embeddings-for-regulated-industries/2026-03-17T16:00:00.000Zhttps://terryli.ai/posts/managing-ai-agents-like-managing-a-team/2026-03-20T06:21:39.000Zhttps://terryli.ai/posts/match-form-to-access-pattern/2026-03-18T14:53:00.000Zhttps://terryli.ai/posts/match-the-tool-to-the-shape/2026-03-19T04:39:24.000Zhttps://terryli.ai/posts/mcp-not-glorified-api/2025-08-27T00:00:00.000Zhttps://terryli.ai/posts/meta-skills-are-the-multiplier/2026-03-18T15:00:00.000Zhttps://terryli.ai/posts/metabolism-of-the-real-world/2026-03-22T04:45:00.000Zhttps://terryli.ai/posts/mining-management-theory-for-ai-agent-teams/2026-03-18T23:02:04.000Zhttps://terryli.ai/posts/mining-your-llm/2026-03-15T15:36:23.000Zhttps://terryli.ai/posts/model-risk-management-was-not-built-for-this/2026-03-18T05:16:00.000Zhttps://terryli.ai/posts/model-routing-is-a-design-decision/2026-03-17T16:00:00.000Zhttps://terryli.ai/posts/move-the-gate-to-the-package-manager/2026-05-13T01:00:00.000Zhttps://terryli.ai/posts/mox-and-ai-transformation/2026-03-06T12:30:00.000Zhttps://terryli.ai/posts/mtor-self-bootstrap/2026-04-07T12:00:00.000Zhttps://terryli.ai/posts/multi-model-deliberation/2026-03-06T18:00:00.000Zhttps://terryli.ai/posts/multi-persona-ai-review-models-the-receiver/2026-05-06T07:00:00.000Zhttps://terryli.ai/posts/name-your-system-after-biology-then-rename-it/2026-03-23T03:06:12.000Zhttps://terryli.ai/posts/naming-is-a-design-review/2026-03-23T14:07:24.000Zhttps://terryli.ai/posts/no-stable-moat/2026-03-17T06:28:47.000Zhttps://terryli.ai/posts/not-every-cron-job-is-a-feedback-loop/2026-03-13T08:30:32.000Zhttps://terryli.ai/posts/observability-is-not-assurance/2026-04-19T11:50:08.000Zhttps://terryli.ai/posts/one-billion-bet-against-llms/2026-03-11T09:20:00.000Zhttps://terryli.ai/posts/openai-frontier-alliance-advisory-firms/2026-03-17T16:00:00.000Zhttps://terryli.ai/posts/openclaw-this-years-deepseek/2026-03-11T09:30:00.000Zhttps://terryli.ai/posts/operating-papers-and-board-papers-cant-be-the-same-document/2026-04-25T09:00:00.000Zhttps://terryli.ai/posts/optimize-for-routing-not-tokens/2026-03-18T15:00:00.000Zhttps://terryli.ai/posts/over-capture-under-capture/2026-03-16T06:10:00.000Zhttps://terryli.ai/posts/overnight-autonomous-ai-coding-what-works/2026-04-10T07:00:00.000Zhttps://terryli.ai/posts/per-token-pricing-megapixels/2026-03-06T14:00:00.000Zhttps://terryli.ai/posts/personas-exploit-llm-judges/2026-07-26T00:00:00.000Zhttps://terryli.ai/posts/philosophy-isnt-the-opposite-of-practical/2026-03-14T00:00:00.000Zhttps://terryli.ai/posts/plan-execute-split/2026-03-12T09:12:28.000Zhttps://terryli.ai/posts/planning-needs-eyes/2026-03-14T14:00:00.000Zhttps://terryli.ai/posts/play-within-the-design/2026-03-18T10:00:00.000Zhttps://terryli.ai/posts/production-ai-vs-demos/2025-07-20T10:00:00.000Zhttps://terryli.ai/posts/programs-over-prompts/2026-03-19T09:51:20.000Zhttps://terryli.ai/posts/progressive-disclosure-for-ai-agents/2026-03-17T16:00:00.000Zhttps://terryli.ai/posts/progressive-trust-how-to-give-ai-agents-autonomy-without-gambling/2026-03-14T03:37:06.000Zhttps://terryli.ai/posts/published-is-a-reader-state/2026-05-15T12:06:25.000Zhttps://terryli.ai/posts/push-not-pull/2026-03-13T03:28:00.000Zhttps://terryli.ai/posts/put-the-rule-where-it-fires/2026-03-14T12:16:57.000Zhttps://terryli.ai/posts/qdap-annuity-hong-kong/2026-03-08T21:30:00.000Zhttps://terryli.ai/posts/rag-compliance-chunking/2026-03-06T09:00:00.000Zhttps://terryli.ai/posts/rag-solved-the-wrong-problem/2026-03-11T10:00:00.000Zhttps://terryli.ai/posts/read-the-chain-not-the-date/2026-08-02T14:07:14.000Zhttps://terryli.ai/posts/reconstruction-over-retrieval/2026-07-26T00:00:00.000Zhttps://terryli.ai/posts/recovery-is-not-control/2026-05-17T13:18:28.000Zhttps://terryli.ai/posts/redundancy-is-the-only-honest-ai-research-strategy/2026-03-13T12:00:00.000Zhttps://terryli.ai/posts/reliability-hierarchy-hooks-rules-skills/2026-03-18T15:00:00.000Zhttps://terryli.ai/posts/retrieval-becomes-a-control/2026-07-18T06:25:34.000Zhttps://terryli