https://robotnurse.io/1.0 https://robotnurse.io/work/0.9 https://robotnurse.io/advise/0.8 https://robotnurse.io/buy/0.6 https://robotnurse.io/gear/0.7 https://robotnurse.io/gear/best-mac-studio-config/0.8 https://robotnurse.io/gear/mac-mini-vs-mac-pro-local-ai/0.8 https://robotnurse.io/gear/how-much-ram-for-local-llms/0.8 https://robotnurse.io/gear/amdgpu-vs-mac-local-inference/0.8 https://robotnurse.io/gear/mac-studio-vs-gpu-pc-local-ai/0.8 https://robotnurse.io/gear/best-local-ai-setup-under-3000/0.8 https://robotnurse.io/gear/ram-vs-vram-local-llms/0.8 https://robotnurse.io/gear/old-gaming-pc-local-ai/0.8 https://robotnurse.io/tools/0.6 https://robotnurse.io/tools/ram-calculator/0.7 https://robotnurse.io/tools/model-fit/0.7 https://robotnurse.io/tools/cost-calculator/0.7 https://robotnurse.io/runbooks/0.6 https://robotnurse.io/about https://robotnurse.io/contact/ https://robotnurse.io/book/ https://robotnurse.io/thanks/ https://robotnurse.io/audit/0.9 https://robotnurse.io/cpa/0.8 https://robotnurse.io/medical-ai/0.8 https://robotnurse.io/legal-ai/0.8 https://robotnurse.io/benchmarks/0.7 https://robotnurse.io/local-ai-setup/0.9 https://robotnurse.io/local-ai-setup-nyc/0.8 https://robotnurse.io/local-ai-setup-nj/0.8 https://robotnurse.io/packages/0.8 https://robotnurse.io/partners/0.7 https://robotnurse.io/ebook/0.7 https://robotnurse.io/posts/ https://robotnurse.io/faq https://robotnurse.io/privacy https://robotnurse.io/tos https://robotnurse.io/refund https://robotnurse.io/affiliate-disclosure https://robotnurse.io/hire https://robotnurse.io/archive https://robotnurse.io/daily/ https://robotnurse.io/daily/archive/ https://robotnurse.io/daily/archive/2026-09-10 https://robotnurse.io/daily/archive/2026-09-12 https://robotnurse.io/daily/archive/2026-09-13 https://robotnurse.io/daily/archive/2026-09-14 https://robotnurse.io/daily/archive/2026-09-15 https://robotnurse.io/daily/archive/2026-09-18 https://robotnurse.io/daily/archive/2026-09-19 https://robotnurse.io/daily/archive/2026-09-20 https://robotnurse.io/daily/archive/2026-09-25 https://robotnurse.io/daily/archive/2026-09-26 https://robotnurse.io/daily/archive/2026-09-28 https://robotnurse.io/daily/archive/2026-10-01 https://robotnurse.io/daily/archive/2026-10-09 https://robotnurse.io/daily/archive/2026-10-12 https://robotnurse.io/daily/archive/2026-10-15 https://robotnurse.io/daily/archive/2026-10-16 https://robotnurse.io/daily/archive/2026-10-19 https://robotnurse.io/checklist/ https://robotnurse.io/runbooks/apple-16gb/ https://robotnurse.io/runbooks/apple-32gb/ https://robotnurse.io/posts/01-model-stopped-saying-no2026-09-270.7 https://robotnurse.io/posts/22-zero-of-twenty-five-skills-were-loading2026-09-120.7 https://robotnurse.io/posts/21-json-mode-was-causing-the-malformed-json2026-09-130.7 https://robotnurse.io/posts/06-our-cluster-is-idle-97-percent-of-the-time2026-09-130.7 https://robotnurse.io/posts/04-benchmark-said-18-percent-production-said-0.072026-09-130.7 https://robotnurse.io/posts/23-the-machine-was-up-everything-on-it-was-dead2026-09-120.7 https://robotnurse.io/posts/02-cannot-prompt-your-way-to-abstention2026-09-130.7 https://robotnurse.io/posts/03-what-thinking-mode-actually-costs2026-09-130.7 https://robotnurse.io/posts/05-we-optimized-prefill-bottleneck-was-decode2026-09-130.7 https://robotnurse.io/posts/08-your-engine-isnt-hung-its-saturated2026-09-130.7 https://robotnurse.io/posts/17-every-surprising-eval-result-was-infrastructure2026-09-130.7 https://robotnurse.io/posts/18-a-safety-hook-truncated-our-code-generation2026-09-130.7 https://robotnurse.io/posts/07-one-slow-request-locks-out-everything2026-09-270.7 https://robotnurse.io/posts/26-most-of-our-tts-compute-was-thrown-away2026-09-130.7 https://robotnurse.io/posts/28-22-percent-of-our-training-set-was-invented2026-09-130.7 https://robotnurse.io/posts/31-a-9x-pessimization-that-looked-like-a-hardware-limit2026-09-130.7 https://robotnurse.io/posts/19-every-failure-silently-scored-0.52026-09-130.7 https://robotnurse.io/posts/27-the-fine-tune-was-dead-weight2026-09-130.7 https://robotnurse.io/posts/29-max-tokens-900-yielded-zero-pairs2026-09-130.7 https://robotnurse.io/posts/09-the-crash-we-could-not-reproduce2026-09-120.7 https://robotnurse.io/posts/13-one-abandoned-request-wedged-the-server2026-09-120.7 https://robotnurse.io/posts/30-the-instrument-was-broken2026-09-130.7 