Top all-time · #8779
Aidan Cooper
Personal blog · aidancooper.co.ukaidata
Aidan Cooper's personal blog at aidancooper.co.uk covers ai and data. Since 2021, 9 different Hacker News readers have linked it 14 times, most often “A Guide to Structured Generation Using Constrained Decoding”. It is still writing: the last post was on 2026-09-27, and you can follow it by feed.
Too few sampled comments (2) to say how often HN recommends it.?
Why HN links it
HN readers link this blog's post on constrained decoding, the technique of restricting LLM output to regex, grammar or JSON-schema formats. It is offered as a starting point for that topic.
Summarised by Claude from a sample of 2 HN comments that link here. Each comment is marked recommending, critical or neutral (citing without judging); how this works.
Building biomedical/health AI systems using large language models. I write about AI/ML, data science, and software development.
Most-linked posts?
- A Guide to Structured Generation Using Constrained Decoding3 people · 3 mentions · first linked 2024-04-09
- Explaining ML Models: A Non-Technical Guide to Interpreting SHAP Analyses3 people · 3 mentions · first linked 2021-11-01
- 4 Pandas Anti-Patterns to Avoid and How to Fix Them2 people · 2 mentions · first linked 2022-07-19
What Hacker News says?
I'm a little confused why you have to specify "strict: true" to get this behavior. It is obviously always desired, I would be surprised for people to ever specify "strict: false". That API design leaves to be desired. I also learned about constrainted decoding[1], that they give a brief explanation about. This is a really clever technique! It will increase reliability as well as reduce latency (less tokens to pick f…
Comment linking /constrained-decoding/
on “Structured Outputs in the API”Submitted Explaining ML Models: A Non-Technical Guide to Interpreting SHAP Analyses as a story · 36 points
HN discussionI don’t know of one definitive resource but “constrained decoding”, regex/grammar-based decoding, json-schema decoding will give you a bunch of hits. Look up vLLM, SGLang and outlines’ implementations for more technical details. This looks pretty decent: https://www.aidancooper.co.uk/constrained-decoding/
Comment linking /constrained-decoding/
on “Jeff – Jev-compatible 0.8B decision models, trained at home, ~30 ms”Submitted A Non-Technical Guide to Interpreting SHAP Analyses as a story · 14 points
HN discussionSubmitted Could Vaccinating Young People Before Old People Save More Lives? as a story · 14 points
HN discussionSubmitted 4 Pandas Anti-Patterns to Avoid and How to Fix Them as a story · 4 points
HN discussionSubmitted A Guide to Structured Generation Using Constrained Decoding as a story · 2 points
HN discussionSubmitted Pandas Antipatterns as a story · 2 points
HN discussionSubmitted The Most Underused Email Feature: Schedule Send as a story · 2 points
HN discussionSubmitted LLM That Beat Frontier Models at Chess as a story · 1 points
HN discussionSubmitted Machine Learning in the Life Sciences Has a Data Problem as a story · 1 points
HN discussionSubmitted Supervised Clustering: How to Use SHAP Values for Better Cluster Analysis as a story · 1 points
HN discussionSubmitted Explaining Machine Learning Models: A Guide to Interpreting SHAP Analyses as a story · 1 points
HN discussionSubmitted Scots Are More Inclined to Skip the Gym and Watch the Footy Than English People as a story · 1 points
HN discussion