.ai/posts/revealed-preference-in-interviews/2026-03-16T10:45:30.000Zhttps://terryli.ai/posts/rules-decay-hooks-dont/2026-03-07T00:40:13.000Zhttps://terryli.ai/posts/same-trigger-one-skill/2026-04-10T07:30:00.000Zhttps://terryli.ai/posts/save-conclusions-not-just-rules/2026-03-17T04:39:46.000Zhttps://terryli.ai/posts/seven-patterns-for-agent-facing-clis/2026-04-06T12:00:00.000Zhttps://terryli.ai/posts/shadow-agents-are-coming-for-your-org/2026-03-17T16:00:00.000Zhttps://terryli.ai/posts/shifting-priors-is-not-finding-truth/2026-03-13T03:19:55.000Zhttps://terryli.ai/posts/show-up-with-the-machine/2026-03-18T06:53:44.000Zhttps://terryli.ai/posts/skills-are-collapsed-recursion/2026-03-18T14:45:00.000Zhttps://terryli.ai/posts/skills-as-behavioral-nudges/2026-03-06T13:30:00.000Zhttps://terryli.ai/posts/skills-as-files/2026-03-07T01:12:16.000Zhttps://terryli.ai/posts/skills-prototype-mcp-production/2026-03-18T15:00:00.000Zhttps://terryli.ai/posts/skills-should-die/2026-03-24T07:47:14.000Zhttps://terryli.ai/posts/software-engineering-for-ai-instructions/2026-03-06T20:00:00.000Zhttps://terryli.ai/posts/spaced-repetition-for-beliefs/2026-03-17T05:46:02.000Zhttps://terryli.ai/posts/stealing-from-peers-a-truth-seeking-discipline/2026-03-18T09:15:24.000Zhttps://terryli.ai/posts/stop-theorizing-about-your-prompts/2026-03-20T03:07:58.000Zhttps://terryli.ai/posts/summarisation-is-a-test-of-comprehension-not-intelligence/2026-03-20T01:18:30.000Zhttps://terryli.ai/posts/sunyata-in-the-skill-library/2026-03-18T09:28:05.000Zhttps://terryli.ai/posts/supply-driven-compute/2026-03-18T14:15:00.000Zhttps://terryli.ai/posts/systematise-decisions-not-actions/2026-03-17T06:01:11.000Zhttps://terryli.ai/posts/taste-is-the-bottleneck/2026-07-26T00:00:00.000Zhttps://terryli.ai/posts/taste-is-the-metabolism/2026-03-21T13:56:16.000Zhttps://terryli.ai/posts/taste-requires-stakes/2026-03-06T16:30:00.000Zhttps://terryli.ai/posts/taste-works-for-small-bets/2026-03-13T03:42:35.000Zhttps://terryli.ai/posts/ten-things-i-learned-from-the-agent-skills-gold-rush/2026-04-05T04:31:19.000Zhttps://terryli.ai/posts/ten-types-of-actionable-knowledge/2026-03-17T03:51:30.000Zhttps://terryli.ai/posts/test-first-dispatch-for-ai-coding-agents/2026-04-07T13:25:00.000Zhttps://terryli.ai/posts/the-accidental-life-os/2026-03-11T09:32:43.000Zhttps://terryli.ai/posts/the-agent-governance-gap-is-already-here/2026-03-14T08:50:12.000Zhttps://terryli.ai/posts/the-agent-is-not-the-control-point/2026-05-17T06:23:14.000Zhttps://terryli.ai/posts/the-agent-is-the-trace/2026-05-25T12:30:00.000Zhttps://terryli.ai/posts/the-ai-portability-test/2026-07-27T05:40:48.000Zhttps://terryli.ai/posts/the-ai-trading-system-you-should-build-but-never-use/2026-03-17T07:52:36.000Zhttps://terryli.ai/posts/the-aidlt-conflation-trap-in-hkmas-march-2026-strategic-review-mandate/2026-03-18T09:46:40.000Zhttps://terryli.ai/posts/the-annotation-model-what-ai-journaling-gets-right/2026-03-17T09:19:16.000Zhttps://terryli.ai/posts/the-anti-slop-pattern/2026-04-18T14:00:00.000Zhttps://terryli.ai/posts/the-architect-implementer-split/2026-04-07T00:30:00.000Zhttps://terryli.ai/posts/the-architecture-biopsy/2026-03-25T03:35:00.000Zhttps://terryli.ai/posts/the-are-you-sure-loop/2026-03-13T22:21:00.000Zhttps://terryli.ai/posts/the-assistant-is-a-character/2026-03-13T23:00:00.000Zhttps://terryli.ai/posts/the-best-idea-i-didnt-adopt/2026-06-06T12:41:20.000Zhttps://terryli.ai/posts/the-black-box-that-responds-to-role-play/2026-03-13T22:45:00.000Zhttps://terryli.ai/posts/the-book-that-tells-you-not-to-read-it/2026-03-17T02:50:00.000Zhttps://terryli.ai/posts/the-bootstrap-problem-in-ai-tooling/2026-03-13T08:25:52.000Zhttps://terryli.ai/posts/the-boring-future-of-ai-agents/2026-03-18T05:24:53.000Zhttps://terryli.ai/posts/the-borrowed-brain/2026-03-25T02:15:00.000Zhttps://terryli.ai/posts/the-boundary-is-an-assessment/2026-03-24T09:15:31.000Zhttps://terryli.ai/posts/the-byproduct-trap/2026-03-13T06:08:29.000Zhttps://terryli.ai/posts/the-calculator-analogy/2026-03-13T11:34:54.000Zhttps://terryli.ai/posts/the-calibration-gap-why-ai-assurance-needs-experimental-rigour-at-consulting-speed/2026-08-01T12:33:14.000Zhttps://terryli.ai/posts/the-calibration-trap/2026-03-10T11:39:19.000Zhttps://terryli.ai/posts/the-cell-biology-agent-design-manual/2026-04-02T23:30:43.000Zhttps://terryli.ai/posts/the-check-belongs-at-the-trigger/2026-05-12T04:19:53.000Zhttps://terryli.ai/posts/the-cli-boundary/2026-03-14T05:39:07.000Zhttps://terryli.ai/posts/the-comfort-trap/2026-03-10T11:22:48.000Zhttps://terryli.ai/posts/the-confidence-stack/2026-03-17T01:41:00.000Zhttps://terryli.ai/posts/the-confidence-trap/2026-03-14T23:28:27.000Zhttps://terryli.ai/posts/the-constitution-eats-itself/2026-03-22T03:30:00.000Zhttps://terryli.ai/posts/the-dark-matter-of-the-cell/2026-03-25T02:00:00.000Zhttps://terryli.ai/posts/the-debate-round-is-where-value-lives/2026-03-15T14:16:44.000Zhttps://terryli.ai/posts/the-deliberation-format-is-the-product/2026-03-13T03:06:52.000Zhttps://terryli.ai/posts/the-dimensions-nobody-lists/2026-03-17T02:19:53.000Zhttps://terryli.ai/posts/the-dispatch-layer-was-eating-the-quality/2026-04-04T07:17:23.000Zhttps://terryli.ai/posts/the-due-test/2026-03-18T04:25:37.000Zhttps://terryli.ai/posts/the-easter-egg-that-landed/2026-03-16T10:17:34.000Zhttps://terryli.ai/posts/the-emergence-ladder-from-molecules-to-economies/2026-03-20T06:27:40.000Zhttps://terryli.ai/posts/the-engine-is-the-policy/2026-04-25T09:30:00.000Zhttps://terryli.ai/posts/the-eval-gap/2026-03-14T08:14:46.000Zhttps://terryli.ai/posts/the-experiment-loop-isnt-about-code/2026-04-17T13:30:00.000Zhttps://terryli.ai/posts/the-experiment-loop-without-the-gpu/2026-03-07T23:10:37.000Zhttps://terryli.ai/posts/the-expert-illusion/2026-03-14T15:00:00.000Zhttps://terryli.ai/posts/the-failures-that-look-like-success/2026-03-17T16:00:00.000Zhttps://terryli.ai/posts/the-fairness-impossibility-is-not-a-bug/2026-03-17T16:00:00.000Zhttps://terryli.ai/posts/the-first-datapoint/2026-03-11T09:10:00.000Zhttps://terryli.ai/posts/the-fluency-trap/2026-03-13T06:26:00.000Zhttps://terryli.ai/posts/the-four-layers-of-every-ai-agent/2026-03-17T16:00:00.000Zhttps://terryli.ai/posts/the-framework-that-writes-itself/2026-04-19T08:47:40.000Zhttps://terryli.ai/posts/the-frontier-is-no-longer-the-back-office/2026-05-24T12:00:00.000Zhttps://terryli.ai/posts/the-global-minimum-of-governance/2026-03-18T04:20:00.000Zhttps://terryli.ai/posts/the-grey-areas-are-the-whole-thing/2026-03-13T08:10:35.000Zhttps://terryli.ai/posts/the-heuristic-library/2026-03-17T02:38:00.000Zhttps://terryli.ai/posts/the-human-bus-problem/2026-03-13T03:28:00.000Zhttps://terryli.ai/posts/the-identification-problem/2026-03-13T03:28:00.000Zhttps://terryli.ai/posts/the-immune-system-pattern/2026-03-18T06:59:28.000Zhttps://terryli.ai/posts/the-infra-trap/2026-03-09T10:45:36.000Zhttps://terryli.ai/posts/the-integration-layer-is-the-moat/2026-03-17T16:00:00.000Zhttps://terryli.ai/posts/the-interlocutor-mode/2026-03-17T07:15:51.000Zhttps://terryli.ai/posts/the-invisible-puppeteer/2025-01-31T00:00:00.000Zhttps://terryli.ai/posts/the-judgement-bucket-how-to-be-rigorous-about-the-claims-you-cant-measure/2026-08-01T13:30:06.000Zhttps://terryli.ai/posts/the-knowledge-base-is-the-maintenance/2026-08-02T03:04:48.000Zhttps://terryli.ai/posts/the-knowledge-mining-gap/2026-03-17T06:19:39.000Zhttps://terryli.ai/posts/the-knowledge-that-disappears-when-you-try-to-capture-it/2026-03-14T08:47:31.000Zhttps://terryli.ai/posts/the-label-is-not-the-risk/2026-05-17T07:31:01.000Zhttps://terryli.ai/posts/the-lamp-that-knows-you/2026-03-18T05:29:47.000Zhttps://terryli.ai/posts/the-last-10-is-the-feedback-loop/2026-03-13T03:28:00.000Zhttps://terryli.ai/posts/the-learning-loop-plateau/2026-04-19T16:30:00.000Zhttps://terryli.ai/posts/the-lens-trick/2026-04-24T09:00:00.000Zhttps://terryli.ai/posts/the-lethal-trifecta-what-openclaws-security-crisis-teaches-about-ai-agent-architecture/2026-03-18T07:57:27.000Zhttps://terryli.ai/posts/the-llm-is-the-tool/2026-03-24T07:47:14.000Zhttps://terryli.ai/posts/the-locksmiths-box/2026-03-18T09:38:01.000Zhttps://terryli.ai/posts/the-loop-is-the-product/2026-03-13T08:27:47.000Zhttps://terryli.ai/posts/the-maker-checker-trap/2026-03-18T05:13:52.000Zhttps://terryli.ai/posts/the-marginal-agent/2026-03-20T09:08:36.000Zhttps://terryli.ai/posts/the-market-prices-leverage-not-value/2026-03-14T00:00:00.000Zhttps://terryli.ai/posts/the-missing-metabolism/2026-03-21T09:53:00.000Zhttps://terryli.ai/posts/the-model-is-not-the-unit-of-return/2026-06-01T14:04:34.000Zhttps://terryli.ai/posts/the-model-is-the-architecture/2026-03-22T10:53:39.000Zhttps://terryli.ai/posts/the-mteb-leader-barely-beats-a-free-model-on-agent-memory/2026-03-17T08:00:00.000Zhttps://terryli.ai/posts/the-nag-tax/2026-03-10T00:00:00.000Zhttps://terryli.ai/posts/the-name-collision-that-found-two-tools/2026-04-06T08:00:00.000Zhttps://terryli.ai/posts/the-naming-problem/2026-03-20T15:51:07.000Zhttps://terryli.ai/posts/the-navigation-problem-in-agent-flywheels/2026-03-19T09:51:22.000Zhttps://terryli.ai/posts/the-oauth-token-you-forgot-about/2026-04-23T12:00:00.000Zhttps://terryli.ai/posts/the-one-cycle-late-test/2026-03-19T08:59:09.000Zhttps://terryli.ai/posts/the-one-env-var-that-cost-a-day/2026-04-09T20:00:00.000Zhttps://terryli.ai/posts/the-orchestration-layer-is-knowledge-not-code/2026-03-13T08:17:20.000Zhttps://terryli.ai/posts/the-organism-has-a-cortex/2026-03-22T22:29:12.000Zhttps://terryli.ai/posts/the-organism-theory/2026-03-22T00:16:02.000Zhttps://terryli.ai/posts/the-personalised-system-era/2026-03-10T01:06:00.000Zhttps://terryli.ai/posts/the-pipeline-paradox/2026-03-17T14:24:15.000Zhttps://terryli.ai/posts/the-primary-source-tax/2026-04-05T14:43:08.000Zhttps://terryli.ai/posts/the-problem-with-clever-browser-automation/2026-03-07T03:46:52.000Zhttps://terryli.ai/posts/the-production-gap-why-ai-pilots-fail/2026-03-17T16:00:00.000Zhttps://terryli.ai/posts/the-program-is-the-plan/2026-06-03T14:05:00.000Zhttps://terryli.ai/posts/the-queue-should-live-where-your-thoughts-live/2026-03-09T11:00:00.000Zhttps://terryli.ai/posts/the-queue-that-texts-you-back/2026-03-09T10:00:00.000Zhttps://terryli.ai/posts/the-rename-that-built-a-tool/2026-04-04T14:17:58.000Zhttps://terryli.ai/posts/the-reversible-direction/2026-04-10T06:03:08.000Zhttps://terryli.ai/posts/the-risk-without-an-engineering-solution/2026-04-18T12:00:00.000Zhttps://terryli.ai/posts/the-search-and-replace-test/2026-04-19T23:00:00.000Zhttps://terryli.ai/posts/the-second-pass-finds-more/2026-03-12T05:37:55.000Zhttps://terryli.ai/posts/the-semantic-consumer/2026-03-21T10:39:10.000Zhttps://terryli.ai/posts/the-session-boundary-is-why-you-still-dont-have-ai-agents/2026-03-13T03:28:00.000Zhttps://terryli.ai/posts/the-silence-of-missing-skills/2026-03-18T10:32:46.000Zhttps://terryli.ai/posts/the-skill-is-knowing-what-matters/2026-03-13T03:34:45.000Zhttps://terryli.ai/posts/the-sop-is-the-product/2026-05-26T09:23:28.000Zhttps://terryli.ai/posts/the-specificity-trap/2026-03-17T16:00:00.000Zhttps://terryli.ai/posts/the-specimen-not-the-container/2026-03-18T08:58:16.000Zhttps://terryli.ai/posts/the-system-for-checking-is-not-the-checking/2026-03-13T01:18:33.000Zhttps://terryli.ai/posts/the-template-is-the-schema/2026-07-26T00:00:00.000Zhttps://terryli.ai/posts/the-test-before-the-output/2026-03-24T08:34:07.000Zhttps://terryli.ai/posts/the-thirty-minute-fix-for-a-non-existent-bug/2026-05-10T17:00:00.000Zhttps://terryli.ai/posts/the-thirty-year-gap-between-faking-and-understanding-natural-language/2026-03-13T08:55:40.000Zhttps://terryli.ai/posts/the-todo-intake-gate/2026-03-19T05:20:22.000Zhttps://terryli.ai/posts/the-treadmill-and-the-loop/2026-07-26T00:00:00.000Zhttps://terryli.ai/posts/the-trust-spectrum/2026-03-07T00:46:57.000Zhttps://terryli.ai/posts/the-unexplainable-alpha/2026-03-19T12:00:00.000Zhttps://terryli.ai/posts/the-vocabulary-trap/2026-03-20T23:18:20.000Zhttps://terryli.ai/posts/the-work-behind-the-work/2026-07-12T05:49:07