https://blog.yatharthlakhera.in/ 2026-06-26 https://blog.yatharthlakhera.in/archive 2026-06-25 https://blog.yatharthlakhera.in/bio 2026-06-25 https://blog.yatharthlakhera.in/subscribe 2026-06-25 https://blog.yatharthlakhera.in/authors 2026-07-29 https://blog.yatharthlakhera.in/authors/yatharth-lakhera 2026-06-13 https://blog.yatharthlakhera.in/p/the-biggest-speedup-in-local-inference-is-one-you-never-turned-on 2026-07-01 https://blog.yatharthlakhera.in/p/speculative-decoding-isn-t-free-it-s-a-bet-you-can-lose 2026-06-28 https://blog.yatharthlakhera.in/p/the-runtime-is-the-new-api-you-already-picked-one 2026-06-27 https://blog.yatharthlakhera.in/p/your-bandwidth-is-fixed-your-speed-isn-t 2026-06-24 https://blog.yatharthlakhera.in/p/how-models-get-smarter-without-getting-slower 2026-06-24 https://blog.yatharthlakhera.in/p/your-model-has-two-kinds-of-slow 2026-06-18 https://blog.yatharthlakhera.in/p/i-fixed-every-memory-problem-it-was-still-slow 2026-06-18 https://blog.yatharthlakhera.in/p/the-model-fit-in-memory-the-conversation-didn-t 2026-06-18 https://blog.yatharthlakhera.in/p/what-a-layer-in-an-llm-actually-is 2026-06-14 https://blog.yatharthlakhera.in/p/for-running-llms-your-ram-number-is-a-lie-look-at-the-vram 2026-06-14 https://blog.yatharthlakhera.in/p/i-bought-a-mac-mini-to-run-a-9b-model-hello-took-a-minute 2026-06-14 https://blog.yatharthlakhera.in/t/optimization-techniques 2026-06-24 https://blog.yatharthlakhera.in/t/inference-speed 2026-06-24 https://blog.yatharthlakhera.in/t/llm-concepts 2026-06-24 https://blog.yatharthlakhera.in/t/mac-mini-setup 2026-06-24