Top all-time · #5816
Aleksa Gordić
Personal blog · aleksagordic.com/blogaiprogramming
Aleksa Gordić's personal blog at aleksagordic.com covers ai and programming. Since 2025, 14 different Hacker News readers have linked it 18 times, most often “vLLM: Anatomy of a High-Throughput LLM Inference System”. It is still writing: the last post was on 2026-07-14, and you can follow it by feed.
Too few sampled comments (3) to say how often HN recommends it.?
Why HN links it
Explainers on LLM internals, including transformers and vLLM-style inference engines. HN readers link them for technical background on how models and serving work.
Summarised by Claude from a sample of 3 HN comments that link here. Each comment is marked recommending, critical or neutral (citing without judging); how this works.
Aleksa Gordić's website.
Most-linked posts?
- vLLM: Anatomy of a High-Throughput LLM Inference System9 people · 9 mentions · first linked 2025-09-01
- The Transformer: The Life of a Token3 people · 3 mentions · first linked 2026-05-26
- TPU and GPU Clusters: The Anatomy of Collective Communication2 people · 2 mentions · first linked 2026-07-14
- GPUs: Anatomy of high performance matmul kernels2 people · 2 mentions · first linked 2025-09-29
What Hacker News says?
Yes it's 100% prioritization. Through that it's also likely running on more GPUs at once but that's an artifact of prioritization at the datacenter level. Any task coming into an AI datacenter atm is split into fairly fined grained chunks of work and added to queues to be processed. When you add a job with high priority all those chunks will be processed off the queue first by each and every GPU that frees up. It pr…
Comment linking /blog/vllm
on “Speed up responses with fast mode”Submitted GPUs: Anatomy of high performance matmul kernels as a story · 30 points
HN discussionA better blog on Transformers: https://www.aleksagordic.com/blog/transformer
Comment linking /blog/transformer
on “How LLMs Work”Submitted vLLM: Anatomy of a High-Throughput LLM Inference System as a story · 4 points
HN discussionThe whole thing feels AI written, generated from the codebase. For instance, it goes into (nano)vLLM internals and doesn’t mention PagedAttention once (one of the core ideas that vLLM is based on)[1]. Also mentions that Part 2 will cover dense vs MoE’s, which is weird because nanovllm hardcodes a dense Qwen3 into the source. Here are better (imo) explainers about how vLLM works: - https://hamzaelshafie.bearblog.dev/…
Comment linking /blog/vllm
on “Nano-vLLM: How a vLLM-style inference engine works”Submitted VLLM: Anatomy of a High-Throughput LLM Inference System as a story · 4 points
HN discussionSubmitted Nvidia GPUs: Anatomy of high performance matmul kernels as a story · 3 points
HN discussionSubmitted VLLM: Anatomy of a High-Throughput LLM Inference System as a story · 3 points
HN discussionSubmitted VLLM: Anatomy of a High-Throughput LLM Inference System as a story · 3 points
HN discussionSubmitted Neocambria as a story · 3 points
HN discussionSubmitted The Transformer: The Life of a Token as a story · 2 points
HN discussionSubmitted How vLLM Delivers High Throughput LLM Serving - An Engineer’s View as a story · 2 points
HN discussionSubmitted TPU and GPU Clusters: The Anatomy of Collective Communication as a story · 1 points
HN discussionSubmitted TPU and GPU Clusters: The Anatomy of Collective Communication as a story · 1 points
HN discussionSubmitted Inside The Transformer: The Life of a Token as a story · 1 points
HN discussionSubmitted From Oura Ring to Dataset: Analyzing and Visualizing a Year of Biometric Data as a story · 1 points
HN discussionSubmitted Inside vLLM: Anatomy of a High-Throughput LLM Inference System as a story · 1 points
HN discussionSubmitted VLLM: Anatomy of a High-Throughput LLM Inference System as a story · 1 points
HN discussion