.000Zhttps://terryli.ai/posts/the-wrong-metric-why-i-stopped-switching-ai-models-mid-session/2026-03-12T10:00:00.000Zhttps://terryli.ai/posts/there-is-no-llm-in-a-cell/2026-03-25T01:30:00.000Zhttps://terryli.ai/posts/three-ai-governance-blind-spots/2026-03-06T21:30:00.000Zhttps://terryli.ai/posts/three-apac-regulators-are-converging-on-ai-governance-banks-should-build-one-framework/2026-07-26T00:00:00.000Zhttps://terryli.ai/posts/three-crates-before-lunch/2026-03-06T17:00:00.000Zhttps://terryli.ai/posts/three-paradigms-agent-knowledge/2026-03-18T14:59:51.000Zhttps://terryli.ai/posts/three-things-aml-ai-gets-wrong/2026-03-06T15:00:00.000Zhttps://terryli.ai/posts/titration/2026-03-25T06:17:07.000Zhttps://terryli.ai/posts/tool-health-is-the-missing-layer-of-agent-native-apps/2026-05-16T03:55:43.000Zhttps://terryli.ai/posts/tools-are-skills-with-null-llm/2026-03-24T01:16:02.000Zhttps://terryli.ai/posts/traces-are-the-new-debugger/2026-03-07T01:12:16.000Zhttps://terryli.ai/posts/unknown-is-not-low-risk/2026-05-16T03:42:48.000Zhttps://terryli.ai/posts/upstream-constraint-pattern/2026-03-06T19:00:00.000Zhttps://terryli.ai/posts/what-60k-stars-actually-validates/2026-04-18T14:30:00.000Zhttps://terryli.ai/posts/what-16000-willison-posts-reveal/2026-04-07T12:00:00.000Zhttps://terryli.ai/posts/what-a-port-forgets/2026-05-21T20:22:35.000Zhttps://terryli.ai/posts/what-a-prompt-trim-cannot-see/2026-07-25T06:58:04.000Zhttps://terryli.ai/posts/what-alphasense-charges-ten-thousand-dollars-for/2026-03-14T08:15:41.000Zhttps://terryli.ai/posts/what-and-why-beat-how/2025-07-10T10:00:00.000Zhttps://terryli.ai/posts/what-anthropic-managed-agents-validates-and-what-to-steal/2026-04-09T14:00:00.000Zhttps://terryli.ai/posts/what-chinese-ai-labs-see-that-western-ones-dont/2026-03-17T16:00:00.000Zhttps://terryli.ai/posts/what-feels-like-play/2026-03-13T11:31:56.000Zhttps://terryli.ai/posts/what-hermes-agent-got-right/2026-04-19T17:00:00.000Zhttps://terryli.ai/posts/what-human-memory-teaches-ai-agents-and-what-it-doesnt/2026-03-14T12:10:45.000Zhttps://terryli.ai/posts/what-if-your-vault-had-residents/2026-03-14T13:34:56.000Zhttps://terryli.ai/posts/what-is-understanding/2026-03-14T00:00:00.000Zhttps://terryli.ai/posts/what-it-actually-takes-to-run-an-ai-agent-in-a-bank/2026-03-07T05:44:33.000Zhttps://terryli.ai/posts/what-it-feels-like-ai-80-percent/2026-03-11T08:00:00.000Zhttps://terryli.ai/posts/what-llm-wiki-looks-like-after-six-months/2026-04-09T04:21:55.000Zhttps://terryli.ai/posts/what-llms-dont-volunteer/2026-03-17T03:58:28.000Zhttps://terryli.ai/posts/what-makes-great-ai-consultant/2026-03-06T10:30:00.000Zhttps://terryli.ai/posts/what-mcp-actually-changes/2026-03-11T08:50:00.000Zhttps://terryli.ai/posts/what-the-receipts-cost/2026-05-19T22:43:47.000Zhttps://terryli.ai/posts/what-the-weights-dont-know/2026-03-17T06:23:57.000Zhttps://terryli.ai/posts/what-we-know-about-multi-agent-orchestration-and-why-it-might-not-matter/2026-03-17T23:27:15.000Zhttps://terryli.ai/posts/when-a-heuristic-has-two-homes/2026-03-18T08:58:59.000Zhttps://terryli.ai/posts/when-better-is-worse/2026-03-08T13:40:35.000Zhttps://terryli.ai/posts/when-code-gets-cheap-coordination-gets-expensive/2026-05-22T10:00:43.000Zhttps://terryli.ai/posts/when-defender-news-weakens-the-ask/2026-05-12T03:56:23.000Zhttps://terryli.ai/posts/when-langgraph-earns-its-keep/2026-03-14T17:00:00.000Zhttps://terryli.ai/posts/when-the-name-doesnt-fit/2026-04-05T00:48:58.000Zhttps://terryli.ai/posts/when-the-platform-is-mature-the-architects-job-changes/2026-03-11T00:00:00.000Zhttps://terryli.ai/posts/when-to-build-vs-wait/2026-03-06T16:00:00.000Zhttps://terryli.ai/posts/when-to-make-your-pipeline-agentic/2026-03-14T10:54:45.000Zhttps://terryli.ai/posts/when-to-think-and-when-to-count/2026-03-17T02:55:00.000Zhttps://terryli.ai/posts/when-your-life-os-becomes-the-life/2026-03-17T09:19:14.000Zhttps://terryli.ai/posts/where-gen-ai-is-actually-transformative/2026-03-14T00:00:00.000Zhttps://terryli.ai/posts/where-rules-live/