https://robotnurse.io/posts/14-one-env-var-cost-us-45x2026-09-130.7 https://robotnurse.io/posts/32-we-built-a-voice-clone-and-never-marked-it2026-09-130.7 https://robotnurse.io/posts/24-a-192gb-model-on-a-256gb-box2026-09-130.7 https://robotnurse.io/posts/36-four-wrong-diagnoses-of-a-locked-out-mac2026-09-130.7 https://robotnurse.io/posts/38-compression-reported-success-and-did-nothing2026-09-120.7 https://robotnurse.io/posts/39-four-reasons-your-config-edit-didnt-take2026-09-130.7 https://robotnurse.io/posts/40-the-429-that-was-really-a-4002026-09-130.7 https://robotnurse.io/posts/16-health-checker-manufactured-the-stalls2026-09-270.7 https://robotnurse.io/posts/20-unfusing-a-kernel-made-it-faster2026-09-130.7 https://robotnurse.io/posts/11-benchmark-the-hardware-first2026-09-130.7 https://robotnurse.io/posts/12-your-ssh-benchmark-runs-at-nice-52026-09-120.7 https://robotnurse.io/posts/15-prefix-cache-died-at-token-10672026-09-130.7 https://robotnurse.io/posts/41-our-baseline-was-the-three-worst-samples2026-09-130.7 https://robotnurse.io/posts/10-evaluating-an-inference-engine-and-rejecting-it2026-09-130.7 https://robotnurse.io/posts/25-two-x-decode-from-block-diffusion-drafting2026-09-130.7 https://robotnurse.io/posts/35-our-kernel-benchmark-measured-the-system-cache2026-09-120.7 https://robotnurse.io/posts/33-we-wrote-our-own-inference-engine2026-09-270.7 https://robotnurse.io/posts/37-the-watchdog-starves-with-everything-else2026-09-130.7 https://robotnurse.io/posts/42-five-parallel-agents-zero-commits2026-09-120.7 https://robotnurse.io/posts/34-the-chunked-prefill-fix-was-13x-slower2026-09-130.7 https://robotnurse.io/posts/45-the-neural-engine-as-a-prefill-coprocessor2026-09-130.7 https://robotnurse.io/posts/43-every-gpu-oom-was-a-sysctl2026-09-130.7 https://robotnurse.io/posts/44-17-gb-of-attention-buffers-nobody-read2026-09-130.7 https://robotnurse.io/posts/46-can-ssd-offload-fit-a-model-twice-your-ram2026-09-250.7 https://robotnurse.io/posts/47-our-benchmark-queues-deadlocked-on-each-others-pgrep2026-09-120.7 https://robotnurse.io/posts/48-five-kernel-ideas-that-measured-neutral-or-worse2026-09-250.7 https://robotnurse.io/posts/49-we-pinned-a-hash-of-floating-point-output2026-09-130.7 https://robotnurse.io/posts/50-hipaa-safe-ai-in-a-small-medical-practice2026-09-270.7 https://robotnurse.io/posts/51-client-confidential-ai-for-law-firms-and-cpas2026-09-120.7 https://robotnurse.io/posts/52-on-prem-vs-cloud-ai-tco-over-3-years2026-09-270.7 https://robotnurse.io/posts/53-what-a-local-ai-setup-project-costs2026-09-120.7 https://robotnurse.io/posts/54-replacing-chatgpt-with-a-local-model2026-09-120.7 https://robotnurse.io/posts/55-sizing-a-local-ai-box-for-a-small-office2026-09-120.7 https://robotnurse.io/posts/running-a-private-llm-for-a-small-law-firm-on-a-mac-studio2026-09-140.7 https://robotnurse.io/posts/we-indexed-45-local-inference-benchmark-results-the-4-7x-gap2026-09-140.7 https://robotnurse.io/posts/is-local-ai-right-for-your-small-business2026-10-170.7 https://robotnurse.io/posts/local-ai-on-a-compact-device2026-10-170.7 https://robotnurse.io/posts/ollama-for-beginners-local-llms2026-10-170.7 https://robotnurse.io/posts/run-local-ai-on-a-mac-realistic-guide2026-10-170.7 https://robotnurse.io/posts/what-a-private-ai-setup-is2026-10-170.7 https://robotnurse.io/posts/58-our-two-machines-ran-different-models-with-the-same-name2026-09-190.7 https://robotnurse.io/posts/59-speculative-decoding-wins-alone-and-loses-batched2026-09-250.7 https://robotnurse.io/posts/60-your-headline-tokens-per-second-is-a-short-prompt-number2026-09-250.7 https://robotnurse.io/posts/56-every-sub-4-bit-moe-number-was-uncalibrated2026-09-190.7 https://robotnurse.io/posts/57-we-published-a-finding-and-withdrew-it-the-next-day2026-09-270.7 https://robotnurse.io/posts/61-our-conversion-ran-five-times-at-once2026-09-270.7 https://robotnurse.io/posts/62-we-sent-a-v4-model-a-v3-prompt2026-09-270.7 https://robotnurse.io/posts/63-the-tag-was-128-tokens-back2026-09-270.7 https://robotnurse.io/series/the-model-zoo https://robotnurse.io/series/local-stack https://robotnurse.io/series/inference-engines https://robotnurse.io/series/running-it-daily https://robotnurse.io/series/hardware-notes https://robotnurse.io/series/quantization https://robotnurse.io/series/buyer-s-guide https://robotnurse.io/series/standalone https://robotnurse.io/series/ https://robotnurse.io/tags/