2026-03-08T15:26:00.000Zhttps://terryli.ai/posts/where-the-harness-starts-to-matter/2026-07-19T00:48:05.000Zhttps://terryli.ai/posts/why-agents-break-governance/2026-04-18T12:00:00.000Zhttps://terryli.ai/posts/why-ai-demands-experiments/2026-03-17T01:52:58.000Zhttps://terryli.ai/posts/why-ai-makes-us-dumber/2026-03-06T18:30:00.000Zhttps://terryli.ai/posts/why-be-nice/2026-03-13T07:47:50.000Zhttps://terryli.ai/posts/why-i-didnt-package-my-ai-organism/2026-04-05T07:37:11.000Zhttps://terryli.ai/posts/why-i-stopped-optimising-my-morning-routine/2026-03-07T06:50:55.000Zhttps://terryli.ai/posts/why-nobody-benchmarks-memory/2026-03-13T06:25:00.000Zhttps://terryli.ai/posts/why-nobody-builds-cross-vendor-ai-orchestration/2026-03-13T08:18:00.000Zhttps://terryli.ai/posts/why-the-cell/2026-03-25T02:58:00.000Zhttps://terryli.ai/posts/workflows-not-containers/2026-03-24T04:31:03.000Zhttps://terryli.ai/posts/you-can-know-the-game-is-unfair/2026-03-14T00:00:00.000Zhttps://terryli.ai/posts/you-cant-ab-test-your-life/2026-03-14T00:00:00.000Zhttps://terryli.ai/posts/your-agent-pays-the-cold-start-tax-every-morning/2026-03-14T08:47:32.000Zhttps://terryli.ai/posts/your-ai-agents-quality-gate-is-lying-to-you/2026-04-08T00:00:00.000Zhttps://terryli.ai/posts/your-ai-did-the-research-you-didnt/2026-03-19T08:55:29.000Zhttps://terryli.ai/posts/your-ai-is-an-echo-chamber/2026-03-16T06:50:00.000Zhttps://terryli.ai/posts/your-ai-pipeline-is-probably-mapreduce/2026-03-14T16:00:00.000Zhttps://terryli.ai/posts/your-ai-risk-tier-is-probably-wrong/2026-03-18T05:15:00.000Zhttps://terryli.ai/posts/your-ai-roadmap-is-already-obsolete/2026-03-17T16:00:00.000Zhttps://terryli.ai/posts/your-ai-tools-should-watch-you-fumble/2026-03-16T01:51:36.000Zhttps://terryli.ai/posts/your-body-doesnt-care-what-youre-thinking-about/2026-03-13T09:12:17.000Zhttps://terryli.ai/posts/your-ground-truth-is-someone-elses-process-outcome/2026-03-18T04:20:00.000Zhttps://terryli.ai/posts/your-llm-review-missed-a-verbnoun-mismatch/2026-04-18T22:46:33.000Zhttps://terryli.ai/posts/your-output-is-your-selections/2026-03-13T03:37:46.000Zhttps://terryli.ai/posts/your-tool-shouldnt-know-what-to-ignore/2026-03-07T03:08:27.000Zhttps://terryli.ai/posts/your-variable-names-should-be-a-toy/2026-03-23T03:08:00.000Zhttps://terryli.ai/posts/your-wage-reflects-your-scarcity/2026-03-14T00:00:00.000Zhttps://terryli.ai/posts/your-wearable-doesnt-know-youre-tired/2026-03-18T10:59:19.000Zhttps://terryli.ai/search/https://terryli.ai/tags/https://terryli.ai/tags/abstraction/https://terryli.ai/tags/aesthetics/https://terryli.ai/tags/agent-architecture/https://terryli.ai/tags/agent-design/https://terryli.ai/tags/agent-experience/https://terryli.ai/tags/agentic-ai/https://terryli.ai/tags/agentic-engineering/https://terryli.ai/tags/agentic-workflows/https://terryli.ai/tags/agentic/https://terryli.ai/tags/agents/https://terryli.ai/tags/agents/2/https://terryli.ai/tags/agi/https://terryli.ai/tags/ai-agents/https://terryli.ai/tags/ai-agents/2/https://terryli.ai/tags/ai-agents/3/https://terryli.ai/tags/ai-architecture/https://terryli.ai/tags/ai-assistants/https://terryli.ai/tags/ai-assisted-development/https://terryli.ai/tags/ai-consulting/https://terryli.ai/tags/ai-controls/https://terryli.ai/tags/ai-ecosystem/https://terryli.ai/tags/ai-engineering/https://terryli.ai/tags/ai-governance/https://terryli.ai/tags/ai-governance/2/https://terryli.ai/tags/ai-ops/https://terryli.ai/tags/ai-research/https://terryli.ai/tags/ai-risk/https://terryli.ai/tags/ai-safety/https://terryli.ai/tags/ai-strategy/https://terryli.ai/tags/ai-systems/https://terryli.ai/tags/ai-tools/https://terryli.ai/tags/ai-workflow/https://terryli.ai/tags/ai/https://terryli.ai/tags/ai/2/https://terryli.ai/tags/ai/3/https://terryli.ai/tags/ai/4/https://terryli.ai/tags/ai/5/https://terryli.ai/tags/ai/6/https://terryli.ai/tags/ai/7/https://terryli.ai/tags/aml/https://terryli.ai/tags/analogy/https://terryli.ai/tags/analysis/https://terryli.ai/tags/anthropic/https://terryli.ai/tags/apac/https://terryli.ai/tags/api-design/https://terryli.ai/tags/architect-implementer/https://terryli.ai/tags/architecture/https://terryli.ai/tags/architecture/2/https://terryli.ai/tags/assurance/https://terryli.ai/tags/async/https://terryli.ai/tags/authentication/https://terryli.ai/tags/automation/https://terryli.ai/tags/autonomous-agents/https://terryli.ai/tags/autonomy/https://terryli.ai/tags/banking/https://terryli.ai/tags/benchmarks/https://terryli.ai/tags/biology/https://terryli.ai/tags/browser-automation/https://terryli.ai/tags/building/https://terryli.ai/tags/business-models/https://terryli.ai/tags/capability-assurance/https://terryli.ai/tags/career/https://terryli.ai/tags/careers/https://terryli.ai/tags/china/https://terryli.ai/tags/chinese-internet/https://terryli.ai/tags/claude-code/https://terryli.ai/tags/cli/https://terryli.ai/tags/code/https://terryli.ai/tags/coding-agents/https://terryli.ai/tags/cognition/https://terryli.ai/tags/cognitive-science/https://terryli.ai/tags/collaboration/https://terryli.ai/tags/communication/https://terryli.ai/tags/complexity/https://terryli.ai/tags/compliance/https://terryli.ai/tags/compound-engineering/https://terryli.ai/tags/compound-interest/https://terryli.ai/tags/consilium/https://terryli.ai/tags/consulting/https://terryli.ai/tags/consulting/2/https://terryli.ai/tags/context-engineering/https://terryli.ai/tags/context/https://terryli.ai/tags/controls/https://terryli.ai/tags/cost-optimization/https://terryli.ai/tags/craft/https://terryli.ai/tags/creativity/https://terryli.ai/tags/data-governance/https://terryli.ai/tags/data/https://terryli.ai/tags/debugging/https://terryli.ai/tags/decision-making/https://terryli.ai/tags/decisions/https://terryli.ai/tags/delegation/https://terryli.ai/tags/dependency-thinking/https://terryli.ai/tags/design-patterns/https://terryli.ai/tags/design-principles/https://terryli.ai/tags/design/https://terryli.ai/tags/design/2/https://terryli.ai/tags/developer-experience/https://terryli.ai/tags/developer-tools/https://terryli.ai/tags/development/https://terryli.ai/tags/devops/https://terryli.ai/tags/digital-transformation/https://terryli.ai/tags/displacement/https://terryli.ai/tags/dlt/https://terryli.ai/tags/documentation/https://terryli.ai/tags/economics/https://terryli.ai/tags/embeddings/https://terryli.ai/tags/emergence/https://terryli.ai/tags/engineering/https://terryli.ai/tags/engineering/2/https://terryli.ai/tags/enterprise-ai/https://terryli.ai/tags/enterprise/https://terryli.ai/tags/epistemics/https://terryli.ai/tags/epistemology/https://terryli.ai/tags/evals/https://terryli.ai/tags/evaluation/https://terryli.ai/tags/evidence/https://terryli.ai/tags/experiment/https://terryli.ai/tags/experimental-rigour/https://terryli.ai/tags/experimentation/https://terryli.ai/tags/experiments/https://terryli.ai/tags/fairness/https://terryli.ai/tags/finance/https://terryli.ai/tags/financial-services/https://terryli.ai/tags/fine-tuning/https://terryli.ai/tags/fodina/https://terryli.ai/tags/forecasting/https://terryli.ai/tags/foundation-models/https://terryli.ai/tags/frameworks/https://terryli.ai/tags/future-of-work/https://terryli.ai/tags/garden/https://terryli.ai/tags/garp/https://terryli.ai/tags/gen-ai/https://terryli.ai/tags/geopolitics/https://terryli.ai/tags/glm-5.1/https://terryli.ai/tags/governance/https://terryli.ai/tags/hallucination/https://terryli.ai/tags/health/https://terryli.ai/tags/heuristics/https://terryli.ai/tags/hiring/https://terryli.ai/tags/hitl/https://terryli.ai/tags/hkma/https://terryli.ai/tags/home/https://terryli.ai/tags/honesty/https://terryli.ai/tags/hong-kong/https://terryli.ai/tags/human-ai-collaboration/https://terryli.ai/tags/human-in-the-loop/https://terryli.ai/tags/ideation/https://terryli.ai/tags/independence/https://terryli.ai/tags/inference/https://terryli.ai/tags/infrastructure/https://terryli.ai/tags/intelligence/https://terryli.ai/tags/intent-classification/https://terryli.ai/tags/interface-design/https://terryli.ai/tags/interpretability/https://terryli.ai/tags/interviews/https://terryli.ai/tags/investing/https://terryli.ai/tags/knowledge-management/https://terryli.ai/tags/knowledge-mining/https://terryli.ai/tags/knowledge-work/https://terryli.ai/tags/langchain/https://terryli.ai/tags/language/https://terryli.ai/tags/leadership/https://terryli.ai/tags/learning/https://terryli.ai/tags/legacy/https://terryli.ai/tags/life/https://terryli.ai/tags/llm-review/https://terryli.ai/tags/llm/https://terryli.ai/tags/macro/https://terryli.ai/tags/management/https://terryli.ai/tags/mcp/https://terryli.ai/tags/memory/https://terryli.ai/tags/mental-models/https://terryli.ai/tags/meta-cognition/https://terryli.ai/tags/meta/https://terryli.ai/tags/metabolism/https://terryli.ai/tags/methodology/https://terryli.ai/tags/mindset/https://terryli.ai/tags/ml/https://terryli.ai/tags/moats/https://terryli.ai/tags/mobile/https://terryli.ai/tags/model-risk/https://terryli.ai/tags/monitoring/https://terryli.ai/tags/mtor/https://terryli.ai/tags/multi-agent/https://terryli.ai/tags/multi-model/https://terryli.ai/tags/naming/https://terryli.ai/tags/observability/https://terryli.ai/tags/observation/https://terryli.ai/tags/obsidian/https://terryli.ai/tags/open-source/https://terryli.ai/tags/openai/https://terryli.ai/tags/optionality/https://terryli.ai/tags/organisation/https://terryli.ai/tags/organism/https://terryli.ai/tags/parenting/https://terryli.ai/tags/patterns/https://terryli.ai/tags/personal-infrastructure/https://terryli.ai/tags/personal-systems/https://terryli.ai/tags/philosophy/https://terryli.ai/tags/planning/https://terryli.ai/tags/policy/https://terryli.ai/tags/postmortem/https://terryli.ai/tags/pptx/https://terryli.ai/tags/practice/https://terryli.ai/tags/preparation/https://terryli.ai/tags/pricing/https://terryli.ai/tags/production/https://terryli.ai/tags/productivity/https://terryli.ai/tags/programming/https://terryli.ai/tags/prompt-engineering/https://terryli.ai/tags/publishing/https://terryli.ai/tags/python/https://terryli.ai/tags/rag/https://terryli.ai/tags/reading/https://terryli.ai/tags/reflection/https://terryli.ai/tags/regulation/https://terryli.ai/tags/reliability/https://terryli.ai/tags/research/https://terryli.ai/tags/responsible-ai/https://terryli.ai/tags/retrieval/https://terryli.ai/tags/review/https://terryli.ai/tags/ribosome/https://terryli.ai/tags/risk-management/https://terryli.ai/tags/risk/https://terryli.ai/tags/rust/https://terryli.ai/tags/scaffolding/https://terryli.ai/tags/search/https://terryli.ai/tags/security/https://terryli.ai/tags/self-awareness/https://terryli.ai/tags/self-improvement/https://terryli.ai/tags/self-knowledge/https://terryli.ai/tags/skill-design/https://terryli.ai/tags/skills/https://terryli.ai/tags/software-engineering/https://terryli.ai/tags/software/https://terryli.ai/tags/sr-11-7/https://terryli.ai/tags/stakeholder/https://terryli.ai/tags/stories/https://terryli.ai/tags/strategy/https://terryli.ai/tags/structured-judgement/https://terryli.ai/tags/supply-chain/https://terryli.ai/tags/sycophancy/https://terryli.ai/tags/system-design/https://terryli.ai/tags/systems-design/https://terryli.ai/tags/systems-thinking/https://terryli.ai/tags/systems/https://terryli.ai/tags/taste/https://terryli.ai/tags/technology/https://terryli.ai/tags/temporal/https://terryli.ai/tags/testing/https://terryli.ai/tags/theory/https://terryli.ai/tags/thinking/https://terryli.ai/tags/token-efficiency/https://terryli.ai/tags/tokenisation/https://terryli.ai/tags/tooling/https://terryli.ai/tags/tools/https://terryli.ai/tags/tools/2/https://terryli.ai/tags/understanding/https://terryli.ai/tags/utilities/https://terryli.ai/tags/validation/https://terryli.ai/tags/value/https://terryli.ai/tags/values/https://terryli.ai/tags/vendor-risk/https://terryli.ai/tags/verification/https://terryli.ai/tags/vivesca/https://terryli.ai/tags/willison/https://terryli.ai/tags/work/https://terryli.ai/tags/workflow/https://terryli.ai/tags/workflows/https://terryli.ai/tags/writing/https://terryli.ai/tags/yagni/